abs
stringlengths 44
64
| Download PDF
stringlengths 75
115
| OpenReview
stringlengths 42
42
| title
stringlengths 15
148
| url
stringlengths 44
64
| authors
stringlengths 6
903
| detail_url
stringlengths 44
64
| tags
stringclasses 1
value | abstract
stringlengths 422
5.84k
|
|---|---|---|---|---|---|---|---|---|
https://proceedings.mlr.press/v235/garber24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/garber24a/garber24a.pdf
|
https://openreview.net/forum?id=RnbobOgbn0
|
Projection-Free Online Convex Optimization with Time-Varying Constraints
|
https://proceedings.mlr.press/v235/garber24a.html
|
Dan Garber, Ben Kretzu
|
https://proceedings.mlr.press/v235/garber24a.html
|
ICML 2024
|
We consider the setting of online convex optimization with adversarial time-varying constraints in which actions must be feasible w.r.t. a fixed constraint set, and are also required on average to approximately satisfy additional time-varying constraints. Motivated by scenarios in which the fixed feasible set (hard constraint) is difficult to project on, we consider projection-free algorithms that access this set only through a linear optimization oracle (LOO). We present an algorithm that, on a sequence of length $T$ and using overall $T$ calls to the LOO, guarantees $\tilde{O}(T^{3/4})$ regret w.r.t. the losses and $O(T^{7/8})$ constraints violation (ignoring all quantities except for $T$). In particular, these bounds hold w.r.t. any interval of the sequence. This algorithm however also requires access to an oracle for minimizing a strongly convex nonsmooth function over a Euclidean ball. We present a more efficient algorithm that does not require the latter optimization oracle but only first-order access to the time-varying constraints, and achieves similar bounds w.r.t. the entire sequence. We extend the latter to the setting of bandit feedback and obtain similar bounds (as a function of $T$) in expectation.
|
https://proceedings.mlr.press/v235/garcin24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/garcin24a/garcin24a.pdf
|
https://openreview.net/forum?id=uku9r6RROl
|
DRED: Zero-Shot Transfer in Reinforcement Learning via Data-Regularised Environment Design
|
https://proceedings.mlr.press/v235/garcin24a.html
|
Samuel Garcin, James Doran, Shangmin Guo, Christopher G. Lucas, Stefano V Albrecht
|
https://proceedings.mlr.press/v235/garcin24a.html
|
ICML 2024
|
Autonomous agents trained using deep reinforcement learning (RL) often lack the ability to successfully generalise to new environments, even when these environments share characteristics with the ones they have encountered during training. In this work, we investigate how the sampling of individual environment instances, or levels, affects the zero-shot generalisation (ZSG) ability of RL agents. We discover that, for deep actor-critic architectures sharing their base layers, prioritising levels according to their value loss minimises the mutual information between the agent’s internal representation and the set of training levels in the generated training data. This provides a novel theoretical justification for the regularisation achieved by certain adaptive sampling strategies. We then turn our attention to unsupervised environment design (UED) methods, which assume control over level generation. We find that existing UED methods can significantly shift the training distribution, which translates to low ZSG performance. To prevent both overfitting and distributional shift, we introduce data-regularised environment design (DRED). DRED generates levels using a generative model trained to approximate the ground truth distribution of an initial set of level parameters. Through its grounding, DRED achieves significant improvements in ZSG over adaptive level sampling strategies and UED methods.
|
https://proceedings.mlr.press/v235/gardner24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/gardner24a/gardner24a.pdf
|
https://openreview.net/forum?id=HvwOtYzHBX
|
LLark: A Multimodal Instruction-Following Language Model for Music
|
https://proceedings.mlr.press/v235/gardner24a.html
|
Joshua P Gardner, Simon Durand, Daniel Stoller, Rachel M Bittner
|
https://proceedings.mlr.press/v235/gardner24a.html
|
ICML 2024
|
Music has a unique and complex structure which is challenging for both expert humans and existing AI systems to understand, and presents unique challenges relative to other forms of audio. We present LLark, an instruction-tuned multimodal model for music understanding. We detail our process for dataset creation, which involves augmenting the annotations of diverse open-source music datasets and converting them to a unified instruction-tuning format. We propose a multimodal architecture for LLark, integrating a pretrained generative model for music with a pretrained language model. In evaluations on three types of tasks (music understanding, captioning, reasoning), we show that LLark matches or outperforms existing baselines in music understanding, and that humans show a high degree of agreement with its responses in captioning and reasoning tasks. LLark is trained entirely from open-source music data and models, and we make our training code available along with the release of this paper. Additional results and audio examples are at https://bit.ly/llark, and our source code is available at https://github.com/spotify-research/llark.
|
https://proceedings.mlr.press/v235/garg24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/garg24a/garg24a.pdf
|
https://openreview.net/forum?id=WQbDS9RydY
|
Memorization Through the Lens of Curvature of Loss Function Around Samples
|
https://proceedings.mlr.press/v235/garg24a.html
|
Isha Garg, Deepak Ravikumar, Kaushik Roy
|
https://proceedings.mlr.press/v235/garg24a.html
|
ICML 2024
|
Deep neural networks are over-parameterized and easily overfit to and memorize the datasets that they train on. In the extreme case, it has been shown that networks can memorize a randomly labeled dataset. In this paper, we propose using the curvature of the loss function around each training sample, averaged over training epochs, as a measure of memorization of a sample. We show that this curvature metric effectively captures memorization statistics, both qualitatively and quantitatively in popular image datasets. We provide quantitative validation of the proposed metric against memorization scores released by Feldman & Zhang (2020). Further, experiments on mislabeled data detection show that corrupted samples are learned with high curvature and using curvature for identifying mislabelled examples outperforms existing approaches. Qualitatively, we find that high curvature samples correspond to long-tailed, mislabeled, or conflicting instances, indicating a likelihood of memorization. Notably, this analysis helps us find, to the best of our knowledge, a novel failure mode on the CIFAR100 and ImageNet datasets: that of duplicated images with differing labels.
|
https://proceedings.mlr.press/v235/gatmiry24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/gatmiry24a/gatmiry24a.pdf
|
https://openreview.net/forum?id=VUTyzH63Xa
|
Simplicity Bias via Global Convergence of Sharpness Minimization
|
https://proceedings.mlr.press/v235/gatmiry24a.html
|
Khashayar Gatmiry, Zhiyuan Li, Sashank J. Reddi, Stefanie Jegelka
|
https://proceedings.mlr.press/v235/gatmiry24a.html
|
ICML 2024
|
The remarkable generalization ability of neural networks is usually attributed to the implicit bias of SGD, which often yields models with lower complexity using simpler (e.g. linear) and low-rank features. Recent works have provided empirical and theoretical evidence for the bias of particular variants of SGD (such as label noise SGD) toward flatter regions of the loss landscape. Despite the folklore intuition that flat solutions are ’simple’, the connection with the simplicity of the final trained model (e.g. low-rank) is not well understood. In this work, we take a step toward bridging this gap by studying the simplicity structure that arises from minimizers of the sharpness for a class of two-layer neural networks. We show that, for any high dimensional training data and certain activations, with small enough step size, label noise SGD always converges to a network that replicates a single linear feature across all neurons; thereby implying a simple rank one feature matrix. To obtain this result, our main technical contribution is to show that label noise SGD always minimizes the sharpness on the manifold of models with zero loss for two-layer networks. Along the way, we discover a novel property — a local geodesic convexity — of the trace of Hessian of the loss at approximate stationary points on the manifold of zero loss, which links sharpness to the geometry of the manifold. This tool may be of independent interest.
|
https://proceedings.mlr.press/v235/gatmiry24b.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/gatmiry24b/gatmiry24b.pdf
|
https://openreview.net/forum?id=o8AaRKbP9K
|
Can Looped Transformers Learn to Implement Multi-step Gradient Descent for In-context Learning?
|
https://proceedings.mlr.press/v235/gatmiry24b.html
|
Khashayar Gatmiry, Nikunj Saunshi, Sashank J. Reddi, Stefanie Jegelka, Sanjiv Kumar
|
https://proceedings.mlr.press/v235/gatmiry24b.html
|
ICML 2024
|
Transformers to do reasoning and few-shot learning, without any fine-tuning, is widely conjectured to stem from their ability to implicitly simulate a multi-step algorithms – such as gradient descent – with their weights in a single forward pass. Recently, there has been progress in understanding this complex phenomenon from an expressivity point of view, by demonstrating that Transformers can express such multi-step algorithms. However, our knowledge about the more fundamental aspect of its learnability, beyond single layer models, is very limited. In particular, can training Transformers enable convergence to algorithmic solutions? In this work we resolve this for in context linear regression with linear looped Transformers – a multi-layer model with weight sharing that is conjectured to have an inductive bias to learn fix-point iterative algorithms. More specifically, for this setting we show that the global minimizer of the population training loss implements multi-step preconditioned gradient descent, with a preconditioner that adapts to the data distribution. Furthermore, we show a fast convergence for gradient flow on the regression loss, despite the non-convexity of the landscape, by proving a novel gradient dominance condition. To our knowledge, this is the first theoretical analysis for multi-layer Transformer in this setting. We further validate our theoretical findings through synthetic experiments.
|
https://proceedings.mlr.press/v235/gaur24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/gaur24a/gaur24a.pdf
|
https://openreview.net/forum?id=rJxFvAs7pq
|
Closing the Gap: Achieving Global Convergence (Last Iterate) of Actor-Critic under Markovian Sampling with Neural Network Parametrization
|
https://proceedings.mlr.press/v235/gaur24a.html
|
Mudit Gaur, Amrit Bedi, Di Wang, Vaneet Aggarwal
|
https://proceedings.mlr.press/v235/gaur24a.html
|
ICML 2024
|
The current state-of-the-art theoretical analysis of Actor-Critic (AC) algorithms significantly lags in addressing the practical aspects of AC implementations. This crucial gap needs bridging to bring the analysis in line with practical implementations of AC. To address this, we advocate for considering the MMCLG criteria: Multi-layer neural network parametrization for actor/critic, Markovian sampling, Continuous state-action spaces, the performance of the Last iterate, and Global optimality. These aspects are practically significant and have been largely overlooked in existing theoretical analyses of AC algorithms. In this work, we address these gaps by providing the first comprehensive theoretical analysis of AC algorithms that encompasses all five crucial practical aspects (covers MMCLG criteria). We establish global convergence sample complexity bounds of $\tilde{\mathcal{O}}\left( \epsilon^{-3} \right)$. We achieve this result through our novel use of the weak gradient domination property of MDP’s and our unique analysis of the error in critic estimation.
|
https://proceedings.mlr.press/v235/gautam24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/gautam24a/gautam24a.pdf
|
https://openreview.net/forum?id=VHO4nE7v41
|
Variance-reduced Zeroth-Order Methods for Fine-Tuning Language Models
|
https://proceedings.mlr.press/v235/gautam24a.html
|
Tanmay Gautam, Youngsuk Park, Hao Zhou, Parameswaran Raman, Wooseok Ha
|
https://proceedings.mlr.press/v235/gautam24a.html
|
ICML 2024
|
Fine-tuning language models (LMs) has demonstrated success in a wide array of downstream tasks. However, as LMs are scaled up, the memory requirements for backpropagation become prohibitively high. Zeroth-order (ZO) optimization methods can leverage memory-efficient forward passes to estimate gradients. More recently, MeZO, an adaptation of ZO-SGD, has been shown to consistently outperform zero-shot and in-context learning when combined with suitable task prompts. In this work, we couple ZO methods with variance reduction techniques to enhance stability and convergence for inference-based LM fine-tuning. We introduce Memory-Efficient Zeroth-Order Stochastic Variance-Reduced Gradient (MeZO-SVRG) and demonstrate its efficacy across multiple LM fine-tuning tasks, eliminating the reliance on task-specific prompts. Evaluated across a range of both masked and autoregressive LMs on benchmark GLUE tasks, MeZO-SVRG outperforms MeZO with up to 20% increase in test accuracies in both full- and partial-parameter fine-tuning settings. MeZO-SVRG benefits from reduced computation time as it often surpasses MeZO’s peak test accuracy with a $2\times$ reduction in GPU-hours. MeZO-SVRG significantly reduces the required memory footprint compared to first-order SGD, i.e. by $2\times$ for autoregressive models. Our experiments highlight that MeZO-SVRG’s memory savings progressively improve compared to SGD with larger batch sizes.
|
https://proceedings.mlr.press/v235/gavranovic24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/gavranovic24a/gavranovic24a.pdf
|
https://openreview.net/forum?id=EIcxV7T0Sy
|
Position: Categorical Deep Learning is an Algebraic Theory of All Architectures
|
https://proceedings.mlr.press/v235/gavranovic24a.html
|
Bruno Gavranović, Paul Lessard, Andrew Joseph Dudzik, Tamara Von Glehn, João Guilherme Madeira Araújo, Petar Veličković
|
https://proceedings.mlr.press/v235/gavranovic24a.html
|
ICML 2024
|
We present our position on the elusive quest for a general-purpose framework for specifying and studying deep learning architectures. Our opinion is that the key attempts made so far lack a coherent bridge between specifying constraints which models must satisfy and specifying their implementations. Focusing on building a such a bridge, we propose to apply category theory—precisely, the universal algebra of monads valued in a 2-category of parametric maps—as a single theory elegantly subsuming both of these flavours of neural network design. To defend our position, we show how this theory recovers constraints induced by geometric deep learning, as well as implementations of many architectures drawn from the diverse landscape of neural networks, such as RNNs. We also illustrate how the theory naturally encodes many standard constructs in computer science and automata theory.
|
https://proceedings.mlr.press/v235/ge24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/ge24a/ge24a.pdf
|
https://openreview.net/forum?id=tya725xlZ3
|
Masked Face Recognition with Generative-to-Discriminative Representations
|
https://proceedings.mlr.press/v235/ge24a.html
|
Shiming Ge, Weijia Guo, Chenyu Li, Zhang Junzheng, Yong Li, Dan Zeng
|
https://proceedings.mlr.press/v235/ge24a.html
|
ICML 2024
|
Masked face recognition is important for social good but challenged by diverse occlusions that cause insufficient or inaccurate representations. In this work, we propose a unified deep network to learn generative-to-discriminative representations for facilitating masked face recognition. To this end, we split the network into three modules and learn them on synthetic masked faces in a greedy module-wise pretraining manner. First, we leverage a generative encoder pretrained for face inpainting and finetune it to represent masked faces into category-aware descriptors. Attribute to the generative encoder’s ability in recovering context information, the resulting descriptors can provide occlusion-robust representations for masked faces, mitigating the effect of diverse masks. Then, we incorporate a multi-layer convolutional network as a discriminative reformer and learn it to convert the category-aware descriptors into identity-aware vectors, where the learning is effectively supervised by distilling relation knowledge from off-the-shelf face recognition model. In this way, the discriminative reformer together with the generative encoder serves as the pretrained backbone, providing general and discriminative representations towards masked faces. Finally, we cascade one fully-connected layer following by one softmax layer into a feature classifier and finetune it to identify the reformed identity-aware vectors. Extensive experiments on synthetic and realistic datasets demonstrate the effectiveness of our approach in recognizing masked faces.
|
https://proceedings.mlr.press/v235/ge24b.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/ge24b/ge24b.pdf
|
https://openreview.net/forum?id=JV84NVo1em
|
Safe and Robust Subgame Exploitation in Imperfect Information Games
|
https://proceedings.mlr.press/v235/ge24b.html
|
Zhenxing Ge, Zheng Xu, Tianyu Ding, Linjian Meng, Bo An, Wenbin Li, Yang Gao
|
https://proceedings.mlr.press/v235/ge24b.html
|
ICML 2024
|
Opponent exploitation is an important task for players to exploit the weaknesses of others in games. Existing approaches mainly focus on balancing between exploitation and exploitability but are often vulnerable to modeling errors and deceptive adversaries. To address this problem, our paper offers a novel perspective on the safety of opponent exploitation, named Adaptation Safety. This concept leverages the insight that strategies, even those not explicitly aimed at opponent exploitation, may inherently be exploitable due to computational complexities, rendering traditional safety overly rigorous. In contrast, adaptation safety requires that the strategy should not be more exploitable than it would be in scenarios where opponent exploitation is not considered. Building on such adaptation safety, we further propose an Opponent eXploitation Search (OX-Search) framework by incorporating real-time search techniques for efficient online opponent exploitation. Moreover, we provide theoretical analyses to show the adaptation safety and robust exploitation of OX-Search, even with inaccurate opponent models. Empirical evaluations in popular poker games demonstrate OX-Search’s superiority in both exploitability and exploitation compared to previous methods.
|
https://proceedings.mlr.press/v235/gedon24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/gedon24a/gedon24a.pdf
|
https://openreview.net/forum?id=M4ejBhNNrn
|
No Double Descent in Principal Component Regression: A High-Dimensional Analysis
|
https://proceedings.mlr.press/v235/gedon24a.html
|
Daniel Gedon, Antonio H. Ribeiro, Thomas B. Schön
|
https://proceedings.mlr.press/v235/gedon24a.html
|
ICML 2024
|
Understanding the generalization properties of large-scale models necessitates incorporating realistic data assumptions into the analysis. Therefore, we consider Principal Component Regression (PCR)—combining principal component analysis and linear regression—on data from a low-dimensional manifold. We present an analysis of PCR when the data is sampled from a spiked covariance model, obtaining fundamental asymptotic guarantees for the generalization risk of this model. Our analysis is based on random matrix theory and allows us to provide guarantees for high-dimensional data. We additionally present an analysis of the distribution shift between training and test data. The results allow us to disentangle the effects of (1) the number of parameters, (2) the data-generating model and, (3) model misspecification on the generalization risk. The use of PCR effectively regularizes the model and prevents the interpolation peak of the double descent. Our theoretical findings are empirically validated in simulation, demonstrating their practical relevance.
