publication_date
stringlengths 20
20
| doc_url
stringlengths 33
34
| id
stringlengths 33
34
| authors
stringlengths 5
20.1k
| update_date
stringlengths 20
20
| category_all
stringlengths 5
138
| abstract
stringlengths 123
3.09k
| category_primary
stringclasses 78
values | title
stringlengths 10
211
|
|---|---|---|---|---|---|---|---|---|
2024-02-09T01:18:16Z | http://arxiv.org/pdf/2402.06126v3 | http://arxiv.org/abs/2402.06126v3 | Haizhong Zheng, Xiaoyan Bai, Xueshen Liu, Z. Morley Mao, Beidi Chen, Fan Lai, Atul Prakash | 2024-06-03T18:28:58Z | cs.CL, cs.AI, cs.LG | Large Language Models (LLMs) have achieved remarkable success with their
billion-level parameters, yet they incur high inference overheads. The
emergence of activation sparsity in LLMs provides a natural approach to reduce
this cost by involving only parts of the parameters for inference. However,
existing methods only focus on utilizing this naturally formed activation
sparsity in a post-training setting, overlooking the potential for further
amplifying this inherent sparsity. In this paper, we hypothesize that LLMs can
learn to be efficient by achieving more structured activation sparsity. To
achieve this, we introduce a novel training algorithm, Learn-To-be-Efficient
(LTE), designed to train efficiency-aware LLMs to learn to activate fewer
neurons and achieve a better trade-off between sparsity and performance.
Furthermore, unlike SOTA MoEfication methods, which mainly focus on ReLU-based
models, LTE can also be applied to LLMs like LLaMA using non-ReLU activations.
Extensive evaluation on language understanding, language generation, and
instruction tuning tasks show that LTE consistently outperforms SOTA baselines.
Along with our hardware-aware custom kernel implementation, LTE reduces
LLaMA2-7B inference latency by 25% at 50% sparsity. | cs.CL | Learn To be Efficient: Build Structured Sparsity in Large Language Models |
2024-02-09T01:15:42Z | http://arxiv.org/pdf/2402.06125v1 | http://arxiv.org/abs/2402.06125v1 | Joshua Zingale, Jugal Kalita | 2024-02-09T01:15:42Z | cs.CL | Controlled text generation (CTG) seeks to guide large language model (LLM)
output to produce text that conforms to desired criteria. The current study
presents a novel CTG algorithm that enforces adherence toward specific
rhetorical relations in an LLM sentence-completion context by a parser-driven
decoding scheme that requires no model fine-tuning. The method is validated
both with automatic and human evaluation. The code is accessible on GitHub. | cs.CL | Language Model Sentence Completion with a Parser-Driven Rhetorical Control Method |
2024-02-09T01:10:25Z | http://arxiv.org/pdf/2402.06120v1 | http://arxiv.org/abs/2402.06120v1 | Shima Imani, Hamid Palangi | 2024-02-09T01:10:25Z | cs.CL | Large Language Models (LLMs) have demonstrated impressive performance across
a wide range of applications; however, assessing their reasoning capabilities
remains a significant challenge. In this paper, we introduce a framework
grounded in group and symmetry principles, which have played a crucial role in
fields such as physics and mathematics, and offer another way to evaluate their
capabilities. While the proposed framework is general, to showcase the benefits
of employing these properties, we focus on arithmetic reasoning and investigate
the performance of these models on four group properties: closure, identity,
inverse, and associativity. Our findings reveal that LLMs studied in this work
struggle to preserve group properties across different test regimes. In the
closure test, we observe biases towards specific outputs and an abrupt
degradation in their performance from 100% to 0% after a specific sequence
length. They also perform poorly in the identity test, which represents adding
irrelevant information in the context, and show sensitivity when subjected to
inverse test, which examines the robustness of the model with respect to
negation. In addition, we demonstrate that breaking down problems into smaller
steps helps LLMs in the associativity test that we have conducted. To support
these tests we have developed a synthetic dataset which will be released. | cs.CL | Exploring Group and Symmetry Principles in Large Language Models |
2024-02-08T23:02:04Z | http://arxiv.org/pdf/2402.06094v1 | http://arxiv.org/abs/2402.06094v1 | Ming Shen | 2024-02-08T23:02:04Z | cs.CL | Although supervised finetuning (SFT) has emerged as an essential technique to
align large language models with humans, it is considered superficial, with
style learning being its nature. At the same time, recent works indicate the
importance of data selection for SFT, showing that finetuning with high-quality
and diverse subsets of the original dataset leads to superior downstream
performance. In this work, we rethink the intuition behind data selection for
SFT. Considering SFT is superficial, we propose that essential demonstrations
for SFT should focus on reflecting human-like interactions instead of data
quality or diversity. However, it is not straightforward to directly assess to
what extent a demonstration reflects human styles. Towards an initial attempt
in this direction, we find selecting instances with long responses is
surprisingly more effective for SFT than utilizing full datasets or instances
selected based on quality and diversity. We hypothesize that such a simple
heuristic implicitly mimics a crucial aspect of human-style conversation:
detailed responses are usually more helpful. | cs.CL | Rethinking Data Selection for Supervised Fine-Tuning |