|
https://proceedings.mlr.press/v235/geirhos24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/geirhos24a/geirhos24a.pdf
|
https://openreview.net/forum?id=s0Jvdolv2I
|
Don’t trust your eyes: on the (un)reliability of feature visualizations
|
https://proceedings.mlr.press/v235/geirhos24a.html
|
Robert Geirhos, Roland S. Zimmermann, Blair Bilodeau, Wieland Brendel, Been Kim
|
https://proceedings.mlr.press/v235/geirhos24a.html
|
ICML 2024
|
How do neural networks extract patterns from pixels? Feature visualizations attempt to answer this important question by visualizing highly activating patterns through optimization. Today, visualization methods form the foundation of our knowledge about the internal workings of neural networks, as a type of mechanistic interpretability. Here we ask: How reliable are feature visualizations? We start our investigation by developing network circuits that trick feature visualizations into showing arbitrary patterns that are completely disconnected from normal network behavior on natural input. We then provide evidence for a similar phenomenon occurring in standard, unmanipulated networks: feature visualizations are processed very differently from standard input, casting doubt on their ability to "explain" how neural networks process natural images. This can be used as a sanity check for feature visualizations. We underpin our empirical findings by theory proving that the set of functions that can be reliably understood by feature visualization is extremely small and does not include general black-box neural networks. Therefore, a promising way forward could be the development of networks that enforce certain structures in order to ensure more reliable feature visualizations.
|
https://proceedings.mlr.press/v235/geist24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/geist24a/geist24a.pdf
|
https://openreview.net/forum?id=L0VoOdjCUb
|
Learning with 3D rotations, a hitchhiker’s guide to SO(3)
|
https://proceedings.mlr.press/v235/geist24a.html
|
Andreas René Geist, Jonas Frey, Mikel Zhobro, Anna Levina, Georg Martius
|
https://proceedings.mlr.press/v235/geist24a.html
|
ICML 2024
|
Many settings in machine learning require the selection of a rotation representation. However, choosing a suitable representation from the many available options is challenging. This paper acts as a survey and guide through rotation representations. We walk through their properties that harm or benefit deep learning with gradient-based optimization. By consolidating insights from rotation-based learning, we provide a comprehensive overview of learning functions with rotation representations. We provide guidance on selecting representations based on whether rotations are in the model’s input or output and whether the data primarily comprises small angles.
|
https://proceedings.mlr.press/v235/genalti24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/genalti24a/genalti24a.pdf
|
https://openreview.net/forum?id=bPsohGR6gD
|
Graph-Triggered Rising Bandits
|
https://proceedings.mlr.press/v235/genalti24a.html
|
Gianmarco Genalti, Marco Mussi, Nicola Gatti, Marcello Restelli, Matteo Castiglioni, Alberto Maria Metelli
|
https://proceedings.mlr.press/v235/genalti24a.html
|
ICML 2024
|
In this paper, we propose a novel generalization of rested and restless bandits where the evolution of the arms’ expected rewards is governed by a graph defined over the arms. An edge connecting a pair of arms $(i,j)$ represents the fact that a pull of arm $i$ triggers the evolution of arm $j$, and vice versa. Interestingly, rested and restless bandits are both special cases of our model for some suitable (degenerate) graphs. Still, the model can represent way more general and interesting scenarios. We first tackle the problem of computing the optimal policy when no specific structure is assumed on the graph, showing that it is NP-hard. Then, we focus on a specific structure forcing the graph to be composed of a set of fully connected subgraphs (i.e., cliques), and we prove that the optimal policy can be easily computed in closed form. Then, we move to the learning problem presenting regret minimization algorithms for deterministic and stochastic cases. Our regret bounds highlight the complexity of the learning problem by incorporating instance-dependent terms that encode specific properties of the underlying graph structure. Moreover, we illustrate how the knowledge of the underlying graph is not necessary for achieving the no-regret property.
|
https://proceedings.mlr.press/v235/geng24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/geng24a/geng24a.pdf
|
https://openreview.net/forum?id=QXEx16jWdN
|
Improving Adversarial Energy-Based Model via Diffusion Process
|
https://proceedings.mlr.press/v235/geng24a.html
|
Cong Geng, Tian Han, Peng-Tao Jiang, Hao Zhang, Jinwei Chen, Søren Hauberg, Bo Li
|
https://proceedings.mlr.press/v235/geng24a.html
|
ICML 2024
|
Generative models have shown strong generation ability while efficient likelihood estimation is less explored. Energy-based models (EBMs) define a flexible energy function to parameterize unnormalized densities efficiently but are notorious for being difficult to train. Adversarial EBMs introduce a generator to form a minimax training game to avoid expensive MCMC sampling used in traditional EBMs, but a noticeable gap between adversarial EBMs and other strong generative models still exists. Inspired by diffusion-based models, we embedded EBMs into each denoising step to split a long-generated process into several smaller steps. Besides, we employ a symmetric Jeffrey divergence and introduce a variational posterior distribution for the generator’s training to address the main challenges that exist in adversarial EBMs. Our experiments show significant improvement in generation compared to existing adversarial EBMs, while also providing a useful energy function for efficient density estimation.
|
https://proceedings.mlr.press/v235/geng24b.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/geng24b/geng24b.pdf
|
https://openreview.net/forum?id=AJGwSx0RUV
|
Reinforcement Learning within Tree Search for Fast Macro Placement
|
https://proceedings.mlr.press/v235/geng24b.html
|
Zijie Geng, Jie Wang, Ziyan Liu, Siyuan Xu, Zhentao Tang, Mingxuan Yuan, Jianye Hao, Yongdong Zhang, Feng Wu
|
https://proceedings.mlr.press/v235/geng24b.html
|
ICML 2024
|
Macro placement is a crucial step in modern chip design, and reinforcement learning (RL) has recently emerged as a promising technique for improving the placement quality. However, existing RL-based techniques are hindered by their low sample efficiency, requiring numerous online rollouts or substantial offline expert data to achieve bootstrap, which are often impractical in industrial scenarios. To address this challenge, we propose a novel sample-efficient framework, namely EfficientPlace, for fast macro placement. EfficientPlace integrates a global tree search algorithm to strategically direct the optimization process, as well as a RL agent for local policy learning to advance the tree search. Experiments on commonly used benchmarks demonstrate that EfficientPlace achieves remarkable placement quality within a short timeframe, outperforming recent state-of-the-art approaches.
|
https://proceedings.mlr.press/v235/georgiev24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/georgiev24a/georgiev24a.pdf
|
https://openreview.net/forum?id=2FHWFG5ahw
|
Adaptive Horizon Actor-Critic for Policy Learning in Contact-Rich Differentiable Simulation
|
https://proceedings.mlr.press/v235/georgiev24a.html
|
Ignat Georgiev, Krishnan Srinivasan, Jie Xu, Eric Heiden, Animesh Garg
|
https://proceedings.mlr.press/v235/georgiev24a.html
|
ICML 2024
|
Model-Free Reinforcement Learning (MFRL), leveraging the policy gradient theorem, has demonstrated considerable success in continuous control tasks. However, these approaches are plagued by high gradient variance due to zeroth-order gradient estimation, resulting in suboptimal policies. Conversely, First-Order Model-Based Reinforcement Learning (FO-MBRL) methods employing differentiable simulation provide gradients with reduced variance but are susceptible to sampling error in scenarios involving stiff dynamics, such as physical contact. This paper investigates the source of this error and introduces Adaptive Horizon Actor-Critic (AHAC), an FO-MBRL algorithm that reduces gradient error by adapting the model-based horizon to avoid stiff dynamics. Empirical findings reveal that AHAC outperforms MFRL baselines, attaining 40% more reward across a set of locomotion tasks and efficiently scaling to high-dimensional control environments with improved wall-clock-time efficiency. adaptive-horizon-actor-critic.github.io
|
https://proceedings.mlr.press/v235/gerken24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/gerken24a/gerken24a.pdf
|
https://openreview.net/forum?id=plXXbXjvQ9
|
Emergent Equivariance in Deep Ensembles
|
https://proceedings.mlr.press/v235/gerken24a.html
|
Jan E Gerken, Pan Kessel
|
https://proceedings.mlr.press/v235/gerken24a.html
|
ICML 2024
|
We show that deep ensembles become equivariant for all inputs and at all training times by simply using data augmentation. Crucially, equivariance holds off-manifold and for any architecture in the infinite width limit. The equivariance is emergent in the sense that predictions of individual ensemble members are not equivariant but their collective prediction is. Neural tangent kernel theory is used to derive this result and we verify our theoretical insights using detailed numerical experiments.
|
https://proceedings.mlr.press/v235/ghandeharioun24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/ghandeharioun24a/ghandeharioun24a.pdf
|
https://openreview.net/forum?id=5uwBzcn885
|
Patchscopes: A Unifying Framework for Inspecting Hidden Representations of Language Models
|
https://proceedings.mlr.press/v235/ghandeharioun24a.html
|
Asma Ghandeharioun, Avi Caciularu, Adam Pearce, Lucas Dixon, Mor Geva
|
https://proceedings.mlr.press/v235/ghandeharioun24a.html
|
ICML 2024
|
Understanding the internal representations of large language models (LLMs) can help explain models’ behavior and verify their alignment with human values. Given the capabilities of LLMs in generating human-understandable text, we propose leveraging the model itself to explain its internal representations in natural language. We introduce a framework called Patchscopes and show how it can be used to answer a wide range of questions about an LLM’s computation. We show that many prior interpretability methods based on projecting representations into the vocabulary space and intervening on the LLM computation can be viewed as instances of this framework. Moreover, several of their shortcomings such as failure in inspecting early layers or lack of expressivity can be mitigated by Patchscopes. Beyond unifying prior inspection techniques, Patchscopes also opens up new possibilities such as using a more capable model to explain the representations of a smaller model, and multihop reasoning error correction.
|
https://proceedings.mlr.press/v235/ghazi24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/ghazi24a/ghazi24a.pdf
|
https://openreview.net/forum?id=zfmwAaB9Nw
|
Individualized Privacy Accounting via Subsampling with Applications in Combinatorial Optimization
|
https://proceedings.mlr.press/v235/ghazi24a.html
|
Badih Ghazi, Pritish Kamath, Ravi Kumar, Pasin Manurangsi, Adam Sealfon
|
https://proceedings.mlr.press/v235/ghazi24a.html
|
ICML 2024
|
In this work, we give a new technique for analyzing individualized privacy accounting via the following simple observation: if an algorithm is one-sided add-DP, then its subsampled variant satisfies two-sided DP. From this, we obtain several improved algorithms for private combinatorial optimization problems, including decomposable submodular maximization and set cover. Our error guarantees are asymptotically tight and our algorithm satisfies pure-DP while previously known algorithms (Gupta et al., 2010; Chaturvedi et al., 2021) are approximate-DP. We also show an application of our technique beyond combinatorial optimization by giving a pure-DP algorithm for the shifting heavy hitter problem in a stream; previously, only an approximate-DP algorithm was known (Kaplan et al., 2021; Cohen & Lyu, 2023).
|
https://proceedings.mlr.press/v235/ghimire24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/ghimire24a/ghimire24a.pdf
|
https://openreview.net/forum?id=bcN7KSB2YS
|
State-Constrained Zero-Sum Differential Games with One-Sided Information
|
https://proceedings.mlr.press/v235/ghimire24a.html
|
Mukesh Ghimire, Lei Zhang, Zhe Xu, Yi Ren
|
https://proceedings.mlr.press/v235/ghimire24a.html
|
ICML 2024
|
We study zero-sum differential games with state constraints and one-sided information, where the informed player (Player 1) has a categorical payoff type unknown to the uninformed player (Player 2). The goal of Player 1 is to minimize his payoff without violating the constraints, while that of Player 2 is to violate the state constraints if possible, or to maximize the payoff otherwise. One example of the game is a man-to-man matchup in football. Without state constraints, Cardaliaguet (2007) showed that the value of such a game exists and is convex to the common belief of players. Our theoretical contribution is an extension of this result to games with state constraints and the derivation of the primal and dual subdynamic principles necessary for computing behavioral strategies. Different from existing works that are concerned about the scalability of no-regret learning in games with discrete dynamics, our study reveals the underlying structure of strategies for belief manipulation resulting from information asymmetry and state constraints. This structure will be necessary for scalable learning on games with continuous actions and long time windows. We use a simplified football game to demonstrate the utility of this work, where we reveal player positions and belief states in which the attacker should (or should not) play specific random deceptive moves to take advantage of information asymmetry, and compute how the defender should respond.
|
https://proceedings.mlr.press/v235/ghosal24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/ghosal24a/ghosal24a.pdf
|
https://openreview.net/forum?id=cPsn9AcOYh
|
Understanding Finetuning for Factual Knowledge Extraction
|
https://proceedings.mlr.press/v235/ghosal24a.html
|
Gaurav Rohit Ghosal, Tatsunori Hashimoto, Aditi Raghunathan
|
https://proceedings.mlr.press/v235/ghosal24a.html
|
ICML 2024
|
In this work, we study the impact of QA fine-tuning data on downstream factuality. We show that fine-tuning on lesser-known facts that are poorly stored during pretraining yields significantly worse factuality than fine-tuning on well-known facts, even when all facts are seen during pretraining. We prove this phenomenon theoretically, showing that training on lesser-known facts can lead the model to ignore subject entity names and instead output a generic plausible response even when the relevant factual knowledge is encoded in the model. On three question answering benchmarks (PopQA, Entity Questions, and MMLU) and two language models (Llama-2-7B and Mistral-7B), we find that (i) finetuning on a completely factual but lesser-known subset of the data deteriorates downstream factuality (5-10%) and (ii) finetuning on a subset of better-known examples matches or outperforms finetuning on the entire dataset. Ultimately, our results shed light on the interaction between pretrained knowledge and finetuning data and demonstrate the importance of taking into account how facts are stored in the pretrained model when fine-tuning for knowledge-intensive tasks.
|
https://proceedings.mlr.press/v235/ghosh24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/ghosh24a/ghosh24a.pdf
|
https://openreview.net/forum?id=XkHJo8iXGQ
|
A Closer Look at the Limitations of Instruction Tuning
|
https://proceedings.mlr.press/v235/ghosh24a.html
|
Sreyan Ghosh, Chandra Kiran Reddy Evuru, Sonal Kumar, Ramaneswaran S, Deepali Aneja, Zeyu Jin, Ramani Duraiswami, Dinesh Manocha
|
https://proceedings.mlr.press/v235/ghosh24a.html
|
ICML 2024
|
Instruction Tuning (IT), the process of training large language models (LLMs) using instruction-response pairs, has emerged as the predominant method for transforming base pre-trained LLMs into open-domain conversational agents. While IT has achieved notable success and widespread adoption, its limitations and shortcomings remain underexplored. In this paper, through rigorous experiments and an in-depth analysis of the changes LLMs undergo through IT, we reveal various limitations of IT. In particular, we show that (1) IT fails to enhance knowledge or skills in LLMs. LoRA fine-tuning is limited to learning response initiation and style tokens, and full-parameter fine-tuning leads to knowledge degradation. (2) Copying response patterns from IT datasets derived from knowledgeable sources leads to a decline in response quality. (3) Full-parameter fine-tuning increases hallucination by inaccurately borrowing tokens from conceptually similar instances in the IT dataset for generating responses. (4) Popular methods to improve IT do not lead to performance improvements over a simple LoRA fine-tuned model. Our findings reveal that responses generated solely from pre-trained knowledge consistently outperform responses by models that learn any form of new knowledge from IT on open-source datasets. We hope the insights and challenges revealed in this paper inspire future work in related directions.
|
https://proceedings.mlr.press/v235/ghosh24b.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/ghosh24b/ghosh24b.pdf
|
https://openreview.net/forum?id=eo88noTbb5
|
Agnostic Learning of Mixed Linear Regressions with EM and AM Algorithms
|
https://proceedings.mlr.press/v235/ghosh24b.html
|
Avishek Ghosh, Arya Mazumdar
|
https://proceedings.mlr.press/v235/ghosh24b.html
|
ICML 2024
|
Mixed linear regression is a well-studied problem in parametric statistics and machine learning. Given a set of samples, tuples of covariates and labels, the task of mixed linear regression is to find a small list of linear relationships that best fit the samples. Usually it is assumed that the label is generated stochastically by randomly selecting one of two or more linear functions, applying this chosen function to the covariates, and potentially introducing noise to the result. In that situation, the objective is to estimate the ground-truth linear functions up to some parameter error. The popular expectation maximization (EM) and alternating minimization (AM) algorithms have been previously analyzed for this. In this paper, we consider the more general problem of agnostic learning of mixed linear regression from samples, without such generative models. In particular, we show that the AM and EM algorithms, under standard conditions of separability and good initialization, lead to agnostic learning in mixed linear regression by converging to the population loss minimizers, for suitably defined loss functions. In some sense, this shows the strength of AM and EM algorithms that converges to “optimal solutions” even in the absence of realizable generative models.
|
https://proceedings.mlr.press/v235/ghosh24c.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/ghosh24c/ghosh24c.pdf
|
https://openreview.net/forum?id=sO5qtpvsUZ
|
Optimal Eye Surgeon: Finding image priors through sparse generators at initialization
|
https://proceedings.mlr.press/v235/ghosh24c.html
|
Avrajit Ghosh, Xitong Zhang, Kenneth K. Sun, Qing Qu, Saiprasad Ravishankar, Rongrong Wang
|
https://proceedings.mlr.press/v235/ghosh24c.html
|
ICML 2024
|
We introduce Optimal Eye Surgeon (OES), a framework for pruning and training deep image generator networks. Typically, untrained deep convolutional networks, which include image sampling operations, serve as effective image priors. However, they tend to overfit to noise in image restoration tasks due to being overparameterized. OES addresses this by adaptively pruning networks at random initialization to a level of underparameterization. This process effectively captures low-frequency image components even without training, by just masking. When trained to fit noisy image, these pruned subnetworks, which we term Sparse-DIP, resist overfitting to noise. This benefit arises from underparameterization and the regularization effect of masking, constraining them in the manifold of image priors. We demonstrate that subnetworks pruned through OES surpass other leading pruning methods, such as the Lottery Ticket Hypothesis, which is known to be suboptimal for image recovery tasks. Our extensive experiments demonstrate the transferability of OES-masks and the characteristics of sparse-subnetworks for image generation. Code is available at https://github.com/Avra98/Optimal-Eye-Surgeon.
|
https://proceedings.mlr.press/v235/gillman24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/gillman24a/gillman24a.pdf
|
https://openreview.net/forum?id=i0nVanexij
|
Self-Correcting Self-Consuming Loops for Generative Model Training
|
https://proceedings.mlr.press/v235/gillman24a.html
|
Nate Gillman, Michael Freeman, Daksh Aggarwal, Chia-Hong Hsu, Calvin Luo, Yonglong Tian, Chen Sun
|
https://proceedings.mlr.press/v235/gillman24a.html
|
ICML 2024
|
As synthetic data becomes higher quality and proliferates on the internet, machine learning models are increasingly trained on a mix of human- and machine-generated data. Despite the successful stories of using synthetic data for representation learning, using synthetic data for generative model training creates “self-consuming loops” which may lead to training instability or even collapse, unless certain conditions are met. Our paper aims to stabilize self-consuming generative model training. Our theoretical results demonstrate that by introducing an idealized correction function, which maps a data point to be more likely under the true data distribution, self-consuming loops can be made exponentially more stable. We then propose self-correction functions, which rely on expert knowledge (e.g. the laws of physics programmed in a simulator), and aim to approximate the idealized corrector automatically and at scale. We empirically validate the effectiveness of self-correcting self-consuming loops on the challenging human motion synthesis task, and observe that it successfully avoids model collapse, even when the ratio of synthetic data to real data is as high as 100%.
|
https://proceedings.mlr.press/v235/glaser24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/glaser24a/glaser24a.pdf
|
https://openreview.net/forum?id=2dlmcTXfcY
|
Kernel-Based Evaluation of Conditional Biological Sequence Models
|
https://proceedings.mlr.press/v235/glaser24a.html
|
Pierre Glaser, Steffanie Paul, Alissa M Hummer, Charlotte Deane, Debora Susan Marks, Alan Nawzad Amin
|
https://proceedings.mlr.press/v235/glaser24a.html
|
ICML 2024
|
We propose a set of kernel-based tools to evaluate the designs and tune the hyperparameters of conditional sequence models, with a focus on problems in computational biology. The backbone of our tools is a new measure of discrepancy between the true conditional distribution and the model’s estimate, called the Augmented Conditional Maximum Mean Discrepancy (ACMMD). Provided that the model can be sampled from, the ACMMD can be estimated unbiasedly from data to quantify absolute model fit, integrated within hypothesis tests, and used to evaluate model reliability. We demonstrate the utility of our approach by analyzing a popular protein design model, ProteinMPNN. We are able to reject the hypothesis that ProteinMPNN fits its data for various protein families, and tune the model’s temperature hyperparameter to achieve a better fit.
|
https://proceedings.mlr.press/v235/gloeckle24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/gloeckle24a/gloeckle24a.pdf
|
https://openreview.net/forum?id=pEWAcejiU2
|
Better & Faster Large Language Models via Multi-token Prediction
|
https://proceedings.mlr.press/v235/gloeckle24a.html
|
Fabian Gloeckle, Badr Youbi Idrissi, Baptiste Roziere, David Lopez-Paz, Gabriel Synnaeve
|
https://proceedings.mlr.press/v235/gloeckle24a.html
|
ICML 2024
|
Large language models such as GPT and Llama are trained with a next-token prediction loss. In this work, we suggest that training language models to predict multiple future tokens at once results in higher sample efficiency. More specifically, at each position in the training corpus, we ask the model to predict the following $n$ tokens using $n$ independent output heads, operating on top of a shared model trunk. Considering multi-token prediction as an auxiliary training task, we measure improved downstream capabilities with no overhead in training time for both code and natural language models. The method is increasingly useful for larger model sizes, and keeps its appeal when training for multiple epochs. Gains are especially pronounced on generative benchmarks like coding, where our models consistently outperform strong baselines by several percentage points. Our 13B parameter models solves 12% more problems on Human Eval and 17% more on MBPP than comparable next-token models. Experiments on small algorithmic tasks demonstrate that multi-token prediction is favorable for the development of induction heads and algorithmic reasoning capabilities. As an additional benefit, models trained with 4-token prediction are up to $3\times$ faster at inference, even with large batch sizes.
|
https://proceedings.mlr.press/v235/gloeckler24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/gloeckler24a/gloeckler24a.pdf
|
https://openreview.net/forum?id=DL79HYCFFq
|
All-in-one simulation-based inference
|
https://proceedings.mlr.press/v235/gloeckler24a.html
|
Manuel Gloeckler, Michael Deistler, Christian Dietrich Weilbach, Frank Wood, Jakob H. Macke
|
https://proceedings.mlr.press/v235/gloeckler24a.html
|
ICML 2024
|
Amortized Bayesian inference trains neural networks to solve stochastic inference problems using model simulations, thereby making it possible to rapidly perform Bayesian inference for any newly observed data. However, current simulation-based amortized inference methods are simulation-hungry and inflexible: They require the specification of a fixed parametric prior, simulator, and inference tasks ahead of time. Here, we present a new amortized inference method—the Simformer—which overcomes these limitations. By training a probabilistic diffusion model with transformer architectures, the Simformer outperforms current state-of-the-art amortized inference approaches on benchmark tasks and is substantially more flexible: It can be applied to models with function-valued parameters, it can handle inference scenarios with missing or unstructured data, and it can sample arbitrary conditionals of the joint distribution of parameters and data, including both posterior and likelihood. We showcase the performance and flexibility of the Simformer on simulators from ecology, epidemiology, and neuroscience, and demonstrate that it opens up new possibilities and application domains for amortized Bayesian inference on simulation-based models.
|
https://proceedings.mlr.press/v235/glukhov24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/glukhov24a/glukhov24a.pdf
|
https://openreview.net/forum?id=j5csKrtyAe
|
Position: Fundamental Limitations of LLM Censorship Necessitate New Approaches
|
https://proceedings.mlr.press/v235/glukhov24a.html
|
David Glukhov, Ilia Shumailov, Yarin Gal, Nicolas Papernot, Vardan Papyan
|
https://proceedings.mlr.press/v235/glukhov24a.html
|
ICML 2024
|
Large language models (LLMs) have exhibited impressive capabilities in comprehending complex instructions. However, their blind adherence to provided instructions has led to concerns regarding risks of malicious use. Existing defence mechanisms, such as model fine-tuning or output censorship methods have proven to be fallible at ensuring that LLMs do not return semantically impermissible responses. We present fundamental limitations of verifying the semantic properties of LLM outputs and identifying compositional threats, illustrating inherent challenges of current approaches to censoring LLM outputs. Specifically, we demonstrate that semantic censorship can be perceived as an undecidable problem, and semantic properties of LLM outputs can become impossible to verify when the LLM is capable of providing "encrypted" outputs. We further show challenges of censorship can extend beyond just semantic censorship, as attackers can reconstruct impermissible outputs from a collection of permissible ones. Consequently, we call for a re-evaluation of the problem of censorship and its goals, stressing the need for new definitions and approaches to censorship. In addition, we provide an initial attempt toward achieving this goal through syntactic censorship, drawing from a security perspective to design censorship methods that can provide guarantees.
|
https://proceedings.mlr.press/v235/goldblum24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/goldblum24a/goldblum24a.pdf
|
https://openreview.net/forum?id=EaJ7nqJ2Fa
|
Position: The No Free Lunch Theorem, Kolmogorov Complexity, and the Role of Inductive Biases in Machine Learning
|
https://proceedings.mlr.press/v235/goldblum24a.html
|
Micah Goldblum, Marc Anton Finzi, Keefer Rowan, Andrew Gordon Wilson
|
https://proceedings.mlr.press/v235/goldblum24a.html
|
ICML 2024
|
No free lunch theorems for supervised learning state that no learner can solve all problems or that all learners achieve exactly the same accuracy on average over a uniform distribution on learning problems. Accordingly, these theorems are often referenced in support of the notion that individual problems require specially tailored inductive biases. While virtually all uniformly sampled datasets have high complexity, real-world problems disproportionately generate low-complexity data, and we argue that neural network models share this same preference, formalized using Kolmogorov complexity. Notably, we show that architectures designed for a particular domain, such as computer vision, can compress datasets on a variety of seemingly unrelated domains. Our experiments show that pre-trained and even randomly initialized language models prefer to generate low-complexity sequences. Whereas no free lunch theorems seemingly indicate that individual problems require specialized learners, we explain how tasks that often require human intervention such as picking an appropriately sized model when labeled data is scarce or plentiful can be automated into a single learning algorithm. These observations justify the trend in deep learning of unifying seemingly disparate problems with an increasingly small set of machine learning models.
|
https://proceedings.mlr.press/v235/gong24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/gong24a/gong24a.pdf
|
https://openreview.net/forum?id=YuNFJSEkTi
|
CasCast: Skillful High-resolution Precipitation Nowcasting via Cascaded Modelling
|
https://proceedings.mlr.press/v235/gong24a.html
|
Junchao Gong, Lei Bai, Peng Ye, Wanghan Xu, Na Liu, Jianhua Dai, Xiaokang Yang, Wanli Ouyang
|
https://proceedings.mlr.press/v235/gong24a.html
|
ICML 2024
|
Precipitation nowcasting based on radar data plays a crucial role in extreme weather prediction and has broad implications for disaster management. Despite progresses have been made based on deep learning, two key challenges of precipitation nowcasting are not well-solved: (i) the modeling of complex precipitation system evolutions with different scales, and (ii) accurate forecasts for extreme precipitation. In this work, we propose CasCast, a cascaded framework composed of a deterministic and a probabilistic part to decouple the predictions for mesoscale precipitation distributions and small-scale patterns. Then, we explore training the cascaded framework at the high resolution and conducting the probabilistic modeling in a low dimensional latent space with a frame-wise-guided diffusion transformer for enhancing the optimization of extreme events while reducing computational costs. Extensive experiments on three benchmark radar precipitation datasets show that CasCast achieves competitive performance. Especially, CasCast significantly surpasses the baseline (up to +91.8%) for regional extreme-precipitation nowcasting.
|
https://proceedings.mlr.press/v235/gong24b.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/gong24b/gong24b.pdf
|
https://openreview.net/forum?id=drjjxmi2Ha
|
Does Label Smoothing Help Deep Partial Label Learning?
|
https://proceedings.mlr.press/v235/gong24b.html
|
Xiuwen Gong, Nitin Bisht, Guandong Xu
|
https://proceedings.mlr.press/v235/gong24b.html
|
ICML 2024
|
Although deep partial label learning (deep PLL) classifiers have shown their competitive performance, they are heavily influenced by the noisy false-positive labels leading to poorer performance as the training progresses. Meanwhile, existing deep PLL research lacks theoretical guarantee on the analysis of correlation between label noise (or ambiguity degree) and classification performance. This paper addresses the above limitations with label smoothing (LS) from both theoretical and empirical aspects. In theory, we prove lower and upper bounds of the expected risk to show that label smoothing can help deep PLL. We further derive the optimal smoothing rate to investigate the conditions, i.e., when label smoothing benefits deep PLL. In practice, we design a benchmark solution and a novel optimization algorithm called Label Smoothing-based Partial Label Learning (LS-PLL). Extensive experimental results on benchmark PLL datasets and various deep architectures validate that label smoothing does help deep PLL in improving classification performance and learning distinguishable representations, and the best results can be achieved when the empirical smoothing rate approximately approaches the optimal smoothing rate in theoretical findings. Code is publicly available at https://github.com/kalpiree/LS-PLL.
|
https://proceedings.mlr.press/v235/gong24c.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/gong24c/gong24c.pdf
|
https://openreview.net/forum?id=cBWVJh5Fvf
|
AST-T5: Structure-Aware Pretraining for Code Generation and Understanding
|
https://proceedings.mlr.press/v235/gong24c.html
|
Linyuan Gong, Mostafa Elhoushi, Alvin Cheung
|
https://proceedings.mlr.press/v235/gong24c.html
|
ICML 2024
|
Large language models (LLMs) have made significant advancements in code-related tasks, yet many LLMs treat code as simple sequences, neglecting its structured nature. We introduce AST-T5, a novel pretraining paradigm that leverages the Abstract Syntax Tree (AST) for enhanced code generation, transpilation, and understanding. Using dynamic programming, our AST-Aware Segmentation retains code structure, while our AST-Aware Span Corruption objective equips the model to reconstruct various code structures. Unlike other models, AST-T5 avoids complex program analyses or architectural changes, so it integrates seamlessly with any encoder-decoder Transformer. Evaluations show that AST-T5 consistently outperforms similar-sized LMs across various code-related tasks including HumanEval and MBPP. Structure-awareness makes AST-T5 particularly powerful in code-to-code tasks, surpassing CodeT5 by 2 points in exact match score for the Bugs2Fix task and by 3 points in exact match score for Java-C# Transpilation in CodeXGLUE. Our code and model are publicly available at https://github.com/gonglinyuan/ast_t5.
|
https://proceedings.mlr.press/v235/gong24d.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/gong24d/gong24d.pdf
|
https://openreview.net/forum?id=36rWa8zVkh
|
A Nearly Optimal Single Loop Algorithm for Stochastic Bilevel Optimization under Unbounded Smoothness
|
https://proceedings.mlr.press/v235/gong24d.html
|
Xiaochuan Gong, Jie Hao, Mingrui Liu
|
https://proceedings.mlr.press/v235/gong24d.html
|
ICML 2024
|
This paper studies the problem of stochastic bilevel optimization where the upper-level function is nonconvex with potentially unbounded smoothness and the lower-level function is strongly convex. This problem is motivated by meta-learning applied to sequential data, such as text classification using recurrent neural networks, where the smoothness constant of the upper-level loss function scales linearly with the gradient norm and can be potentially unbounded. Existing algorithm crucially relies on the nested loop design, which requires significant tuning efforts and is not practical. In this paper, we address this issue by proposing a Single Loop bIlevel oPtimizer (SLIP). The proposed algorithm first updates the lower-level variable by a few steps of stochastic gradient descent, and then simultaneously updates the upper-level variable by normalized stochastic gradient descent with momentum and the lower-level variable by stochastic gradient descent. Under standard assumptions, we show that our algorithm finds an $\epsilon$-stationary point within $\widetilde{O}(1/\epsilon^4)$[Here $\widetilde{O}(\cdot)$ compresses logarithmic factors of $1/\epsilon$ and $1/\delta$, where $\delta\in(0,1)$ denotes the failure probability.] oracle calls of stochastic gradient or Hessian-vector product, both in expectation and with high probability. This complexity result is nearly optimal up to logarithmic factors without mean-square smoothness of the stochastic gradient oracle. Our proof relies on (i) a refined characterization and control of the lower-level variable and (ii) establishing a novel connection between bilevel optimization and stochastic optimization under distributional drift. Our experiments on various tasks show that our algorithm significantly outperforms strong baselines in bilevel optimization.