2024-02-08T21:47:16Z | http://arxiv.org/pdf/2402.06073v2 | http://arxiv.org/abs/2402.06073v2 | Di Cao, Xianchen Wang, Junfeng Zhou, Jiakai Zhang, Yanjing Lei, Wenpeng Chen | 2024-02-12T15:28:38Z | cs.CL, cs.SD, eess.AS | Traditional Time Delay Neural Networks (TDNN) have achieved state-of-the-art
performance at the cost of high computational complexity and slower inference
speed, making them difficult to implement in an industrial environment. The
Densely Connected Time Delay Neural Network (D-TDNN) with Context Aware Masking
(CAM) module has proven to be an efficient structure to reduce complexity while
maintaining system performance. In this paper, we propose a fast and
lightweight model, LightCAM, which further adopts a depthwise separable
convolution module (DSM) and uses multi-scale feature aggregation (MFA) for
feature fusion at different levels. Extensive experiments are conducted on
VoxCeleb dataset, the comparative results show that it has achieved an EER of
0.83 and MinDCF of 0.0891 in VoxCeleb1-O, which outperforms the other
mainstream speaker verification methods. In addition, complexity analysis
further demonstrates that the proposed architecture has lower computational
cost and faster inference speed. | cs.CL | LightCAM: A Fast and Light Implementation of Context-Aware Masking based D-TDNN for Speaker Verification |
2024-02-08T20:35:31Z | http://arxiv.org/pdf/2402.16877v1 | http://arxiv.org/abs/2402.16877v1 | Austin Xu, Will Monroe, Klinton Bicknell | 2024-02-08T20:35:31Z | cs.IR, cs.AI, cs.CL, cs.LG | We study the problem of zero-shot exercise retrieval in the context of online
language learning, to give learners the ability to explicitly request
personalized exercises via natural language. Using real-world data collected
from language learners, we observe that vector similarity approaches poorly
capture the relationship between exercise content and the language that
learners use to express what they want to learn. This semantic gap between
queries and content dramatically reduces the effectiveness of general-purpose
retrieval models pretrained on large scale information retrieval datasets like
MS MARCO. We leverage the generative capabilities of large language models to
bridge the gap by synthesizing hypothetical exercises based on the learner's
input, which are then used to search for relevant exercises. Our approach,
which we call mHyER, overcomes three challenges: (1) lack of relevance labels
for training, (2) unrestricted learner input content, and (3) low semantic
similarity between input and retrieval candidates. mHyER outperforms several
strong baselines on two novel benchmarks created from crowdsourced data and
publicly available data. | cs.IR | Large Language Model Augmented Exercise Retrieval for Personalized Language Learning |
2024-02-08T20:35:06Z | http://arxiv.org/pdf/2402.06044v3 | http://arxiv.org/abs/2402.06044v3 | Hainiu Xu, Runcong Zhao, Lixing Zhu, Jinhua Du, Yulan He | 2024-06-03T10:48:16Z | cs.AI, cs.CL | Neural Theory-of-Mind (N-ToM), machine's ability to understand and keep track
of the mental states of others, is pivotal in developing socially intelligent
agents. However, prevalent N-ToM benchmarks have several shortcomings,
including the presence of ambiguous and artificial narratives, absence of
personality traits and preferences, a lack of questions addressing characters'
psychological mental states, and limited diversity in the questions posed. In
response to these issues, we construct OpenToM, a new benchmark for assessing
N-ToM with (1) longer and clearer narrative stories, (2) characters with
explicit personality traits, (3) actions that are triggered by character
intentions, and (4) questions designed to challenge LLMs' capabilities of
modeling characters' mental states of both the physical and psychological
world. Using OpenToM, we reveal that state-of-the-art LLMs thrive at modeling
certain aspects of mental states in the physical world but fall short when
tracking characters' mental states in the psychological world. | cs.AI | OpenToM: A Comprehensive Benchmark for Evaluating Theory-of-Mind Reasoning Capabilities of Large Language Models |
2024-02-08T20:24:44Z | http://arxiv.org/pdf/2402.06041v1 | http://arxiv.org/abs/2402.06041v1 | Beatrice Savoldi, Andrea Piergentili, Dennis Fucci, Matteo Negri, Luisa Bentivogli | 2024-02-08T20:24:44Z | cs.CL | Gender-neutral translation (GNT) that avoids biased and undue binary
assumptions is a pivotal challenge for the creation of more inclusive
translation technologies. Advancements for this task in Machine Translation
(MT), however, are hindered by the lack of dedicated parallel data, which are
necessary to adapt MT systems to satisfy neutral constraints. For such a
scenario, large language models offer hitherto unforeseen possibilities, as
they come with the distinct advantage of being versatile in various (sub)tasks
when provided with explicit instructions. In this paper, we explore this
potential to automate GNT by comparing MT with the popular GPT-4 model. Through
extensive manual analyses, our study empirically reveals the inherent
limitations of current MT systems in generating GNTs and provides valuable
insights into the potential and challenges associated with prompting for
neutrality. | cs.CL | A Prompt Response to the Demand for Automatic Gender-Neutral Translation |
2024-02-08T19:57:29Z | http://arxiv.org/pdf/2402.06025v6 | http://arxiv.org/abs/2402.06025v6 | Wasu Top Piriyakulkij, Cassidy Langenfeld, Tuan Anh Le, Kevin Ellis | 2024-05-22T21:24:19Z | cs.AI, cs.CL | We give a model of how to infer natural language rules by doing experiments.