|
https://proceedings.mlr.press/v235/gong24e.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/gong24e/gong24e.pdf
|
https://openreview.net/forum?id=y5L8W0KRUX
|
Evolution-Inspired Loss Functions for Protein Representation Learning
|
https://proceedings.mlr.press/v235/gong24e.html
|
Chengyue Gong, Adam Klivans, James Madigan Loy, Tianlong Chen, Qiang Liu, Daniel Jesus Diaz
|
https://proceedings.mlr.press/v235/gong24e.html
|
ICML 2024
|
AI-based frameworks for protein engineering use self-supervised learning (SSL) to obtain representations for downstream mutation effect predictions. The most common training objective for these methods is wildtype accuracy: given a sequence or structure where a wildtype residue has been masked, predict the missing amino acid. Wildtype accuracy, however, does not align with the primary goal of protein engineering, which is to suggest a mutation rather than to identify what already appears in nature. Here we present Evolutionary Ranking (EvoRank), a training objective that incorporates evolutionary information derived from multiple sequence alignments (MSAs) to learn more diverse protein representations. EvoRank corresponds to ranking amino-acid likelihoods in the probability distribution induced by an MSA. This objective forces models to learn the underlying evolutionary dynamics of a protein. Across a variety of phenotypes and datasets, we demonstrate that EvoRank leads to dramatic improvements in zero-shot performance and can compete with models fine-tuned on experimental data. This is particularly important in protein engineering, where it is expensive to obtain data for fine-tuning.
|
https://proceedings.mlr.press/v235/gong24f.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/gong24f/gong24f.pdf
|
https://openreview.net/forum?id=jKYyFbH8ap
|
Evaluation of LLMs on Syntax-Aware Code Fill-in-the-Middle Tasks
|
https://proceedings.mlr.press/v235/gong24f.html
|
Linyuan Gong, Sida Wang, Mostafa Elhoushi, Alvin Cheung
|
https://proceedings.mlr.press/v235/gong24f.html
|
ICML 2024
|
We introduce Syntax-Aware Fill-in-the-Middle (SAFIM), a new benchmark for evaluating Large Language Models (LLMs) on the code Fill-in-the-Middle (FIM) task. This benchmark focuses on syntax-aware completions of program structures such as code blocks and conditional expressions, and includes 17,720 examples from multiple programming languages, sourced from recent code submissions after April 2022 to minimize data contamination. SAFIM provides a robust framework with various prompt designs and novel syntax-aware post-processing techniques, facilitating accurate and fair comparisons across LLMs. Our comprehensive evaluation of 15 LLMs shows that FIM pretraining not only enhances FIM proficiency but also improves Left-to-Right (L2R) inference using LLMs. Our findings challenge conventional beliefs and suggest that pretraining methods and data quality have more impact than model size. SAFIM thus serves as a foundational platform for future research in effective pretraining strategies for code LLMs. The evaluation toolkit and dataset are available at https://github.com/gonglinyuan/safim, and the leaderboard is available at https://safimbenchmark.com.
|
https://proceedings.mlr.press/v235/gong24g.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/gong24g/gong24g.pdf
|
https://openreview.net/forum?id=lrPrkWXqzd
|
E$^2$GAN: Efficient Training of Efficient GANs for Image-to-Image Translation
|
https://proceedings.mlr.press/v235/gong24g.html
|
Yifan Gong, Zheng Zhan, Qing Jin, Yanyu Li, Yerlan Idelbayev, Xian Liu, Andrey Zharkov, Kfir Aberman, Sergey Tulyakov, Yanzhi Wang, Jian Ren
|
https://proceedings.mlr.press/v235/gong24g.html
|
ICML 2024
|
One highly promising direction for enabling flexible real-time on-device image editing is utilizing data distillation by leveraging large-scale text-to-image diffusion models to generate paired datasets used for training generative adversarial networks (GANs). This approach notably alleviates the stringent requirements typically imposed by high-end commercial GPUs for performing image editing with diffusion models. However, unlike text-to-image diffusion models, each distilled GAN is specialized for a specific image editing task, necessitating costly training efforts to obtain models for various concepts. In this work, we introduce and address a novel research direction: can the process of distilling GANs from diffusion models be made significantly more efficient? To achieve this goal, we propose a series of innovative techniques. First, we construct a base GAN model with generalized features, adaptable to different concepts through fine-tuning, eliminating the need for training from scratch. Second, we identify crucial layers within the base GAN model and employ Low-Rank Adaptation (LoRA) with a simple yet effective rank search process, rather than fine-tuning the entire base model. Third, we investigate the minimal amount of data necessary for fine-tuning, further reducing the overall training time. Extensive experiments show that we can efficiently empower GANs with the ability to perform real-time high-quality image editing on mobile devices with remarkably reduced training and storage costs for each concept.
|
https://proceedings.mlr.press/v235/gorbunov24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/gorbunov24a/gorbunov24a.pdf
|
https://openreview.net/forum?id=DBI6AuCD4a
|
High-Probability Convergence for Composite and Distributed Stochastic Minimization and Variational Inequalities with Heavy-Tailed Noise
|
https://proceedings.mlr.press/v235/gorbunov24a.html
|
Eduard Gorbunov, Abdurakhmon Sadiev, Marina Danilova, Samuel Horváth, Gauthier Gidel, Pavel Dvurechensky, Alexander Gasnikov, Peter Richtárik
|
https://proceedings.mlr.press/v235/gorbunov24a.html
|
ICML 2024
|
High-probability analysis of stochastic first-order optimization methods under mild assumptions on the noise has been gaining a lot of attention in recent years. Typically, gradient clipping is one of the key algorithmic ingredients to derive good high-probability guarantees when the noise is heavy-tailed. However, if implemented naively, clipping can spoil the convergence of the popular methods for composite and distributed optimization (Prox-SGD/Parallel SGD) even in the absence of any noise. Due to this reason, many works on high-probability analysis consider only unconstrained non-distributed problems, and the existing results for composite/distributed problems do not include some important special cases (like strongly convex problems) and are not optimal. To address this issue, we propose new stochastic methods for composite and distributed optimization based on the clipping of stochastic gradient differences and prove tight high-probability convergence results (including nearly optimal ones) for the new methods. In addition, we also develop new methods for composite and distributed variational inequalities and analyze the high-probability convergence of these methods.
|
https://proceedings.mlr.press/v235/goring24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/goring24a/goring24a.pdf
|
https://openreview.net/forum?id=xTYIAD2NND
|
Out-of-Domain Generalization in Dynamical Systems Reconstruction
|
https://proceedings.mlr.press/v235/goring24a.html
|
Niclas Alexander Göring, Florian Hess, Manuel Brenner, Zahra Monfared, Daniel Durstewitz
|
https://proceedings.mlr.press/v235/goring24a.html
|
ICML 2024
|
In science we are interested in finding the governing equations, the dynamical rules, underlying empirical phenomena. While traditionally scientific models are derived through cycles of human insight and experimentation, recently deep learning (DL) techniques have been advanced to reconstruct dynamical systems (DS) directly from time series data. State-of-the-art dynamical systems reconstruction (DSR) methods show promise in capturing invariant and long-term properties of observed DS, but their ability to generalize to unobserved domains remains an open challenge. Yet, this is a crucial property we would expect from any viable scientific theory. In this work, we provide a formal framework that addresses generalization in DSR. We explain why and how out-of-domain (OOD) generalization (OODG) in DSR profoundly differs from OODG considered elsewhere in machine learning. We introduce mathematical notions based on topological concepts and ergodic theory to formalize the idea of learnability of a DSR model. We formally prove that black-box DL techniques, without adequate structural priors, generally will not be able to learn a generalizing DSR model. We also show this empirically, considering major classes of DSR algorithms proposed so far, and illustrate where and why they fail to generalize across the whole phase space. Our study provides the first comprehensive mathematical treatment of OODG in DSR, and gives a deeper conceptual understanding of where the fundamental problems in OODG lie and how they could possibly be addressed in practice.
|
https://proceedings.mlr.press/v235/goswami24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/goswami24a/goswami24a.pdf
|
https://openreview.net/forum?id=FVvf69a5rx
|
MOMENT: A Family of Open Time-series Foundation Models
|
https://proceedings.mlr.press/v235/goswami24a.html
|
Mononito Goswami, Konrad Szafer, Arjun Choudhry, Yifu Cai, Shuo Li, Artur Dubrawski
|
https://proceedings.mlr.press/v235/goswami24a.html
|
ICML 2024
|
We introduce MOMENT, a family of open-source foundation models for general-purpose time series analysis. Pre-training large models on time series data is challenging due to (1) the absence of a large and cohesive public time series repository, and (2) diverse time series characteristics which make multi-dataset training onerous. Additionally, (3) experimental benchmarks to evaluate these models, especially in scenarios with limited resources, time, and supervision, are still in their nascent stages. To address these challenges, we compile a large and diverse collection of public time series, called the Time series Pile, and systematically tackle time series-specific challenges to unlock large-scale multi-dataset pre-training. Finally, we build on recent work to design a benchmark to evaluate time series foundation models on diverse tasks and datasets in limited supervision settings. Experiments on this benchmark demonstrate the effectiveness of our pre-trained models with minimal data and task-specific fine-tuning. Finally, we present several interesting empirical observations about large pre-trained time series models. Pre-trained models (AutonLab/MOMENT-1-large) and Time Series Pile (AutonLab/Timeseries-PILE) are available on Huggingface.
|
https://proceedings.mlr.press/v235/gottlieb24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/gottlieb24a/gottlieb24a.pdf
|
https://openreview.net/forum?id=BoPj12CnAn
|
Weighted distance nearest neighbor condensing
|
https://proceedings.mlr.press/v235/gottlieb24a.html
|
Lee-Ad Gottlieb, Timor Sharabi, Roi Weiss
|
https://proceedings.mlr.press/v235/gottlieb24a.html
|
ICML 2024
|
The problem of nearest neighbor condensing has enjoyed a long history of study, both in its theoretical and practical aspects. In this paper, we introduce the problem of weighted distance nearest neighbor condensing, where one assigns weights to each point of the condensed set, and then new points are labeled based on their weighted distance nearest neighbor in the condensed set. We study the theoretical properties of this new model, and show that it can produce dramatically better condensing than the standard nearest neighbor rule, yet is characterized by generalization bounds almost identical to the latter. We then suggest a condensing heuristic for our new problem. We demonstrate Bayes consistency for this heuristic, and also show promising empirical results.
|
https://proceedings.mlr.press/v235/gou24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/gou24a/gou24a.pdf
|
https://openreview.net/forum?id=XLlQb24X2o
|
Test-Time Degradation Adaptation for Open-Set Image Restoration
|
https://proceedings.mlr.press/v235/gou24a.html
|
Yuanbiao Gou, Haiyu Zhao, Boyun Li, Xinyan Xiao, Xi Peng
|
https://proceedings.mlr.press/v235/gou24a.html
|
ICML 2024
|
In contrast to close-set scenarios that restore images from a predefined set of degradations, open-set image restoration aims to handle the unknown degradations that were unforeseen during the pretraining phase, which is less-touched as far as we know. This work study this challenging problem and reveal its essence as unidentified distribution shifts between the test and training data. Recently, test-time adaptation has emerged as a fundamental method to address this inherent disparities. Inspired by it, we propose a test-time degradation adaptation framework for open-set image restoration, which consists of three components, i.e., i) a pre-trained and degradation-agnostic diffusion model for generating clean images, ii) a test-time degradation adapter adapts the unknown degradations based on the input image during the testing phase, and iii) the adapter-guided image restoration guides the model through the adapter to produce the corresponding clean image. Through experiments on multiple degradations, we show that our method achieves comparable even better performance than those task-specific methods. The code is available at https://github.com/XLearning-SCU/2024-ICML-TAO.
|
https://proceedings.mlr.press/v235/grau-moya24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/grau-moya24a/grau-moya24a.pdf
|
https://openreview.net/forum?id=B1ajnQyZgK
|
Learning Universal Predictors
|
https://proceedings.mlr.press/v235/grau-moya24a.html
|
Jordi Grau-Moya, Tim Genewein, Marcus Hutter, Laurent Orseau, Gregoire Deletang, Elliot Catt, Anian Ruoss, Li Kevin Wenliang, Christopher Mattern, Matthew Aitchison, Joel Veness
|
https://proceedings.mlr.press/v235/grau-moya24a.html
|
ICML 2024
|
Meta-learning has emerged as a powerful approach to train neural networks to learn new tasks quickly from limited data by pre-training them on a broad set of tasks. But, what are the limits of meta-learning? In this work, we explore the potential of amortizing the most powerful universal predictor, namely Solomonoff Induction (SI), into neural networks via leveraging (memory-based) meta-learning to its limits. We use Universal Turing Machines (UTMs) to generate training data used to expose networks to a broad range of patterns. We provide theoretical analysis of the UTM data generation processes and meta-training protocols. We conduct comprehensive experiments with neural architectures (e.g. LSTMs, Transformers) and algorithmic data generators of varying complexity and universality. Our results suggest that UTM data is a valuable resource for meta-learning, and that it can be used to train neural networks capable of learning universal prediction strategies.
|
https://proceedings.mlr.press/v235/gravina24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/gravina24a/gravina24a.pdf
|
https://openreview.net/forum?id=gVg8V9isul
|
Long Range Propagation on Continuous-Time Dynamic Graphs
|
https://proceedings.mlr.press/v235/gravina24a.html
|
Alessio Gravina, Giulio Lovisotto, Claudio Gallicchio, Davide Bacciu, Claas Grohnfeldt
|
https://proceedings.mlr.press/v235/gravina24a.html
|
ICML 2024
|
Learning Continuous-Time Dynamic Graphs (C-TDGs) requires accurately modeling spatio-temporal information on streams of irregularly sampled events. While many methods have been proposed recently, we find that most message passing-, recurrent- or self-attention-based methods perform poorly on long-range tasks. These tasks require correlating information that occurred "far" away from the current event, either spatially (higher-order node information) or along the time dimension (events occurred in the past). To address long-range dependencies, we introduce Continuous-Time Graph Anti-Symmetric Network (CTAN). Grounded within the ordinary differential equations framework, our method is designed for efficient propagation of information. In this paper, we show how CTAN’s (i) long-range modeling capabilities are substantiated by theoretical findings and how (ii) its empirical performance on synthetic long-range benchmarks and real-world benchmarks is superior to other methods. Our results motivate CTAN’s ability to propagate long-range information in C-TDGs as well as the inclusion of long-range tasks as part of temporal graph models evaluation.
|
https://proceedings.mlr.press/v235/graziani24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/graziani24a/graziani24a.pdf
|
https://openreview.net/forum?id=io1XSRtcO8
|
The Expressive Power of Path-Based Graph Neural Networks
|
https://proceedings.mlr.press/v235/graziani24a.html
|
Caterina Graziani, Tamara Drucks, Fabian Jogl, Monica Bianchini, Franco Scarselli, Thomas Gärtner
|
https://proceedings.mlr.press/v235/graziani24a.html
|
ICML 2024
|
We systematically investigate the expressive power of path-based graph neural networks. While it has been shown that path-based graph neural networks can achieve strong empirical results, an investigation into their expressive power is lacking. Therefore, we propose PATH-WL, a general class of color refinement algorithms based on paths and shortest path distance information. We show that PATH-WL is incomparable to a wide range of expressive graph neural networks, can count cycles, and achieves strong empirical results on the notoriously difficult family of strongly regular graphs. Our theoretical results indicate that PATH-WL forms a new hierarchy of highly expressive graph neural networks.
|
https://proceedings.mlr.press/v235/grazzi24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/grazzi24a/grazzi24a.pdf
|
https://openreview.net/forum?id=SlRcJvf1yd
|
Nonsmooth Implicit Differentiation: Deterministic and Stochastic Convergence Rates
|
https://proceedings.mlr.press/v235/grazzi24a.html
|
Riccardo Grazzi, Massimiliano Pontil, Saverio Salzo
|
https://proceedings.mlr.press/v235/grazzi24a.html
|
ICML 2024
|
We study the problem of efficiently computing the derivative of the fixed-point of a parametric nondifferentiable contraction map. This problem has wide applications in machine learning, including hyperparameter optimization, meta-learning and data poisoning attacks. We analyze two popular approaches: iterative differentiation (ITD) and approximate implicit differentiation (AID). A key challenge behind the nonsmooth setting is that the chain rule does not hold anymore. We build upon the work by Bolte et al. (2022), who prove linear convergence of nonsmooth ITD under a piecewise Lipschitz smooth assumption. In the deterministic case, we provide a linear rate for AID and an improved linear rate for ITD which closely match the ones for the smooth setting. We further introduce NSID, a new stochastic method to compute the implicit derivative when the contraction map is defined as the composition of an outer map and an inner map which is accessible only through a stochastic unbiased estimator. We establish rates for the convergence of NSID, encompassing the best available rates in the smooth setting. We also present illustrative experiments confirming our analysis.
|
https://proceedings.mlr.press/v235/grcic24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/grcic24a/grcic24a.pdf
|
https://openreview.net/forum?id=b9VfvegTEO
|
Fine-grained Classes and How to Find Them
|
https://proceedings.mlr.press/v235/grcic24a.html
|
Matej Grcic, Artyom Gadetsky, Maria Brbic
|
https://proceedings.mlr.press/v235/grcic24a.html
|
ICML 2024
|
In many practical applications, coarse-grained labels are readily available compared to fine-grained labels that reflect subtle differences between classes. However, existing methods cannot leverage coarse labels to infer fine-grained labels in an unsupervised manner. To bridge this gap, we propose FALCON, a method that discovers fine-grained classes from coarsely labeled data without any supervision at the fine-grained level. FALCON simultaneously infers unknown fine-grained classes and underlying relationships between coarse and fine-grained classes. Moreover, FALCON is a modular method that can effectively learn from multiple datasets labeled with different strategies. We evaluate FALCON on eight image classification tasks and a single-cell classification task. FALCON outperforms baselines by a large margin, achieving 22% improvement over the best baseline on the tieredImageNet dataset with over 600 fine-grained classes.
|
https://proceedings.mlr.press/v235/greenblatt24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/greenblatt24a/greenblatt24a.pdf
|
https://openreview.net/forum?id=KviM5k8pcP
|
AI Control: Improving Safety Despite Intentional Subversion
|
https://proceedings.mlr.press/v235/greenblatt24a.html
|
Ryan Greenblatt, Buck Shlegeris, Kshitij Sachan, Fabien Roger
|
https://proceedings.mlr.press/v235/greenblatt24a.html
|
ICML 2024
|
As large language models (LLMs) become more powerful and are deployed more autonomously, it will be increasingly important to prevent them from causing harmful outcomes. To do so, safety measures either aim at making LLMs try to avoid harmful outcomes or aim at preventing LLMs from causing harmful outcomes, even if they try to cause them. In this paper, we focus on this second layer of defense. We develop and evaluate pipelines of safety techniques (protocols) that try to ensure safety despite intentional subversion - an approach we call AI control. We investigate a setting in which we want to solve a sequence of programming problems without ever submitting subtly wrong code, using access to a powerful but untrusted model (in our case, GPT-4), access to a less powerful trusted model (in our case, GPT-3.5), and limited access to high-quality trusted labor. We investigate a range of protocols and red-team them by exploring strategies that the untrusted model could use to subvert them. We find that using the trusted model to edit untrusted-model code or using the untrusted model as a monitor substantially improves on simple baselines.
|
https://proceedings.mlr.press/v235/grenioux24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/grenioux24a/grenioux24a.pdf
|
https://openreview.net/forum?id=2Gr5wZR6uc
|
Stochastic Localization via Iterative Posterior Sampling
|
https://proceedings.mlr.press/v235/grenioux24a.html
|
Louis Grenioux, Maxence Noble, Marylou Gabrié, Alain Oliviero Durmus
|
https://proceedings.mlr.press/v235/grenioux24a.html
|
ICML 2024
|
Building upon score-based learning, new interest in stochastic localization techniques has recently emerged. In these models, one seeks to noise a sample from the data distribution through a stochastic process, called observation process, and progressively learns a denoiser associated to this dynamics. Apart from specific applications, the use of stochastic localization for the problem of sampling from an unnormalized target density has not been explored extensively. This work contributes to fill this gap. We consider a general stochastic localization framework and introduce an explicit class of observation processes, associated with flexible denoising schedules. We provide a complete methodology, Stochastic Localization via Iterative Posterior Sampling (SLIPS), to obtain approximate samples of these dynamics, and as a by-product, samples from the target distribution. Our scheme is based on a Markov chain Monte Carlo estimation of the denoiser and comes with detailed practical guidelines. We illustrate the benefits and applicability of SLIPS on several benchmarks of multi-modal distributions, including Gaussian mixtures in increasing dimensions, Bayesian logistic regression and a high-dimensional field system from statistical-mechanics.
|
https://proceedings.mlr.press/v235/greshler24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/greshler24a/greshler24a.pdf
|
https://openreview.net/forum?id=NS8z5FinYl
|
A Bayesian Approach to Online Planning
|
https://proceedings.mlr.press/v235/greshler24a.html
|
Nir Greshler, David Ben Eli, Carmel Rabinovitz, Gabi Guetta, Liran Gispan, Guy Zohar, Aviv Tamar
|
https://proceedings.mlr.press/v235/greshler24a.html
|
ICML 2024
|
The combination of Monte Carlo tree search and neural networks has revolutionized online planning. As neural network approximations are often imperfect, we ask whether uncertainty estimates about the network outputs could be used to improve planning. We develop a Bayesian planning approach that facilitates such uncertainty quantification, inspired by classical ideas from the meta-reasoning literature. We propose a Thompson sampling based algorithm for searching the tree of possible actions, for which we prove the first (to our knowledge) finite time Bayesian regret bound, and propose an efficient implementation for a restricted family of posterior distributions. In addition we propose a variant of the Bayes-UCB method applied to trees. Empirically, we demonstrate that on the ProcGen Maze and Leaper environments, when the uncertainty estimates are accurate but the neural network output is inaccurate, our Bayesian approach searches the tree much more effectively. In addition, we investigate whether popular uncertainty estimation methods are accurate enough to yield significant gains in planning.
|
https://proceedings.mlr.press/v235/greydanus24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/greydanus24a/greydanus24a.pdf
|
https://openreview.net/forum?id=n9pru4bJU9
|
Scaling Down Deep Learning with MNIST-1D
|
https://proceedings.mlr.press/v235/greydanus24a.html
|
Samuel James Greydanus, Dmitry Kobak
|
https://proceedings.mlr.press/v235/greydanus24a.html
|
ICML 2024
|
Although deep learning models have taken on commercial and political relevance, key aspects of their training and operation remain poorly understood. This has sparked interest in science of deep learning projects, many of which require large amounts of time, money, and electricity. But how much of this research really needs to occur at scale? In this paper, we introduce MNIST-1D: a minimalist, procedurally generated, low-memory, and low-compute alternative to classic deep learning benchmarks. Although the dimensionality of MNIST-1D is only 40 and its default training set size only 4000, MNIST-1D can be used to study inductive biases of different deep architectures, find lottery tickets, observe deep double descent, metalearn an activation function, and demonstrate guillotine regularization in self-supervised learning. All these experiments can be conducted on a GPU or often even on a CPU within minutes, allowing for fast prototyping, educational use cases, and cutting-edge research on a low budget.
|
https://proceedings.mlr.press/v235/grillotti24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/grillotti24a/grillotti24a.pdf
|
https://openreview.net/forum?id=ISG3l8nXrI
|
Quality-Diversity Actor-Critic: Learning High-Performing and Diverse Behaviors via Value and Successor Features Critics
|
https://proceedings.mlr.press/v235/grillotti24a.html
|
Luca Grillotti, Maxence Faldor, Borja G. León, Antoine Cully
|
https://proceedings.mlr.press/v235/grillotti24a.html
|
ICML 2024
|
A key aspect of intelligence is the ability to demonstrate a broad spectrum of behaviors for adapting to unexpected situations. Over the past decade, advancements in deep reinforcement learning have led to groundbreaking achievements to solve complex continuous control tasks. However, most approaches return only one solution specialized for a specific problem. We introduce Quality-Diversity Actor-Critic (QDAC), an off-policy actor-critic deep reinforcement learning algorithm that leverages a value function critic and a successor features critic to learn high-performing and diverse behaviors. In this framework, the actor optimizes an objective that seamlessly unifies both critics using constrained optimization to (1) maximize return, while (2) executing diverse skills. Compared with other Quality-Diversity methods, QDAC achieves significantly higher performance and more diverse behaviors on six challenging continuous control locomotion tasks. We also demonstrate that we can harness the learned skills to adapt better than other baselines to five perturbed environments. Finally, qualitative analyses showcase a range of remarkable behaviors: adaptive-intelligent-robotics.github.io/QDAC.
|
https://proceedings.mlr.press/v235/gruber24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/gruber24a/gruber24a.pdf
|
https://openreview.net/forum?id=QwgSOwynxD
|
A Bias-Variance-Covariance Decomposition of Kernel Scores for Generative Models
|
https://proceedings.mlr.press/v235/gruber24a.html
|
Sebastian Gregor Gruber, Florian Buettner
|
https://proceedings.mlr.press/v235/gruber24a.html
|
ICML 2024
|
Generative models, like large language models, are becoming increasingly relevant in our daily lives, yet a theoretical framework to assess their generalization behavior and uncertainty does not exist. Particularly, the problem of uncertainty estimation is commonly solved in an ad-hoc and task-dependent manner. For example, natural language approaches cannot be transferred to image generation. In this paper, we introduce the first bias-variance-covariance decomposition for kernel scores. This decomposition represents a theoretical framework from which we derive a kernel-based variance and entropy for uncertainty estimation. We propose unbiased and consistent estimators for each quantity which only require generated samples but not the underlying model itself. Based on the wide applicability of kernels, we demonstrate our framework via generalization and uncertainty experiments for image, audio, and language generation. Specifically, kernel entropy for uncertainty estimation is more predictive of performance on CoQA and TriviaQA question answering datasets than existing baselines and can also be applied to closed-source models.
|
https://proceedings.mlr.press/v235/gruber24b.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/gruber24b/gruber24b.pdf
|
https://openreview.net/forum?id=M407RM0z6h
|
Overcoming Saturation in Density Ratio Estimation by Iterated Regularization
|
https://proceedings.mlr.press/v235/gruber24b.html
|
Lukas Gruber, Markus Holzleitner, Johannes Lehner, Sepp Hochreiter, Werner Zellinger
|
https://proceedings.mlr.press/v235/gruber24b.html
|
ICML 2024
|
Estimating the ratio of two probability densities from finitely many samples, is a central task in machine learning and statistics. In this work, we show that a large class of kernel methods for density ratio estimation suffers from error saturation, which prevents algorithms from achieving fast error convergence rates on highly regular learning problems. To resolve saturation, we introduce iterated regularization in density ratio estimation to achieve fast error rates. Our methods outperform its non-iteratively regularized versions on benchmarks for density ratio estimation as well as on large-scale evaluations for importance-weighted ensembling of deep unsupervised domain adaptation models.
|
https://proceedings.mlr.press/v235/gu24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/gu24a/gu24a.pdf
|
https://openreview.net/forum?id=0THUA66D8Z
|
On the Calibration of Human Pose Estimation
|
https://proceedings.mlr.press/v235/gu24a.html
|
Kerui Gu, Rongyu Chen, Xuanlong Yu, Angela Yao
|
https://proceedings.mlr.press/v235/gu24a.html
|
ICML 2024
|
2D human pose estimation predicts keypoint locations and the corresponding confidence. Calibration-wise, the confidence should be aligned with the pose accuracy. Yet existing pose estimation methods tend to estimate confidence with heuristics such as the maximum value of heatmaps. This work shows, through theoretical analysis and empirical verification, a calibration gap in current pose estimation frameworks. Our derivations directly lead to closed-form adjustments in the confidence based on additionally inferred instance size and visibility. Given the black-box nature of deep neural networks, however, it is not possible to close the gap with only closed-form adjustments. We go one step further and propose a Calibrated ConfidenceNet (CCNet) to explicitly learn network-specific adjustments with a confidence prediction branch. The proposed CCNet, as a lightweight post-hoc addition, improves the calibration of standard off-the-shelf pose estimation frameworks.
|
https://proceedings.mlr.press/v235/gu24b.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/gu24b/gu24b.pdf
|
https://openreview.net/forum?id=fiugPLSXjK
|
EDISON: Enhanced Dictionary-Induced Tensorized Incomplete Multi-View Clustering with Gaussian Error Rank Minimization
|
https://proceedings.mlr.press/v235/gu24b.html
|
Zhibin Gu, Zhendong Li, Songhe Feng
|
https://proceedings.mlr.press/v235/gu24b.html
|
ICML 2024
|
This paper presents an efficient and scalable incomplete multi-view clustering method, referred to as Enhanced Dictionary-Induced tenSorized incomplete multi-view clustering with Gaussian errOr raNk minimization (EDISON). Specifically, EDISON employs an enhanced dictionary representation strategy as the foundation for inferring missing data and constructing anchor graphs, ensuring robustness to less-than-ideal data and maintaining high computational efficiency. Additionally, we introduce Gaussian error rank as a concise approximation of the true tensor rank, facilitating a comprehensive exploration of the diverse information encapsulated by various singular values in tensor data. Additionally, we integrate a hyper-anchor graph Laplacian manifold regularization into the tensor representation, allowing for the simultaneous utilization of inter-view high-order correlations and intra-view local correlations. Extensive experiments demonstrate the superiority of the EDISON model in both effectiveness and efficiency compared to SOTA methods.
|
https://proceedings.mlr.press/v235/gu24c.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/gu24c/gu24c.pdf
|
https://openreview.net/forum?id=Ffpg52swvg
|
CRUXEval: A Benchmark for Code Reasoning, Understanding and Execution
|
https://proceedings.mlr.press/v235/gu24c.html
|
Alex Gu, Baptiste Roziere, Hugh James Leather, Armando Solar-Lezama, Gabriel Synnaeve, Sida Wang
|
https://proceedings.mlr.press/v235/gu24c.html
|
ICML 2024
|
We present Code Reasoning, Understanding, and eXecution Evaluation, a benchmark consisting of 800 Python functions (3-13 lines). Each function comes with an input-output pair, leading to two natural tasks: input prediction and output prediction. First, we propose a general recipe for generating our execution benchmark by sampling from a model, which can be used for more challenging versions of the benchmark if needed. Second, we evaluate twenty code models on our benchmark and discover that many recent high-scoring models on HumanEval show no improvements on our benchmark. Third, we show that simple CoT and fine-tuning schemes can improve performance on our benchmark but remain far from solving it. The best setup, GPT-4 with chain of thought (CoT), achieves a pass@1 of 75% and 81% on input and output prediction, respectively. In contrast, Code Llama 34B achieves a pass@1 of 50% and 46% on input and output prediction. When it comes to reasoning about code, GPT-4 has a huge edge over other models but still fails consistently on some surprisingly simple Python programs.
|
https://proceedings.mlr.press/v235/gu24d.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/gu24d/gu24d.pdf
|
https://openreview.net/forum?id=jw2f9v59g0
|
Data-free Distillation of Diffusion Models with Bootstrapping
|
https://proceedings.mlr.press/v235/gu24d.html
|
Jiatao Gu, Chen Wang, Shuangfei Zhai, Yizhe Zhang, Lingjie Liu, Joshua M. Susskind
|
https://proceedings.mlr.press/v235/gu24d.html
|
ICML 2024
|
Diffusion models have demonstrated great potential for generating diverse images. However, their performance often suffers from slow generation due to iterative denoising. Knowledge distillation has been recently proposed as a remedy which can reduce the number of inference steps to one or a few, without significant quality degradation. However, existing distillation methods either require significant amounts of offline computation for generating synthetic training data from the teacher model, or need to perform expensive online learning with the help of real data. In this work, we present a novel technique called BOOT, that overcomes these limitations with an efficient data-free distillation algorithm. The core idea is to learn a time-conditioned model that predicts the output of a pre-trained diffusion model teacher given any time-step. Such a model can be efficiently trained based on bootstrapping from two consecutive sampled steps. Furthermore, our method can be easily adapted to large-scale text-to-image diffusion models, which are challenging for previous methods given the fact that the training sets are often large and difficult to access. We demonstrate the effectiveness of our approach on several benchmark datasets in the DDIM setting, achieving comparable generation quality while being orders of magnitude faster than the diffusion teacher. The text-to-image results show that the proposed approach is able to handle highly complex distributions, shedding light on more efficient generative modeling.
|
https://proceedings.mlr.press/v235/gu24e.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/gu24e/gu24e.pdf
|
https://openreview.net/forum?id=DYMj03Gbri
|
Agent Smith: A Single Image Can Jailbreak One Million Multimodal LLM Agents Exponentially Fast
|
https://proceedings.mlr.press/v235/gu24e.html
|
Xiangming Gu, Xiaosen Zheng, Tianyu Pang, Chao Du, Qian Liu, Ye Wang, Jing Jiang, Min Lin
|
https://proceedings.mlr.press/v235/gu24e.html
|
ICML 2024
|
A multimodal large language model (MLLM) agent can receive instructions, capture images, retrieve histories from memory, and decide which tools to use. Nonetheless, red-teaming efforts have revealed that adversarial images/prompts can jailbreak an MLLM and cause unaligned behaviors. In this work, we report an even more severe safety issue in multi-agent environments, referred to as infectious jailbreak. It entails the adversary simply jailbreaking a single agent, and without any further intervention from the adversary, (almost) all agents will become infected exponentially fast and exhibit harmful behaviors. To validate the feasibility of infectious jailbreak, we simulate multi-agent environments containing up to one million LLaVA-1.5 agents, and employ randomized pair-wise chat as a proof-of-concept instantiation for multi-agent interaction. Our results show that feeding an (infectious) adversarial image into the memory of any randomly chosen agent is sufficient to achieve infectious jailbreak. Finally, we derive a simple principle for determining whether a defense mechanism can provably restrain the spread of infectious jailbreak, but how to design a practical defense that meets this principle remains an open question to investigate.
|
https://proceedings.mlr.press/v235/guerdan24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/guerdan24a/guerdan24a.pdf
|
https://openreview.net/forum?id=HrzQZXzrN2
|
Predictive Performance Comparison of Decision Policies Under Confounding
|
https://proceedings.mlr.press/v235/guerdan24a.html
|
Luke Guerdan, Amanda Lee Coston, Ken Holstein, Steven Wu
|
https://proceedings.mlr.press/v235/guerdan24a.html
|
ICML 2024
|
Predictive models are often introduced to decision-making tasks under the rationale that they improve performance over an existing decision-making policy. However, it is challenging to compare predictive performance against an existing decision-making policy that is generally under-specified and dependent on unobservable factors. These sources of uncertainty are often addressed in practice by making strong assumptions about the data-generating mechanism. In this work, we propose a method to compare the predictive performance of decision policies under a variety of modern identification approaches from the causal inference and off-policy evaluation literatures (e.g., instrumental variable, marginal sensitivity model, proximal variable). Key to our method is the insight that there are regions of uncertainty that we can safely ignore in the policy comparison. We develop a practical approach for finite-sample estimation of regret intervals under no assumptions on the parametric form of the status quo policy. We verify our framework theoretically and via synthetic data experiments. We conclude with a real-world application using our framework to support a pre-deployment evaluation of a proposed modification to a healthcare enrollment policy.