The model integrates Large Language Models (LLMs) with Monte Carlo algorithms
for probabilistic inference, interleaving online belief updates with experiment
design under information-theoretic criteria. We conduct a human-model
comparison on a Zendo-style task, finding that a critical ingredient for
modeling the human data is to assume that humans also consider fuzzy,
probabilistic rules, in addition to assuming that humans perform
approximately-Bayesian belief updates. We also compare with recent algorithms
for using LLMs to generate and revise hypotheses, finding that our online
inference method yields higher accuracy at recovering the true underlying rule,
and provides better support for designing optimal experiments. | cs.AI | Doing Experiments and Revising Rules with Natural Language and Probabilistic Reasoning |
2024-02-08T19:25:40Z | http://arxiv.org/pdf/2402.06015v2 | http://arxiv.org/abs/2402.06015v2 | Yong Cao, Wenyan Li, Jiaang Li, Yifei Yuan, Antonia Karamolegkou, Daniel Hershcovich | 2024-02-15T10:39:19Z | cs.CL, cs.CV | Pretrained large Vision-Language models have drawn considerable interest in
recent years due to their remarkable performance. Despite considerable efforts
to assess these models from diverse perspectives, the extent of visual cultural
awareness in the state-of-the-art GPT-4V model remains unexplored. To tackle
this gap, we extensively probed GPT-4V using the MaRVL benchmark dataset,
aiming to investigate its capabilities and limitations in visual understanding
with a focus on cultural aspects. Specifically, we introduced three visual
related tasks, i.e. caption classification, pairwise captioning, and culture
tag selection, to systematically delve into fine-grained visual cultural
evaluation. Experimental results indicate that GPT-4V excels at identifying
cultural concepts but still exhibits weaker performance in low-resource
languages, such as Tamil and Swahili. Notably, through human evaluation, GPT-4V
proves to be more culturally relevant in image captioning tasks than the
original MaRVL human annotations, suggesting a promising solution for future
visual cultural benchmark construction. | cs.CL | Exploring Visual Culture Awareness in GPT-4V: A Comprehensive Probing |
2024-02-08T18:59:48Z | http://arxiv.org/pdf/2402.05935v2 | http://arxiv.org/abs/2402.05935v2 | Dongyang Liu, Renrui Zhang, Longtian Qiu, Siyuan Huang, Weifeng Lin, Shitian Zhao, Shijie Geng, Ziyi Lin, Peng Jin, Kaipeng Zhang, Wenqi Shao, Chao Xu, Conghui He, Junjun He, Hao Shao, Pan Lu, Hongsheng Li, Yu Qiao, Peng Gao | 2024-06-26T07:59:03Z | cs.CV, cs.AI, cs.CL, cs.LG | We propose SPHINX-X, an extensive Multimodality Large Language Model (MLLM)
series developed upon SPHINX. To improve the architecture and training
efficiency, we modify the SPHINX framework by removing redundant visual
encoders, bypassing fully-padded sub-images with skip tokens, and simplifying
multi-stage training into a one-stage all-in-one paradigm. To fully unleash the
potential of MLLMs, we assemble a comprehensive multi-domain and multimodal
dataset covering publicly available resources in language, vision, and
vision-language tasks. We further enrich this collection with our curated OCR
intensive and Set-of-Mark datasets, extending the diversity and generality. By
training over different base LLMs including TinyLlama1.1B, InternLM2-7B,
LLaMA2-13B, and Mixtral8x7B, we obtain a spectrum of MLLMs that vary in
parameter size and multilingual capabilities. Comprehensive benchmarking
reveals a strong correlation between the multi-modal performance with the data
and parameter scales. Code and models are released at
https://github.com/Alpha-VLLM/LLaMA2-Accessory | cs.CV | SPHINX-X: Scaling Data and Parameters for a Family of Multi-modal Large Language Models |
2024-02-08T18:59:03Z | http://arxiv.org/pdf/2402.05932v2 | http://arxiv.org/abs/2402.05932v2 | Boyi Li, Yue Wang, Jiageng Mao, Boris Ivanovic, Sushant Veer, Karen Leung, Marco Pavone | 2024-04-10T23:29:18Z | cs.RO, cs.AI, cs.CL | Adapting driving behavior to new environments, customs, and laws is a
long-standing problem in autonomous driving, precluding the widespread
deployment of autonomous vehicles (AVs). In this paper, we present LLaDA, a
simple yet powerful tool that enables human drivers and autonomous vehicles
alike to drive everywhere by adapting their tasks and motion plans to traffic
rules in new locations. LLaDA achieves this by leveraging the impressive
zero-shot generalizability of large language models (LLMs) in interpreting the
traffic rules in the local driver handbook. Through an extensive user study, we
show that LLaDA's instructions are useful in disambiguating in-the-wild
unexpected situations. We also demonstrate LLaDA's ability to adapt AV motion
planning policies in real-world datasets; LLaDA outperforms baseline planning
approaches on all our metrics. Please check our website for more details:
https://boyiliee.github.io/llada. | cs.RO | Driving Everywhere with Large Language Model Policy Adaptation |
2024-02-08T18:58:02Z | http://arxiv.org/pdf/2402.05930v1 | http://arxiv.org/abs/2402.05930v1 | Xing Han Lù, Zdeněk Kasner, Siva Reddy | 2024-02-08T18:58:02Z | cs.CL, cs.CV, cs.LG | We propose the problem of conversational web navigation, where a digital
agent controls a web browser and follows user instructions to solve real-world
tasks in a multi-turn dialogue fashion. To support this problem, we introduce
WEBLINX - a large-scale benchmark of 100K interactions across 2300 expert
demonstrations of conversational web navigation. Our benchmark covers a broad
range of patterns on over 150 real-world websites and can be used to train and
evaluate agents in diverse scenarios. Due to the magnitude of information
present, Large Language Models (LLMs) cannot process entire web pages in
real-time. To solve this bottleneck, we design a retrieval-inspired model that
efficiently prunes HTML pages by ranking relevant elements. We use the selected
elements, along with screenshots and action history, to assess a variety of
models for their ability to replicate human behavior when navigating the web.
Our experiments span from small text-only to proprietary multimodal LLMs. We
find that smaller finetuned decoders surpass the best zero-shot LLMs (including
GPT-4V), but also larger finetuned multimodal models which were explicitly
pretrained on screenshots. However, all finetuned models struggle to generalize
to unseen websites. Our findings highlight the need for large multimodal models
that can generalize to novel settings. Our code, data and models are available
for research: https://mcgill-nlp.github.io/weblinx | cs.CL | WebLINX: Real-World Website Navigation with Multi-Turn Dialogue |
2024-02-08T18:56:40Z | http://arxiv.org/pdf/2402.05926v3 | http://arxiv.org/abs/2402.05926v3 | Zhenqing Ling, Daoyuan Chen, Liuyi Yao, Yaliang Li, Ying Shen | 2024-06-17T16:00:59Z | cs.LG, cs.CL | The confluence of Federated Learning (FL) and Large Language Models (LLMs) is
ushering in a new era in privacy-preserving natural language processing.
However, the intensive memory requirements for fine-tuning LLMs pose
significant challenges, especially when deploying on clients with limited
computational resources. To circumvent this, we explore the novel integration
of Memory-efficient Zeroth-Order Optimization within a federated setting, a
synergy we term as FedMeZO. Our study is the first to examine the theoretical
underpinnings of FedMeZO in the context of LLMs, tackling key questions
regarding the influence of large parameter spaces on optimization behavior, the
establishment of convergence properties, and the identification of critical
parameters for convergence to inform personalized federated strategies. Our
extensive empirical evidence supports the theory, showing that FedMeZO not only
converges faster than traditional first-order methods such as FedAvg but also
significantly reduces GPU memory usage during training to levels comparable to
those during inference. Moreover, the proposed personalized FL strategy that is
built upon the theoretical insights to customize the client-wise learning rate
can effectively accelerate loss reduction. We hope our work can help to bridge
theoretical and practical aspects of federated fine-tuning for LLMs, thereby
stimulating further advancements and research in this area. | cs.LG | On the Convergence of Zeroth-Order Federated Tuning for Large Language Models |
2024-02-08T18:49:09Z | http://arxiv.org/pdf/2402.05913v1 | http://arxiv.org/abs/2402.05913v1 | Abhishek Panigrahi, Nikunj Saunshi, Kaifeng Lyu, Sobhan Miryoosefi, Sashank Reddi, Satyen Kale, Sanjiv Kumar | 2024-02-08T18:49:09Z | cs.CL, cs.LG | Recent developments in large language models have sparked interest in
efficient pretraining methods. A recent effective paradigm is to perform
stage-wise training, where the size of the model is gradually increased over
the course of training (e.g. gradual stacking (Reddi et al., 2023)). While the
resource and wall-time savings are appealing, it has limitations, particularly
the inability to evaluate the full model during earlier stages, and degradation
in model quality due to smaller model capacity in the initial stages. In this
work, we propose an alternative framework, progressive subnetwork training,
that maintains the full model throughout training, but only trains subnetworks
within the model in each step. We focus on a simple instantiation of this
framework, Random Path Training (RaPTr) that only trains a sub-path of layers
in each step, progressively increasing the path lengths in stages. RaPTr
achieves better pre-training loss for BERT and UL2 language models while
requiring 20-33% fewer FLOPs compared to standard training, and is competitive
or better than other efficient training methods. Furthermore, RaPTr shows
better downstream performance on UL2, improving QA tasks and SuperGLUE by 1-5%
compared to standard training and stacking. Finally, we provide a theoretical
basis for RaPTr to justify (a) the increasing complexity of subnetworks in
stages, and (b) the stability in loss across stage transitions due to residual
connections and layer norm. | cs.CL | Efficient Stagewise Pretraining via Progressive Subnetworks |
2024-02-08T18:43:05Z | http://arxiv.org/pdf/2402.05904v1 | http://arxiv.org/abs/2402.05904v1 | Eun Cheol Choi, Emilio Ferrara | 2024-02-08T18:43:05Z | cs.CL, cs.CY, cs.HC, cs.SI | Our society is facing rampant misinformation harming public health and trust.