|
https://proceedings.mlr.press/v235/guha24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/guha24a/guha24a.pdf
|
https://openreview.net/forum?id=Ld255Mbx9F
|
On the Diminishing Returns of Width for Continual Learning
|
https://proceedings.mlr.press/v235/guha24a.html
|
Etash Kumar Guha, Vihan Lakshman
|
https://proceedings.mlr.press/v235/guha24a.html
|
ICML 2024
|
While deep neural networks have demonstrated groundbreaking performance in various settings, these models often suffer from catastrophic forgetting when trained on new tasks in sequence. Several works have empirically demonstrated that increasing the width of a neural network leads to a decrease in catastrophic forgetting but have yet to characterize the exact relationship between width and continual learning. We design one of the first frameworks to analyze Continual Learning Theory and prove that width is directly related to forgetting in Feed-Forward Networks (FFN), demonstrating that the diminishing returns of increasing widths to reduce forgetting. We empirically verify our claims at widths hitherto unexplored in prior studies where the diminishing returns are clearly observed as predicted by our theory.
|
https://proceedings.mlr.press/v235/gui24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/gui24a/gui24a.pdf
|
https://openreview.net/forum?id=7uwLvFvpis
|
Vector Quantization Pretraining for EEG Time Series with Random Projection and Phase Alignment
|
https://proceedings.mlr.press/v235/gui24a.html
|
Haokun Gui, Xiucheng Li, Xinyang Chen
|
https://proceedings.mlr.press/v235/gui24a.html
|
ICML 2024
|
In this paper, we propose a BERT-style self-supervised learning model, VQ-MTM (Vector Quantization Masked Time-Series Modeling), for the EEG time series data analysis. At its core, VQ-MTM comprises a theoretically grounded random-projection quantization module and a phase-aligning module guided by the Time-Phase-Shift Equivariance of Fourier Transform, the two modules can generate well-defined semantic units (akin to words in natural language) for the corrupted and periodic time series, thus offering robust and consistent learning signals for the EEG self-supervised learning. VQ-MTM also owns low model complexity and can easily adapt to large-scale datasets. We conduct experiments on five real-world datasets including two large-scale datasets to verify the efficacy of our proposed model, the experiment results show that VQ-MTM is able to consistently surpass the existing methods by large margins on both seizure detection and classification tasks. Our code is available at https://github.com/HaokunGUI/VQ_MTM.
|
https://proceedings.mlr.press/v235/guille-escuret24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/guille-escuret24a/guille-escuret24a.pdf
|
https://openreview.net/forum?id=60vx5AfM3C
|
No Wrong Turns: The Simple Geometry Of Neural Networks Optimization Paths
|
https://proceedings.mlr.press/v235/guille-escuret24a.html
|
Charles Guille-Escuret, Hiroki Naganuma, Kilian Fatras, Ioannis Mitliagkas
|
https://proceedings.mlr.press/v235/guille-escuret24a.html
|
ICML 2024
|
Understanding the optimization dynamics of neural networks is necessary for closing the gap between theory and practice. Stochastic first-order optimization algorithms are known to efficiently locate favorable minima in deep neural networks. This efficiency, however, contrasts with the non-convex and seemingly complex structure of neural loss landscapes. In this study, we delve into the fundamental geometric properties of sampled gradients along optimization paths. We focus on two key quantities, the restricted secant inequality and error bound, as well as their ratio γ, which hold high significance for first-order optimization. Our analysis reveals that these quantities exhibit predictable, consistent behavior throughout training, despite the stochasticity induced by sampling minibatches. Our findings suggest that not only do optimization trajectories never encounter significant obstacles, but they also maintain stable dynamics during the majority of training. These observed properties are sufficiently expressive to theoretically guarantee linear convergence and prescribe learning rate schedules mirroring empirical practices. We conduct our experiments on image classification, semantic segmentation and language modeling across different batch sizes, network architectures, datasets, optimizers, and initialization seeds. We discuss the impact of each factor. Our work provides novel insights into the properties of neural network loss functions, and opens the door to theoretical frameworks more relevant to prevalent practice.
|
https://proceedings.mlr.press/v235/guinet24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/guinet24a/guinet24a.pdf
|
https://openreview.net/forum?id=4jqOV6NlUz
|
Automated Evaluation of Retrieval-Augmented Language Models with Task-Specific Exam Generation
|
https://proceedings.mlr.press/v235/guinet24a.html
|
Gauthier Guinet, Behrooz Omidvar-Tehrani, Anoop Deoras, Laurent Callot
|
https://proceedings.mlr.press/v235/guinet24a.html
|
ICML 2024
|
We propose a new method to measure the task-specific accuracy of Retrieval-Augmented Large Language Models (RAG). Evaluation is performed by scoring the RAG on an automatically-generated synthetic exam composed of multiple choice questions based on the corpus of documents associated with the task. Our method is an automated, cost-efficient, interpretable, and robust strategy to select the optimal components for a RAG system. We leverage Item Response Theory (IRT) to estimate the quality of an exam and its informativeness on task-specific accuracy. IRT also provides a natural way to iteratively improve the exam by eliminating the exam questions that are not sufficiently informative about a model’s ability. We demonstrate our approach on four new open-ended Question-Answering tasks based on Arxiv abstracts, StackExchange questions, AWS DevOps troubleshooting guides, and SEC filings. In addition, our experiments reveal more general insights into factors impacting RAG performance like size, retrieval mechanism, prompting and fine-tuning. Most notably, our findings show that choosing the right retrieval algorithms often leads to bigger performance gains than simply using a larger language model.
|
https://proceedings.mlr.press/v235/guo24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/guo24a/guo24a.pdf
|
https://openreview.net/forum?id=aQl4xiwVBc
|
SLAB: Efficient Transformers with Simplified Linear Attention and Progressive Re-parameterized Batch Normalization
|
https://proceedings.mlr.press/v235/guo24a.html
|
Jialong Guo, Xinghao Chen, Yehui Tang, Yunhe Wang
|
https://proceedings.mlr.press/v235/guo24a.html
|
ICML 2024
|
Transformers have become foundational architectures for both natural language and computer vision tasks. However, the high computational cost makes it quite challenging to deploy on resource-constraint devices. This paper investigates the computational bottleneck modules of efficient transformer, i.e., normalization layers and attention modules. LayerNorm is commonly used in transformer architectures but is not computational friendly due to statistic calculation during inference. However, replacing LayerNorm with more efficient BatchNorm in transformer often leads to inferior performance and collapse in training. To address this problem, we propose a novel method named PRepBN to progressively replace LayerNorm with re-parameterized BatchNorm in training. Moreover, we propose a simplified linear attention (SLA) module that is simple yet effective to achieve strong performance. Extensive experiments on image classification as well as object detection demonstrate the effectiveness of our proposed method. For example, our SLAB-Swin obtains $83.6%$ top-1 accuracy on ImageNet-1K with $16.2$ms latency, which is $2.4$ms less than that of Flatten-Swin with $0.1%$ higher accuracy. We also evaluated our method for language modeling task and obtain comparable performance and lower latency. Codes are publicly available at https://github.com/xinghaochen/SLAB and https://github.com/mindspore-lab/models/tree/master/research/huawei-noah/SLAB.
|
https://proceedings.mlr.press/v235/guo24b.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/guo24b/guo24b.pdf
|
https://openreview.net/forum?id=LfJgeBNCFI
|
DS-Agent: Automated Data Science by Empowering Large Language Models with Case-Based Reasoning
|
https://proceedings.mlr.press/v235/guo24b.html
|
Siyuan Guo, Cheng Deng, Ying Wen, Hechang Chen, Yi Chang, Jun Wang
|
https://proceedings.mlr.press/v235/guo24b.html
|
ICML 2024
|
In this work, we investigate the potential of large language models (LLMs) based agents to automate data science tasks, with the goal of comprehending task requirements, then building and training the best-fit machine learning models. Despite their widespread success, existing LLM agents are hindered by generating unreasonable experiment plans within this scenario. To this end, we present DS-Agent, a novel automatic framework that harnesses LLM agent and case-based reasoning (CBR). In the development stage, DS-Agent follows the CBR framework to structure an automatic iteration pipeline, which can flexibly capitalize on the expert knowledge from Kaggle, and facilitate consistent performance improvement through the feedback mechanism. Moreover, DS-Agent implements a low-resource deployment stage with a simplified CBR paradigm to adapt past successful solutions from the development stage for direct code generation, significantly reducing the demand on foundational capabilities of LLMs. Empirically, DS-Agent with GPT-4 achieves 100% success rate in the development stage, while attaining 36% improvement on average one pass rate across alternative LLMs in the deployment stage. In both stages, DS-Agent achieves the best rank in performance, costing $1.60 and \$0.13 per run with GPT-4, respectively. Our data and code are open-sourced at https://github.com/guosyjlu/DS-Agent.
|
https://proceedings.mlr.press/v235/guo24c.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/guo24c/guo24c.pdf
|
https://openreview.net/forum?id=LJ34pX1U5g
|
Collaborative Heterogeneous Causal Inference Beyond Meta-analysis
|
https://proceedings.mlr.press/v235/guo24c.html
|
Tianyu Guo, Sai Praneeth Karimireddy, Michael Jordan
|
https://proceedings.mlr.press/v235/guo24c.html
|
ICML 2024
|
Collaboration between different data centers is often challenged by heterogeneity across sites. To account for the heterogeneity, the state-of-the-art method is to re-weight the covariate distributions in each site to match the distribution of the target population. Nevertheless, this method still relies on the concept of traditional meta-analysis after adjusting for the distribution shift. This work proposes a collaborative inverse propensity score weighting estimator for causal inference with heterogeneous data. Instead of adjusting the distribution shift separately, we use weighted propensity score models to collaboratively adjust for the distribution shift. Our method shows significant improvements over the methods based on meta-analysis when heterogeneity increases. By incorporating outcome regression models, we prove the asymptotic normality when the covariates have dimension $d<8$. Our methods preserve privacy at individual sites by implementing federated learning protocols.
|
https://proceedings.mlr.press/v235/guo24d.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/guo24d/guo24d.pdf
|
https://openreview.net/forum?id=qDAAMmGsGw
|
ACM-MILP: Adaptive Constraint Modification via Grouping and Selection for Hardness-Preserving MILP Instance Generation
|
https://proceedings.mlr.press/v235/guo24d.html
|
Ziao Guo, Yang Li, Chang Liu, Wenli Ouyang, Junchi Yan
|
https://proceedings.mlr.press/v235/guo24d.html
|
ICML 2024
|
Data plays a pivotal role in the development of both classic and learning-based methods for Mixed-Integer Linear Programming (MILP). However, the scarcity of data in real-world applications underscores the necessity for MILP instance generation methods. Currently, these methods primarily rely on iterating random single-constraint modifications, disregarding the underlying problem structure with constraint interrelations, thereby leading to compromised quality and solvability. In this paper, we propose ACM-MILP, a framework for MILP instance generation, to achieve adaptive constraint modification and constraint interrelation modeling. It employs an adaptive constraint selection mechanism based on probability estimation within the latent space to preserve instance characteristics. Meanwhile, it detects and groups strongly related constraints through community detection, enabling collective modifications that account for constraint dependencies. Experimental results show significant improvements in problem-solving hardness similarity under our framework. Additionally, in the downstream task, we showcase the efficacy of our generated instances for hyperparameter tuning. Source code is available: https://github.com/Thinklab-SJTU/ACM-MILP.
|
https://proceedings.mlr.press/v235/guo24e.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/guo24e/guo24e.pdf
|
https://openreview.net/forum?id=aPVwOAr1aW
|
On the Embedding Collapse when Scaling up Recommendation Models
|
https://proceedings.mlr.press/v235/guo24e.html
|
Xingzhuo Guo, Junwei Pan, Ximei Wang, Baixu Chen, Jie Jiang, Mingsheng Long
|
https://proceedings.mlr.press/v235/guo24e.html
|
ICML 2024
|
Recent advances in foundation models have led to a promising trend of developing large recommendation models to leverage vast amounts of available data. Still, mainstream models remain embarrassingly small in size and naive enlarging does not lead to sufficient performance gain, suggesting a deficiency in the model scalability. In this paper, we identify the embedding collapse phenomenon as the inhibition of scalability, wherein the embedding matrix tends to occupy a low-dimensional subspace. Through empirical and theoretical analysis, we demonstrate a two-sided effect of feature interaction specific to recommendation models. On the one hand, interacting with collapsed embeddings restricts embedding learning and exacerbates the collapse issue. On the other hand, interaction is crucial in mitigating the fitting of spurious features as a scalability guarantee. Based on our analysis, we propose a simple yet effective multi-embedding design incorporating embedding-set-specific interaction modules to learn embedding sets with large diversity and thus reduce collapse. Extensive experiments demonstrate that this proposed design provides consistent scalability and effective collapse mitigation for various recommendation models. Code is available at this repository: https://github.com/thuml/Multi-Embedding.
|
https://proceedings.mlr.press/v235/guo24f.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/guo24f/guo24f.pdf
|
https://openreview.net/forum?id=kc4dZYJlJG
|
FedRC: Tackling Diverse Distribution Shifts Challenge in Federated Learning by Robust Clustering
|
https://proceedings.mlr.press/v235/guo24f.html
|
Yongxin Guo, Xiaoying Tang, Tao Lin
|
https://proceedings.mlr.press/v235/guo24f.html
|
ICML 2024
|
Federated Learning (FL) is a machine learning paradigm that safeguards privacy by retaining client data on edge devices. However, optimizing FL in practice can be challenging due to the diverse and heterogeneous nature of the learning system. Though recent research has focused on improving the optimization of FL when distribution shifts occur among clients, ensuring global performance when multiple types of distribution shifts occur simultaneously among clients—such as feature distribution shift, label distribution shift, and concept shift—remain under-explored. In this paper, we identify the learning challenges posed by the simultaneous occurrence of diverse distribution shifts and propose a clustering principle to overcome these challenges. Through our research, we find that existing methods fail to address the clustering principle. Therefore, we propose a novel clustering algorithm framework, dubbed as FedRC, which adheres to our proposed clustering principle by incorporating a bi-level optimization problem and a novel objective function. Extensive experiments demonstrate that FedRC significantly outperforms other SOTA cluster-based FL methods. Our code will be publicly available.
|
https://proceedings.mlr.press/v235/guo24g.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/guo24g/guo24g.pdf
|
https://openreview.net/forum?id=sCGRhnuMUJ
|
Compressing Large Language Models by Joint Sparsification and Quantization
|
https://proceedings.mlr.press/v235/guo24g.html
|
Jinyang Guo, Jianyu Wu, Zining Wang, Jiaheng Liu, Ge Yang, Yifu Ding, Ruihao Gong, Haotong Qin, Xianglong Liu
|
https://proceedings.mlr.press/v235/guo24g.html
|
ICML 2024
|
In this paper, we introduce a novel model compression technique named Joint Sparsification and Quantization (JSQ), explicitly tailored for large language models (LLMs). Traditional methods employ either sparsification or quantization individually to compress LLMs, leading to performance degradation at high compression ratios. In contrast, our JSQ approach integrates sparsification and quantization cohesively. As sparsification tend to preserve outliers that is harmful to quantization, we introduce a novel sparsity metric to serves as a bridge between the sparsification and quantization. Moreover, it is proven outliers in LLMs have significant impact but harmful to compression. Current solutions are highly coupled with quantization process, which is not helpful to sparsification. To this end, we also introduce a search-based activation editor to automatically eliminate relatively useless outliers. Comprehensive experiments across various datasets and architectures affirm the efficacy of our JSQ framework. Notably, our JSQ achieves 7.96$\times$ computation reduction without crashing for the representative model LLaMA. This accomplishment stands in stark contrast to the limitations of most state-of-the-art LLM compression methods, which typically fail under such extreme compression ratios. Our code is released at https://github.com/uanu2002/JSQ.
|
https://proceedings.mlr.press/v235/guo24h.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/guo24h/guo24h.pdf
|
https://openreview.net/forum?id=O1hmwi51pp
|
Automated Loss function Search for Class-imbalanced Node Classification
|
https://proceedings.mlr.press/v235/guo24h.html
|
Xinyu Guo, Kai Wu, Xiaoyu Zhang, Jing Liu
|
https://proceedings.mlr.press/v235/guo24h.html
|
ICML 2024
|
Class-imbalanced node classification tasks are prevalent in real-world scenarios. Due to the uneven distribution of nodes across different classes, learning high-quality node representations remains a challenging endeavor. The engineering of loss functions has shown promising potential in addressing this issue. It involves the meticulous design of loss functions, utilizing information about the quantities of nodes in different categories and the network’s topology to learn unbiased node representations. However, the design of these loss functions heavily relies on human expert knowledge and exhibits limited adaptability to specific target tasks. In this paper, we introduce a high-performance, flexible, and generalizable automated loss function search framework to tackle this challenge. Across 15 combinations of graph neural networks and datasets, our framework achieves a significant improvement in performance compared to state-of-the-art methods. Additionally, we observe that homophily in graph-structured data significantly contributes to the transferability of the proposed framework.