To address the societal challenge, we introduce FACT-GPT, a system leveraging
Large Language Models (LLMs) to automate the claim matching stage of
fact-checking. FACT-GPT, trained on a synthetic dataset, identifies social
media content that aligns with, contradicts, or is irrelevant to previously
debunked claims. Our evaluation shows that our specialized LLMs can match the
accuracy of larger models in identifying related claims, closely mirroring
human judgment. This research provides an automated solution for efficient
claim matching, demonstrates the potential of LLMs in supporting fact-checkers,
and offers valuable resources for further research in the field. | cs.CL | FACT-GPT: Fact-Checking Augmentation via Claim Matching with LLMs |
2024-02-08T18:27:22Z | http://arxiv.org/pdf/2402.05889v2 | http://arxiv.org/abs/2402.05889v2 | Shoubin Yu, Jaehong Yoon, Mohit Bansal | 2024-06-12T21:03:33Z | cs.CV, cs.AI, cs.CL | Despite impressive advancements in recent multimodal reasoning approaches,
they are still limited in flexibility and efficiency, as these models typically
process only a few fixed modality inputs and require updates to numerous
parameters. This paper tackles these critical challenges and proposes CREMA, a
generalizable, highly efficient, and modular modality-fusion framework that can
incorporate any new modality to enhance video reasoning. We first augment
multiple informative modalities (such as optical flow, 3D point cloud, audio,
thermal heatmap, and touch map) from given videos without extra human
annotation by leveraging sensors or existing pre-trained models. Next, we
introduce a query transformer with multiple parameter-efficient modules
associated with each accessible modality. It projects diverse modality features
to the LLM token embedding space, allowing the model to integrate different
data types for response generation. Furthermore, we propose a novel progressive
multimodal fusion design supported by a lightweight fusion module and
modality-sequential training strategy. It helps compress information across
various assisting modalities, maintaining computational efficiency in the LLM
while improving performance. We validate our method on 7 video-language
reasoning tasks assisted by diverse modalities, including conventional VideoQA
and Video-Audio/3D/Touch/Thermal QA, and achieve better/equivalent performance
against strong multimodal LLMs, including OneLLM, BLIP-2, and SeViLA while
reducing over 90% trainable parameters. We provide extensive analyses of CREMA,
including the impact of each modality on reasoning domains, the design of the
fusion module, and example visualizations. | cs.CV | CREMA: Generalizable and Efficient Video-Language Reasoning via Multimodal Modular Fusion |
2024-02-08T18:14:33Z | http://arxiv.org/pdf/2402.05880v2 | http://arxiv.org/abs/2402.05880v2 | Nikhil Sharma, Q. Vera Liao, Ziang Xiao | 2024-02-10T17:03:58Z | cs.CL, cs.AI, cs.HC | Large language models (LLMs) powered conversational search systems have
already been used by hundreds of millions of people, and are believed to bring
many benefits over conventional search. However, while decades of research and
public discourse interrogated the risk of search systems in increasing
selective exposure and creating echo chambers -- limiting exposure to diverse
opinions and leading to opinion polarization, little is known about such a risk
of LLM-powered conversational search. We conduct two experiments to
investigate: 1) whether and how LLM-powered conversational search increases
selective exposure compared to conventional search; 2) whether and how LLMs
with opinion biases that either reinforce or challenge the user's view change
the effect. Overall, we found that participants engaged in more biased
information querying with LLM-powered conversational search, and an opinionated
LLM reinforcing their views exacerbated this bias. These results present
critical implications for the development of LLMs and conversational search
systems, and the policy governing these technologies. | cs.CL | Generative Echo Chamber? Effects of LLM-Powered Search Systems on Diverse Information Seeking |