|
https://proceedings.mlr.press/v235/guo24i.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/guo24i/guo24i.pdf
|
https://openreview.net/forum?id=yUxdk32TU6
|
COLD-Attack: Jailbreaking LLMs with Stealthiness and Controllability
|
https://proceedings.mlr.press/v235/guo24i.html
|
Xingang Guo, Fangxu Yu, Huan Zhang, Lianhui Qin, Bin Hu
|
https://proceedings.mlr.press/v235/guo24i.html
|
ICML 2024
|
Jailbreaks on large language models (LLMs) have recently received increasing attention. For a comprehensive assessment of LLM safety, it is essential to consider jailbreaks with diverse attributes, such as contextual coherence and sentiment/stylistic variations, and hence it is beneficial to study controllable jailbreaking, i.e. how to enforce control on LLM attacks. In this paper, we formally formulate the controllable attack generation problem, and build a novel connection between this problem and controllable text generation, a well-explored topic of natural language processing. Based on this connection, we adapt the Energy-based Constrained Decoding with Langevin Dynamics (COLD), a state-of-the-art, highly efficient algorithm in controllable text generation, and introduce the COLD-Attack framework which unifies and automates the search of adversarial LLM attacks under a variety of control requirements such as fluency, stealthiness, sentiment, and left-right-coherence. The controllability enabled by COLD-Attack leads to diverse new jailbreak scenarios which not only cover the standard setting of generating fluent (suffix) attack with continuation constraint, but also allow us to address new controllable attack settings such as revising a user query adversarially with paraphrasing constraint, and inserting stealthy attacks in context with position constraint. Our extensive experiments on various LLMs (Llama-2, Mistral, Vicuna, Guanaco, GPT-3.5, and GPT-4) show COLD-Attack’s broad applicability, strong controllability, high success rate, and attack transferability. Our code is available at https://github.com/Yu-Fangxu/COLD-Attack.
|
https://proceedings.mlr.press/v235/guo24j.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/guo24j/guo24j.pdf
|
https://openreview.net/forum?id=7bg10Jj3bG
|
Temporal Logic Specification-Conditioned Decision Transformer for Offline Safe Reinforcement Learning
|
https://proceedings.mlr.press/v235/guo24j.html
|
Zijian Guo, Weichao Zhou, Wenchao Li
|
https://proceedings.mlr.press/v235/guo24j.html
|
ICML 2024
|
Offline safe reinforcement learning (RL) aims to train a constraint satisfaction policy from a fixed dataset. Current state-of-the-art approaches are based on supervised learning with a conditioned policy. However, these approaches fall short in real-world applications that involve complex tasks with rich temporal and logical structures. In this paper, we propose temporal logic Specification-conditioned Decision Transformer (SDT), a novel framework that harnesses the expressive power of signal temporal logic (STL) to specify complex temporal rules that an agent should follow and the sequential modeling capability of Decision Transformer (DT). Empirical evaluations on the DSRL benchmarks demonstrate the better capacity of SDT in learning safe and high-reward policies compared with existing approaches. In addition, SDT shows good alignment with respect to different desired degrees of satisfaction of the STL specification that it is conditioned on.
|
https://proceedings.mlr.press/v235/gupta24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/gupta24a/gupta24a.pdf
|
https://openreview.net/forum?id=tp6ruPIfIV
|
Diffusion Posterior Sampling is Computationally Intractable
|
https://proceedings.mlr.press/v235/gupta24a.html
|
Shivam Gupta, Ajil Jalal, Aditya Parulekar, Eric Price, Zhiyang Xun
|
https://proceedings.mlr.press/v235/gupta24a.html
|
ICML 2024
|
Diffusion models are a remarkably effective way of learning and sampling from a distribution $p(x)$. In posterior sampling, one is also given a measurement model $p(y \mid x)$ and a measurement $y$, and would like to sample from $p(x \mid y)$. Posterior sampling is useful for tasks such as inpainting, super-resolution, and MRI reconstruction, so a number of recent works have given algorithms to heuristically approximate it; but none are known to converge to the correct distribution in polynomial time. In this paper we show that posterior sampling is computationally intractable: under the most basic assumption in cryptography—that one-way functions exist—there are instances for which every algorithm takes superpolynomial time, even though unconditional sampling is provably fast. We also show that the exponential-time rejection sampling algorithm is essentially optimal under the stronger plausible assumption that there are one-way functions that take exponential time to invert.
|
https://proceedings.mlr.press/v235/gupta24b.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/gupta24b/gupta24b.pdf
|
https://openreview.net/forum?id=wDDprThYeT
|
xT: Nested Tokenization for Larger Context in Large Images
|
https://proceedings.mlr.press/v235/gupta24b.html
|
Ritwik Gupta, Shufan Li, Tyler Zhu, Jitendra Malik, Trevor Darrell, Karttikeya Mangalam
|
https://proceedings.mlr.press/v235/gupta24b.html
|
ICML 2024
|
Modern computer vision pipelines handle large images in one of two sub-optimal ways: down-sampling or cropping. These two methods incur significant losses in the amount of information and context present in an image. There are many downstream applications in which global context matters as much as high frequency details, such as in real-world satellite imagery; in such cases researchers have to make the uncomfortable choice of which information to discard. We introduce xT, a simple framework for vision transformers which effectively aggregates global context with local details and can model large images end-to-end on contemporary GPUs. We select a set of benchmark datasets across classic vision tasks which accurately reflect a vision model’s ability to understand truly large images and incorporate fine details over large scales and assess our method’s improvement on them. xT is a streaming, two-stage architecture that adapts existing vision backbones and long sequence language models to effectively model large images without quadratic memory growth. We are able to increase accuracy by up to 8.6% on challenging classification tasks and F1 score by 11.6 on context-dependent segmentation on images as large as 29,000 x 29,000 pixels.
|
https://proceedings.mlr.press/v235/gupta24c.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/gupta24c/gupta24c.pdf
|
https://openreview.net/forum?id=WCVC5wGZyz
|
GistScore: Learning Better Representations for In-Context Example Selection with Gist Bottlenecks
|
https://proceedings.mlr.press/v235/gupta24c.html
|
Shivanshu Gupta, Clemens Rosenbaum, Ethan R. Elenberg
|
https://proceedings.mlr.press/v235/gupta24c.html
|
ICML 2024
|
In-Context Learning (ICL) is the ability of Large Language Models (LLMs) to perform new tasks when conditioned on prompts comprising a few task examples. However, ICL performance can be critically sensitive to the choice of examples. To dynamically select the best examples for every test input, we propose Example Gisting, a novel approach for training example encoders through supervised finetuning with an attention bottleneck between the inputs and outputs. These gist models form the basis for GistScore, a novel metric for scoring and selecting informative examples. Further, we experiment with two variations: (1) finetuning gist models for each dataset and (2) multi-task training a single model on a large collection of datasets. The latter can be used for new tasks out-of-the-box, enabling a training-free ICL pipeline. Evaluations with 21 datasets spanning 9 tasks and 8 diverse LLMs show that our fine-tuned models get state-of-the-art ICL performance with over 20% absolute gain over off-the-shelf retrievers and 5% over the best prior methods. Further, our multi-task model generalizes well to new tasks, datasets, and prompt templates. Selection using this model matches or outperforms prior methods while being three orders of magnitude faster than the strongest training-free baseline.
|
https://proceedings.mlr.press/v235/gushchin24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/gushchin24a/gushchin24a.pdf
|
https://openreview.net/forum?id=EWJn6hfZ4J
|
Light and Optimal Schrödinger Bridge Matching
|
https://proceedings.mlr.press/v235/gushchin24a.html
|
Nikita Gushchin, Sergei Kholkin, Evgeny Burnaev, Alexander Korotin
|
https://proceedings.mlr.press/v235/gushchin24a.html
|
ICML 2024
|
Schrödinger Bridges (SB) have recently gained the attention of the ML community as a promising extension of classic diffusion models which is also interconnected to the Entropic Optimal Transport (EOT). Recent solvers for SB exploit the pervasive bridge matching procedures. Such procedures aim to recover a stochastic process transporting the mass between distributions given only a transport plan between them. In particular, given the EOT plan, these procedures can be adapted to solve SB. This fact is heavily exploited by recent works giving rives to matching-based SB solvers. The cornerstone here is recovering the EOT plan: recent works either use heuristical approximations (e.g., the minibatch OT) or establish iterative matching procedures which by the design accumulate the error during the training. We address these limitations and propose a novel procedure to learn SB which we call the optimal Schrödinger bridge matching. It exploits the optimal parameterization of the diffusion process and provably recovers the SB process (a) with a single bridge matching step and (b) with arbitrary transport plan as the input. Furthermore, we show that the optimal bridge matching objective coincides with the recently discovered energy-based modeling (EBM) objectives to learn EOT/SB. Inspired by this observation, we develop a light solver (which we call LightSB-M) to implement optimal matching in practice using the Gaussian mixture parameterization of the adjusted Schrödinger potential. We experimentally showcase the performance of our solver in a range of practical tasks.
|
https://proceedings.mlr.press/v235/h-zargarbashi24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/h-zargarbashi24a/h-zargarbashi24a.pdf
|
https://openreview.net/forum?id=MrNq6rbcUi
|
Robust Yet Efficient Conformal Prediction Sets
|
https://proceedings.mlr.press/v235/h-zargarbashi24a.html
|
Soroush H. Zargarbashi, Mohammad Sadegh Akhondzadeh, Aleksandar Bojchevski
|
https://proceedings.mlr.press/v235/h-zargarbashi24a.html
|
ICML 2024
|
Conformal prediction (CP) can convert any model’s output into prediction sets guaranteed to include the true label with any user-specified probability. However, same as the model itself, CP is vulnerable to adversarial test examples (evasion) and perturbed calibration data (poisoning). We derive provably robust sets by bounding the worst-case change in conformity scores. Our tighter bounds lead to more efficient sets. We cover both continuous and discrete (sparse) data and our guarantees work both for evasion and poisoning attacks (on both features and labels).
|
https://proceedings.mlr.press/v235/haddadan24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/haddadan24a/haddadan24a.pdf
|
https://openreview.net/forum?id=Sz9mAYuqlE
|
Optimally Improving Cooperative Learning in a Social Setting
|
https://proceedings.mlr.press/v235/haddadan24a.html
|
Shahrzad Haddadan, Cheng Xin, Jie Gao
|
https://proceedings.mlr.press/v235/haddadan24a.html
|
ICML 2024
|
We consider a cooperative learning scenario where a collection of networked agents with individually owned classifiers dynamically update their predictions, for the same classification task, through communication or observations of each other’s predictions. Clearly if highly influential vertices use erroneous classifiers, there will be a negative effect on the accuracy of all the agents in the network. We ask the following question: how can we optimally fix the prediction of a few classifiers so as maximize the overall accuracy in the entire network. To this end we consider an aggregate and an egalitarian objective function. We show a polynomial time algorithm for optimizing the aggregate objective function, and show that optimizing the egalitarian objective function is NP-hard. Furthermore, we develop approximation algorithms for the egalitarian improvement. The performance of all of our algorithms are guaranteed by mathematical analysis and backed by experiments on synthetic and real data.
|
https://proceedings.mlr.press/v235/hagnberger24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/hagnberger24a/hagnberger24a.pdf
|
https://openreview.net/forum?id=sF9epWkNUG
|
Vectorized Conditional Neural Fields: A Framework for Solving Time-dependent Parametric Partial Differential Equations
|
https://proceedings.mlr.press/v235/hagnberger24a.html
|
Jan Hagnberger, Marimuthu Kalimuthu, Daniel Musekamp, Mathias Niepert
|
https://proceedings.mlr.press/v235/hagnberger24a.html
|
ICML 2024
|
Transformer models are increasingly used for solving Partial Differential Equations (PDEs). Several adaptations have been proposed, all of which suffer from the typical problems of Transformers, such as quadratic memory and time complexity. Furthermore, all prevalent architectures for PDE solving lack at least one of several desirable properties of an ideal surrogate model, such as (i) generalization to PDE parameters not seen during training, (ii) spatial and temporal zero-shot super-resolution, (iii) continuous temporal extrapolation, (iv) support for 1D, 2D, and 3D PDEs, and (v) efficient inference for longer temporal rollouts. To address these limitations, we propose Vectorized Conditional Neural Fields (VCNeFs), which represent the solution of time-dependent PDEs as neural fields. Contrary to prior methods, however, VCNeFs compute, for a set of multiple spatio-temporal query points, their solutions in parallel and model their dependencies through attention mechanisms. Moreover, VCNeF can condition the neural field on both the initial conditions and the parameters of the PDEs. An extensive set of experiments demonstrates that VCNeFs are competitive with and often outperform existing ML-based surrogate models.
|
https://proceedings.mlr.press/v235/hahm24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/hahm24a/hahm24a.pdf
|
https://openreview.net/forum?id=ufCptn28vG
|
Isometric Representation Learning for Disentangled Latent Space of Diffusion Models
|
https://proceedings.mlr.press/v235/hahm24a.html
|
Jaehoon Hahm, Junho Lee, Sunghyun Kim, Joonseok Lee
|
https://proceedings.mlr.press/v235/hahm24a.html
|
ICML 2024
|
The latent space of diffusion model mostly still remains unexplored, despite its great success and potential in the field of generative modeling. In fact, the latent space of existing diffusion models are entangled, with a distorted mapping from its latent space to image space. To tackle this problem, we present Isometric Diffusion, equipping a diffusion model with a geometric regularizer to guide the model to learn a geometrically sound latent space of the training data manifold. This approach allows diffusion models to learn a more disentangled latent space, which enables smoother interpolation, more accurate inversion, and more precise control over attributes directly in the latent space. Our extensive experiments consisting of image interpolations, image inversions, and linear editing show the effectiveness of our method.
|
https://proceedings.mlr.press/v235/hahn24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/hahn24a/hahn24a.pdf
|
https://openreview.net/forum?id=foPMkomvk1
|
Pursuing Overall Welfare in Federated Learning through Sequential Decision Making
|
https://proceedings.mlr.press/v235/hahn24a.html
|
Seok-Ju Hahn, Gi-Soo Kim, Junghye Lee
|
https://proceedings.mlr.press/v235/hahn24a.html
|
ICML 2024
|
In traditional federated learning, a single global model cannot perform equally well for all clients. Therefore, the need to achieve the client-level fairness in federated system has been emphasized, which can be realized by modifying the static aggregation scheme for updating the global model to an adaptive one, in response to the local signals of the participating clients. Our work reveals that existing fairness-aware aggregation strategies can be unified into an online convex optimization framework, in other words, a central server’s sequential decision making process. To enhance the decision making capability, we propose simple and intuitive improvements for suboptimal designs within existing methods, presenting $\texttt{AAggFF}$. Considering practical requirements, we further subdivide our method tailored for the cross-device and the cross-silo settings, respectively. Theoretical analyses guarantee sublinear regret upper bounds for both settings: $\mathcal{O}(\sqrt{T \log{K}})$ for the cross-device setting, and $\mathcal{O}(K \log{T})$ for the cross-silo setting, with $K$ clients and $T$ federation rounds. Extensive experiments demonstrate that the federated system equipped with $\texttt{AAggFF}$ achieves better degree of client-level fairness than existing methods in both practical settings.
|
https://proceedings.mlr.press/v235/hajj24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/hajj24a/hajj24a.pdf
|
https://openreview.net/forum?id=GfNyqrwECJ
|
Incorporating probabilistic domain knowledge into deep multiple instance learning
|
https://proceedings.mlr.press/v235/hajj24a.html
|
Ghadi S. Al Hajj, Aliaksandr Hubin, Chakravarthi Kanduri, Milena Pavlovic, Knut Dagestad Rand, Michael Widrich, Anne Schistad Solberg, Victor Greiff, Johan Pensar, Günter Klambauer, Geir Kjetil Sandve
|
https://proceedings.mlr.press/v235/hajj24a.html
|
ICML 2024
|
Deep learning methods, including deep multiple instance learning methods, have been criticized for their limited ability to incorporate domain knowledge. A reason that knowledge incorporation is challenging in deep learning is that the models usually lack a mapping between their model components and the entities of the domain, making it a non-trivial task to incorporate probabilistic prior information. In this work, we show that such a mapping between domain entities and model components can be defined for a multiple instance learning setting and propose a framework DeeMILIP that encompasses multiple strategies to exploit this mapping for prior knowledge incorporation. We motivate and formalize these strategies from a probabilistic perspective. Experiments on an immune-based diagnostics case show that our proposed strategies allow to learn generalizable models even in settings with weak signals, limited dataset size, and limited compute.