2024-02-08T17:57:11Z | http://arxiv.org/pdf/2402.05868v2 | http://arxiv.org/abs/2402.05868v2 | Guo Lin, Wenyue Hua, Yongfeng Zhang | 2024-02-12T16:26:14Z | cs.CL, cs.AI, cs.CR, cs.IR, cs.LG | Cloud-based large language models (LLMs) such as ChatGPT have increasingly
become integral to daily operations, serving as vital tools across various
applications. While these models offer substantial benefits in terms of
accessibility and functionality, they also introduce significant privacy
concerns: the transmission and storage of user data in cloud infrastructures
pose substantial risks of data breaches and unauthorized access to sensitive
information; even if the transmission and storage of data is encrypted, the LLM
service provider itself still knows the real contents of the data, preventing
individuals or entities from confidently using such LLM services. To address
these concerns, this paper proposes a simple yet effective mechanism EmojiCrypt
to protect user privacy. It uses Emoji to encrypt the user inputs before
sending them to LLM, effectively rendering them indecipherable to human or
LLM's examination while retaining the original intent of the prompt, thus
ensuring the model's performance remains unaffected. We conduct experiments on
three tasks, personalized recommendation, sentiment analysis, and tabular data
analysis. Experiment results reveal that EmojiCrypt can encrypt personal
information within prompts in such a manner that not only prevents the
discernment of sensitive data by humans or LLM itself, but also maintains or
even improves the precision without further tuning, achieving comparable or
even better task accuracy than directly prompting the LLM without prompt
encryption. These results highlight the practicality of adopting encryption
measures that safeguard user privacy without compromising the functional
integrity and performance of LLMs. Code and dataset are available at
https://github.com/agiresearch/EmojiCrypt. | cs.CL | EmojiCrypt: Prompt Encryption for Secure Communication with Large Language Models |
2024-02-08T17:54:23Z | http://arxiv.org/pdf/2402.05864v1 | http://arxiv.org/abs/2402.05864v1 | Xuandong Zhao, Lei Li, Yu-Xiang Wang | 2024-02-08T17:54:23Z | cs.CL, cs.CR, cs.LG | In this paper, we propose a new decoding method called Permute-and-Flip (PF)
decoder. It enjoys robustness properties similar to the standard sampling
decoder, but is provably up to 2x better in its quality-robustness tradeoff
than sampling and never worse than any other decoder. We also design a
cryptographic watermarking scheme analogous to Aaronson's Gumbel watermark, but
naturally tailored for PF decoder. The watermarking scheme does not change the
distribution to sample, while allowing arbitrarily low false positive rate and
high recall whenever the generated text has high entropy. Our experiments show
that the PF decoder (and its watermarked counterpart) significantly
outperform(s) naive sampling (and it's Gumbel watermarked counterpart) in terms
of perplexity, while retaining the same robustness (and detectability), hence
making it a promising new approach for LLM decoding. The code is available at
https://github.com/XuandongZhao/pf-decoding | cs.CL | Permute-and-Flip: An optimally robust and watermarkable decoder for LLMs |
2024-02-08T17:51:48Z | http://arxiv.org/pdf/2402.05863v1 | http://arxiv.org/abs/2402.05863v1 | Federico Bianchi, Patrick John Chia, Mert Yuksekgonul, Jacopo Tagliabue, Dan Jurafsky, James Zou | 2024-02-08T17:51:48Z | cs.AI, cs.CL, cs.GT | Negotiation is the basis of social interactions; humans negotiate everything
from the price of cars to how to share common resources. With rapidly growing
interest in using large language models (LLMs) to act as agents on behalf of
human users, such LLM agents would also need to be able to negotiate. In this
paper, we study how well LLMs can negotiate with each other. We develop
NegotiationArena: a flexible framework for evaluating and probing the
negotiation abilities of LLM agents. We implemented three types of scenarios in
NegotiationArena to assess LLM's behaviors in allocating shared resources
(ultimatum games), aggregate resources (trading games) and buy/sell goods
(price negotiations). Each scenario allows for multiple turns of flexible
dialogues between LLM agents to allow for more complex negotiations.
Interestingly, LLM agents can significantly boost their negotiation outcomes by
employing certain behavioral tactics. For example, by pretending to be desolate
and desperate, LLMs can improve their payoffs by 20\% when negotiating against
the standard GPT-4. We also quantify irrational negotiation behaviors exhibited
by the LLM agents, many of which also appear in humans. Together,
\NegotiationArena offers a new environment to investigate LLM interactions,
enabling new insights into LLM's theory of mind, irrationality, and reasoning
abilities. | cs.AI | How Well Can LLMs Negotiate? NegotiationArena Platform and Analysis |
2024-02-08T17:10:12Z | http://arxiv.org/pdf/2402.06690v1 | http://arxiv.org/abs/2402.06690v1 | Mitodru Niyogi | 2024-02-08T17:10:12Z | cs.SE, cs.CL, cs.LG, cs.PL | Natural language (NL) to code suggestion systems assist developers in
Integrated Development Environments (IDEs) by translating NL utterances into
compilable code snippet. The current approaches mainly involve hard-coded,
rule-based systems based on semantic parsing. These systems make heavy use of
hand-crafted rules that map patterns in NL or elements in its syntax parse tree
to various query constructs and can only work on a limited subset of NL with a
restricted NL syntax. These systems are unable to extract semantic information
from the coding intents of the developer, and often fail to infer types, names,
and the context of the source code to get accurate system-level code
suggestions. In this master thesis, we present sequence-to-sequence deep
learning models and training paradigms to map NL to general-purpose programming
languages that can assist users with suggestions of source code snippets, given
a NL intent, and also extend auto-completion functionality of the source code
to users while they are writing source code. The developed architecture
incorporates contextual awareness into neural models which generate source code
tokens directly instead of generating parse trees/abstract meaning
representations from the source code and converting them back to source code.
The proposed pretraining strategy and the data augmentation techniques improve
the performance of the proposed architecture. The proposed architecture has
been found to exceed the performance of a neural semantic parser, TranX, based
on the BLEU-4 metric by 10.82%. Thereafter, a finer analysis for the parsable
code translations from the NL intent for CoNaLA challenge was introduced. The
proposed system is bidirectional as it can be also used to generate NL code
documentation given source code. Lastly, a RoBERTa masked language model for
Python was proposed to extend the developed system for code completion. | cs.SE | Neural Models for Source Code Synthesis and Completion |
2024-02-08T17:06:45Z | http://arxiv.org/pdf/2402.05827v1 | http://arxiv.org/abs/2402.05827v1 | Xinbei Ma, Tianjie Ju, Jiyang Qiu, Zhuosheng Zhang, Hai Zhao, Lifeng Liu, Yulong Wang | 2024-02-08T17:06:45Z | cs.CL | Large language models (LLMs) have played a pivotal role in building
communicative AI to imitate human behaviors but face the challenge of efficient
customization. To tackle this challenge, recent studies have delved into the
realm of model editing, which manipulates specific memories of language models
and changes the related language generation. However, the robustness of model
editing remains an open question. This work seeks to understand the strengths
and limitations of editing methods, thus facilitating robust, realistic
applications of communicative AI. Concretely, we conduct extensive analysis to
address the three key research questions. Q1: Can edited LLMs behave
consistently resembling communicative AI in realistic situations? Q2: To what
extent does the rephrasing of prompts lead LLMs to deviate from the edited
knowledge memory? Q3: Which knowledge features are correlated with the
performance and robustness of editing? Our experimental results uncover a
substantial disparity between existing editing methods and the practical
application of LLMs. On rephrased prompts that are complex and flexible but
common in realistic applications, the performance of editing experiences a
significant decline. Further analysis shows that more popular knowledge is
memorized better, easier to recall, and more challenging to edit effectively. | cs.CL | Is it Possible to Edit Large Language Models Robustly? |
2024-02-08T16:55:21Z | http://arxiv.org/pdf/2402.05819v1 | http://arxiv.org/abs/2402.05819v1 | Hung-Chieh Fang, Nai-Xuan Ye, Yi-Jen Shih, Puyuan Peng, Hsuan-Fu Wang, Layne Berry, Hung-yi Lee, David Harwath | 2024-02-08T16:55:21Z | eess.AS, cs.CL, cs.LG | Recent advances in self-supervised speech models have shown significant
improvement in many downstream tasks. However, these models predominantly
centered on frame-level training objectives, which can fall short in spoken
language understanding tasks that require semantic comprehension. Existing
works often rely on additional speech-text data as intermediate targets, which
is costly in the real-world setting. To address this challenge, we propose
Pseudo-Word HuBERT (PW-HuBERT), a framework that integrates pseudo word-level
targets into the training process, where the targets are derived from a
visually-ground speech model, notably eliminating the need for speech-text
paired data. Our experimental results on four spoken language understanding
(SLU) benchmarks suggest the superiority of our model in capturing semantic
information. | eess.AS | Integrating Self-supervised Speech Model with Pseudo Word-level Targets from Visually-grounded Speech Model |