|
https://proceedings.mlr.press/v235/halawi24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/halawi24a/halawi24a.pdf
|
https://openreview.net/forum?id=6PqWuSuWvX
|
Covert Malicious Finetuning: Challenges in Safeguarding LLM Adaptation
|
https://proceedings.mlr.press/v235/halawi24a.html
|
Danny Halawi, Alexander Wei, Eric Wallace, Tony Tong Wang, Nika Haghtalab, Jacob Steinhardt
|
https://proceedings.mlr.press/v235/halawi24a.html
|
ICML 2024
|
Black-box finetuning is an emerging interface for adapting state-of-the-art language models to user needs. However, such access may also let malicious actors undermine model safety. To demonstrate the challenge of defending finetuning interfaces, we introduce covert malicious finetuning, a method to compromise model safety via finetuning while evading detection. Our method constructs a malicious dataset where every individual datapoint appears innocuous, but finetuning on the dataset teaches the model to respond to encoded harmful requests with encoded harmful responses. Applied to GPT-4, our method produces a finetuned model that acts on harmful instructions 99% of the time and avoids detection by defense mechanisms such as dataset inspection, safety evaluations, and input/output classifiers. Our findings question whether black-box finetuning access can be secured against sophisticated adversaries.
|
https://proceedings.mlr.press/v235/hallak24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/hallak24a/hallak24a.pdf
|
https://openreview.net/forum?id=OndZHBUA1G
|
A Study of First-Order Methods with a Deterministic Relative-Error Gradient Oracle
|
https://proceedings.mlr.press/v235/hallak24a.html
|
Nadav Hallak, Kfir Yehuda Levy
|
https://proceedings.mlr.press/v235/hallak24a.html
|
ICML 2024
|
This paper studies the theoretical guarantees of the classical projected gradient and conditional gradient methods applied to constrained optimization problems with biased relative-error gradient oracles. These oracles are used in various settings, such as distributed optimization systems or derivative-free optimization, and are particularly common when gradients are compressed, quantized, or estimated via finite differences computations. Several settings are investigated: Optimization over the box with a coordinate-wise erroneous gradient oracle, optimization over a general compact convex set, and three more specific scenarios. Convergence guarantees are established with respect to the relative-error magnitude, and in particular, we show that the conditional gradient is invariant to relative-error when applied over the box with a coordinate-wise erroneous gradient oracle, and the projected gradient maintains its convergence guarantees when optimizing a nonconvex objective function.
|
https://proceedings.mlr.press/v235/hamed24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/hamed24a/hamed24a.pdf
|
https://openreview.net/forum?id=HsseRq2FAx
|
Dr. Strategy: Model-Based Generalist Agents with Strategic Dreaming
|
https://proceedings.mlr.press/v235/hamed24a.html
|
Hany Hamed, Subin Kim, Dongyeong Kim, Jaesik Yoon, Sungjin Ahn
|
https://proceedings.mlr.press/v235/hamed24a.html
|
ICML 2024
|
Model-based reinforcement learning (MBRL) has been a primary approach to ameliorating the sample efficiency issue as well as to make a generalist agent. However, there has not been much effort toward enhancing the strategy of dreaming itself. Therefore, it is a question whether and how an agent can “dream better” in a more structured and strategic way. In this paper, inspired by the observation from cognitive science suggesting that humans use a spatial divide-and-conquer strategy in planning, we propose a new MBRL agent, called Dr. Strategy, which is equipped with a novel Dreaming Strategy. The proposed agent realizes a version of divide-and-conquer-like strategy in dreaming. This is achieved by learning a set of latent landmarks and then utilizing these to learn a landmark-conditioned highway policy. With the highway policy, the agent can first learn in the dream to move to a landmark, and from there it tackles the exploration and achievement task in a more focused way. In experiments, we show that the proposed model outperforms prior pixel-based MBRL methods in various visually complex and partially observable navigation tasks.
|
https://proceedings.mlr.press/v235/han24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/han24a/han24a.pdf
|
https://openreview.net/forum?id=vFATIZXlCm
|
UGrid: An Efficient-And-Rigorous Neural Multigrid Solver for Linear PDEs
|
https://proceedings.mlr.press/v235/han24a.html
|
Xi Han, Fei Hou, Hong Qin
|
https://proceedings.mlr.press/v235/han24a.html
|
ICML 2024
|
Numerical solvers of Partial Differential Equations (PDEs) are of fundamental significance to science and engineering. To date, the historical reliance on legacy techniques has circumscribed possible integration of big data knowledge and exhibits sub-optimal efficiency for certain PDE formulations, while data-driven neural methods typically lack mathematical guarantee of convergence and correctness. This paper articulates a mathematically rigorous neural solver for linear PDEs. The proposed UGrid solver, built upon the principled integration of U-Net and MultiGrid, manifests a mathematically rigorous proof of both convergence and correctness, and showcases high numerical accuracy, as well as strong generalization power to various input geometry/values and multiple PDE formulations. In addition, we devise a new residual loss metric, which enables unsupervised training and affords more stability and a larger solution space over the legacy losses.
|
https://proceedings.mlr.press/v235/han24b.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/han24b/han24b.pdf
|
https://openreview.net/forum?id=v8MgLJ7kbL
|
Model Assessment and Selection under Temporal Distribution Shift
|
https://proceedings.mlr.press/v235/han24b.html
|
Elise Han, Chengpiao Huang, Kaizheng Wang
|
https://proceedings.mlr.press/v235/han24b.html
|
ICML 2024
|
We investigate model assessment and selection in a changing environment, by synthesizing datasets from both the current time period and historical epochs. To tackle unknown and potentially arbitrary temporal distribution shift, we develop an adaptive rolling window approach to estimate the generalization error of a given model. This strategy also facilitates the comparison between any two candidate models by estimating the difference of their generalization errors. We further integrate pairwise comparisons into a single-elimination tournament, achieving near-optimal model selection from a collection of candidates. Theoretical analyses and empirical experiments underscore the adaptivity of our proposed methods to the non-stationarity in data.
|
https://proceedings.mlr.press/v235/han24c.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/han24c/han24c.pdf
|
https://openreview.net/forum?id=bdKaQmrM81
|
Riemannian coordinate descent algorithms on matrix manifolds
|
https://proceedings.mlr.press/v235/han24c.html
|
Andi Han, Pratik Jawanpuria, Bamdev Mishra
|
https://proceedings.mlr.press/v235/han24c.html
|
ICML 2024
|
Many machine learning applications are naturally formulated as optimization problems on Riemannian manifolds. The main idea behind Riemannian optimization is to maintain the feasibility of the variables while moving along a descent direction on the manifold. This results in updating all the variables at every iteration. In this work, we provide a general framework for developing computationally efficient coordinate descent (CD) algorithms on matrix manifolds that allows updating only a few variables at every iteration while adhering to the manifold constraint. In particular, we propose CD algorithms for various manifolds such as Stiefel, Grassmann, (generalized) hyperbolic, symplectic, and symmetric positive (semi)definite. While the cost per iteration of the proposed CD algorithms is low, we further develop a more efficient variant via a first-order approximation of the objective function. We analyze their convergence and complexity, and empirically illustrate their efficacy in several applications.
|
https://proceedings.mlr.press/v235/han24d.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/han24d/han24d.pdf
|
https://openreview.net/forum?id=JOrLz5d7OW
|
Prototypical Transformer As Unified Motion Learners
|
https://proceedings.mlr.press/v235/han24d.html
|
Cheng Han, Yawen Lu, Guohao Sun, James Chenhao Liang, Zhiwen Cao, Qifan Wang, Qiang Guan, Sohail Dianat, Raghuveer Rao, Tong Geng, Zhiqiang Tao, Dongfang Liu
|
https://proceedings.mlr.press/v235/han24d.html
|
ICML 2024
|
In this work, we introduce the Prototypical Transformer (ProtoFormer), a general and unified framework that approaches various motion tasks from a prototype perspective. ProtoFormer seamlessly integrates prototype learning with Transformer by thoughtfully considering motion dynamics, introducing two innovative designs. First, Cross-Attention Prototyping discovers prototypes based on signature motion patterns, providing transparency in understanding motion scenes. Second, Latent Synchronization guides feature representation learning via prototypes, effectively mitigating the problem of motion uncertainty. Empirical results demonstrate that our approach achieves competitive performance on popular motion tasks such as optical flow and scene depth. Furthermore, it exhibits generality across various downstream tasks, including object tracking and video stabilization.
|
https://proceedings.mlr.press/v235/han24e.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/han24e/han24e.pdf
|
https://openreview.net/forum?id=cUMOVfOIve
|
SIN: Selective and Interpretable Normalization for Long-Term Time Series Forecasting
|
https://proceedings.mlr.press/v235/han24e.html
|
Lu Han, Han-Jia Ye, De-Chuan Zhan
|
https://proceedings.mlr.press/v235/han24e.html
|
ICML 2024
|
In real-world applications, time series data frequently exhibit non-stationarity, with statistics changing over time. This variability undermines the forecasting accuracy of deep learning models that are trained on historical data but deployed for future prediction. A common approach to mitigate this issue involves normalizing the data to counteract statistical drift, followed by denormalization on the prediction. However, existing methods often employ heuristic normalization techniques that do not fully account for the unique characteristics of the series. Our paper addresses the critical question in this context: which statistics should be removed and restored? We argue that the statistics selected for normalization should exhibit both local invariance and global variability to ensure their correctness and helpfulness. To this end, we propose the Selective and Interpretable Normalization methodology, dubbed SIN. This approach maximizes the covariance between a given look-back window and its subsequent future values, thereby identifying key statistics for normalization and simultaneously learning the corresponding normalization transformations. The interpretable framework can be used to explain the success and limitations of some popular normalization methods. By integrating SIN, we demonstrate improvements in the performance of several prevalent forecasting models, thereby validating the utility of our approach.
|
https://proceedings.mlr.press/v235/han24f.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/han24f/han24f.pdf
|
https://openreview.net/forum?id=fRG45xL1WT
|
Large Language Models Can Automatically Engineer Features for Few-Shot Tabular Learning
|
https://proceedings.mlr.press/v235/han24f.html
|
Sungwon Han, Jinsung Yoon, Sercan O Arik, Tomas Pfister
|
https://proceedings.mlr.press/v235/han24f.html
|
ICML 2024
|
Large Language Models (LLMs), with their remarkable ability to tackle challenging and unseen reasoning problems, hold immense potential for tabular learning, that is vital for many real-world applications. In this paper, we propose a novel in-context learning framework, FeatLLM, which employs LLMs as feature engineers to produce an input data set that is optimally suited for tabular predictions. The generated features are used to infer class likelihood with a simple downstream machine learning model, such as linear regression and yields high performance few-shot learning. The proposed FeatLLM framework only uses this simple predictive model with the discovered features at inference time. Compared to existing LLM-based approaches, FeatLLM eliminates the need to send queries to the LLM for each sample at inference time. Moreover, it merely requires API-level access to LLMs, and overcomes prompt size limitations. As demonstrated across numerous tabular datasets from a wide range of domains, FeatLLM generates high-quality rules, significantly (10% on average) outperforming alternatives such as TabLLM and STUNT.
|
https://proceedings.mlr.press/v235/han24g.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/han24g/han24g.pdf
|
https://openreview.net/forum?id=KycvgOCBBR
|
Improving Group Robustness on Spurious Correlation Requires Preciser Group Inference
|
https://proceedings.mlr.press/v235/han24g.html
|
Yujin Han, Difan Zou
|
https://proceedings.mlr.press/v235/han24g.html
|
ICML 2024
|
Standard empirical risk minimization (ERM) models may prioritize learning spurious correlations between spurious features and true labels, leading to poor accuracy on groups where these correlations do not hold. Mitigating this issue often requires expensive spurious attribute (group) labels or relies on trained ERM models to infer group labels when group information is unavailable. However, the significant performance gap in worst-group accuracy between using pseudo group labels and using oracle group labels inspires us to consider further improving group robustness through preciser group inference. Therefore, we propose GIC, a novel method that accurately infers group labels, resulting in improved worst-group performance. GIC trains a spurious attribute classifier based on two key properties of spurious correlations: (1) high correlation between spurious attributes and true labels, and (2) variability in this correlation between datasets with different group distributions. Empirical studies on multiple datasets demonstrate the effectiveness of GIC in inferring group labels, and combining GIC with various downstream invariant learning methods improves worst-group accuracy, showcasing its powerful flexibility. Additionally, through analyzing the misclassifications in GIC, we identify an interesting phenomenon called semantic consistency, which may contribute to better decoupling the association between spurious attributes and labels, thereby mitigating spurious correlation. The code for GIC is available at https://github.com/yujinhanml/GIC9.
|
https://proceedings.mlr.press/v235/hang24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/hang24a/hang24a.pdf
|
https://openreview.net/forum?id=Irkcamqg4d
|
Binary Decomposition: A Problem Transformation Perspective for Open-Set Semi-Supervised Learning
|
https://proceedings.mlr.press/v235/hang24a.html
|
Jun-Yi Hang, Min-Ling Zhang
|
https://proceedings.mlr.press/v235/hang24a.html
|
ICML 2024
|
Semi-supervised learning (SSL) is a classical machine learning paradigm dealing with labeled and unlabeled data. However, it often suffers performance degradation in real-world open-set scenarios, where unlabeled data contains outliers from novel categories that do not appear in labeled data. Existing studies commonly tackle this challenging open-set SSL problem with detect-and-filter strategy, which attempts to purify unlabeled data by detecting and filtering outliers. In this paper, we propose a novel binary decomposition strategy, which refrains from error-prone procedure of outlier detection by directly transforming the original open-set SSL problem into a number of standard binary SSL problems. Accordingly, a concise yet effective approach named BDMatch is presented. BDMatch confronts two attendant issues brought by binary decomposition, i.e. class-imbalance and representation-compromise, with adaptive logit adjustment and label-specific feature learning respectively. Comprehensive experiments on diversified benchmarks clearly validate the superiority of BDMatch as well as the effectiveness of our binary decomposition strategy.
|
https://proceedings.mlr.press/v235/hans24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/hans24a/hans24a.pdf
|
https://openreview.net/forum?id=axl3FAkpik
|
Spotting LLMs With Binoculars: Zero-Shot Detection of Machine-Generated Text
|
https://proceedings.mlr.press/v235/hans24a.html
|
Abhimanyu Hans, Avi Schwarzschild, Valeriia Cherepanova, Hamid Kazemi, Aniruddha Saha, Micah Goldblum, Jonas Geiping, Tom Goldstein
|
https://proceedings.mlr.press/v235/hans24a.html
|
ICML 2024
|
Detecting text generated by modern large language models is thought to be hard, as both LLMs and humans can exhibit a wide range of complex behaviors. However, we find that a score based on contrasting two closely related language models is highly accurate at separating human-generated and machine-generated text. Based on this mechanism, we propose a novel LLM detector that only requires simple calculations using a pair of pre-trained LLMs. The method, called Binoculars, achieves state-of-the-art accuracy without any training data. It is capable of spotting machine text from a range of modern LLMs without any model-specific modifications. We comprehensively evaluate Binoculars on a number of text sources and in varied situations. Over a wide range of document types, Binoculars detects over 90% of generated samples from ChatGPT (and other LLMs) at a false positive rate of 0.01%, despite not being trained on any ChatGPT data. Code available at https://github.com/ahans30/Binoculars.
|
https://proceedings.mlr.press/v235/hansen24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/hansen24a/hansen24a.pdf
|
https://openreview.net/forum?id=vFk9fqXLst
|
Interpreting Equivariant Representations
|
https://proceedings.mlr.press/v235/hansen24a.html
|
Andreas Abildtrup Hansen, Anna Calissano, Aasa Feragen
|
https://proceedings.mlr.press/v235/hansen24a.html
|
ICML 2024
|
Latent representations are extensively used for tasks like visualization, interpolation, or feature extraction in deep learning models. This paper demonstrates the importance of considering the inductive bias imposed by an equivariant model when using latent representations as neglecting these biases can lead to decreased performance in downstream tasks. We propose principles for choosing invariant projections of latent representations and show their effectiveness in two examples: A permutation equivariant variational auto-encoder for molecular graph generation, where an invariant projection can be designed to maintain information without loss, and for a rotation-equivariant representation in image classification, where random invariant projections proves to retain a high degree of information. In both cases, the analysis of invariant latent representations proves superior to their equivariant counterparts. Finally, we illustrate that the phenomena documented here for equivariant neural networks have counterparts in standard neural networks where invariance is encouraged via augmentation.
|
https://proceedings.mlr.press/v235/hao24a.html
|
https://raw.githubusercontent.com/mlresearch/v235/main/assets/hao24a/hao24a.pdf
|
https://openreview.net/forum?id=uubBZKM99Y
|
Flora: Low-Rank Adapters Are Secretly Gradient Compressors
|
https://proceedings.mlr.press/v235/hao24a.html
|
Yongchang Hao, Yanshuai Cao, Lili Mou
|
https://proceedings.mlr.press/v235/hao24a.html
|
ICML 2024
|
Despite large neural networks demonstrating remarkable abilities to complete different tasks, they require excessive memory usage to store the optimization states for training. To alleviate this, the low-rank adaptation (LoRA) is proposed to reduce the optimization states by training fewer parameters. However, LoRA restricts overall weight update matrices to be low-rank, limiting the model performance. In this work, we investigate the dynamics of LoRA and identify that it can be approximated by a random projection. Based on this observation, we propose Flora, which is able to achieve high-rank updates by resampling the projection matrices while enjoying the sublinear space complexity of optimization states. We conduct experiments across different tasks and model architectures to verify the effectiveness of our approach.
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.