2024-02-08T16:50:01Z | http://arxiv.org/pdf/2402.05813v1 | http://arxiv.org/abs/2402.05813v1 | Lingzhi Wang, Xingshan Zeng, Jinsong Guo, Kam-Fai Wong, Georg Gottlob | 2024-02-08T16:50:01Z | cs.CL, cs.AI | The aim of this study is to investigate Machine Unlearning (MU), a burgeoning
field focused on addressing concerns related to neural models inadvertently
retaining personal or sensitive data. Here, a novel approach is introduced to
achieve precise and selective forgetting within language models. Unlike
previous methodologies that adopt completely opposing training objectives, this
approach aims to mitigate adverse effects on language model performance,
particularly in generation tasks. Furthermore, two innovative evaluation
metrics are proposed: Sensitive Information Extraction Likelihood (S-EL) and
Sensitive Information Memory Accuracy (S-MA), designed to gauge the
effectiveness of sensitive information elimination. To reinforce the forgetting
framework, an effective method for annotating sensitive scopes is presented,
involving both online and offline strategies. The online selection mechanism
leverages language probability scores to ensure computational efficiency, while
the offline annotation entails a robust two-stage process based on Large
Language Models (LLMs). | cs.CL | Selective Forgetting: Advancing Machine Unlearning Techniques and Evaluation in Language Models |
2024-02-08T16:49:41Z | http://arxiv.org/pdf/2402.05812v3 | http://arxiv.org/abs/2402.05812v3 | Sahil Kale, Gautam Khaire, Jay Patankar | 2024-05-09T10:29:16Z | cs.CL | Frequently Asked Questions (FAQs) refer to the most common inquiries about
specific content. They serve as content comprehension aids by simplifying
topics and enhancing understanding through succinct presentation of
information. In this paper, we address FAQ generation as a well-defined Natural
Language Processing task through the development of an end-to-end system
leveraging text-to-text transformation models. We present a literature review
covering traditional question-answering systems, highlighting their limitations
when applied directly to the FAQ generation task. We propose a system capable
of building FAQs from textual content tailored to specific domains, enhancing
their accuracy and relevance. We utilise self-curated algorithms to obtain an
optimal representation of information to be provided as input and also to rank
the question-answer pairs to maximise human comprehension. Qualitative human
evaluation showcases the generated FAQs as well-constructed and readable while
also utilising domain-specific constructs to highlight domain-based nuances and
jargon in the original content. | cs.CL | FAQ-Gen: An automated system to generate domain-specific FAQs to aid content comprehension |
2024-02-08T16:46:26Z | http://arxiv.org/pdf/2402.05808v2 | http://arxiv.org/abs/2402.05808v2 | Zhiheng Xi, Wenxiang Chen, Boyang Hong, Senjie Jin, Rui Zheng, Wei He, Yiwen Ding, Shichun Liu, Xin Guo, Junzhe Wang, Honglin Guo, Wei Shen, Xiaoran Fan, Yuhao Zhou, Shihan Dou, Xiao Wang, Xinbo Zhang, Peng Sun, Tao Gui, Qi Zhang, Xuanjing Huang | 2024-03-17T09:02:02Z | cs.AI, cs.CL, cs.LG | In this paper, we propose R$^3$: Learning Reasoning through Reverse
Curriculum Reinforcement Learning (RL), a novel method that employs only
outcome supervision to achieve the benefits of process supervision for large
language models. The core challenge in applying RL to complex reasoning is to
identify a sequence of actions that result in positive rewards and provide
appropriate supervision for optimization. Outcome supervision provides sparse
rewards for final results without identifying error locations, whereas process
supervision offers step-wise rewards but requires extensive manual annotation.
R$^3$ overcomes these limitations by learning from correct demonstrations.
Specifically, R$^3$ progressively slides the start state of reasoning from a
demonstration's end to its beginning, facilitating easier model exploration at
all stages. Thus, R$^3$ establishes a step-wise curriculum, allowing outcome
supervision to offer step-level signals and precisely pinpoint errors. Using
Llama2-7B, our method surpasses RL baseline on eight reasoning tasks by $4.1$
points on average. Notebaly, in program-based reasoning on GSM8K, it exceeds
the baseline by $4.2$ points across three backbone models, and without any
extra data, Codellama-7B + R$^3$ performs comparable to larger models or
closed-source models. | cs.AI | Training Large Language Models for Reasoning through Reverse Curriculum Reinforcement Learning |
2024-02-08T16:36:11Z | http://arxiv.org/pdf/2402.05794v1 | http://arxiv.org/abs/2402.05794v1 | Marcellus Amadeus, William Alberto Cruz Castañeda, Wilmer Lobato, Niasche Aquino | 2024-02-08T16:36:11Z | cs.CL, cs.AI | Speech technologies rely on capturing a speaker's voice variability while
obtaining comprehensive language information. Textual prompts and sentence
selection methods have been proposed in the literature to comprise such
adequate phonetic data, referred to as a phonetically rich \textit{corpus}.
However, they are still insufficient for acoustic modeling, especially critical
for languages with limited resources. Hence, this paper proposes a novel
approach and outlines the methodological aspects required to create a
\textit{corpus} with broad phonetic coverage for a low-resourced language,
Brazilian Portuguese. Our methodology includes text dataset collection up to a
sentence selection algorithm based on triphone distribution. Furthermore, we
propose a new phonemic classification according to acoustic-articulatory speech
features since the absolute number of distinct triphones, or low-probability
triphones, does not guarantee an adequate representation of every possible
combination. Using our algorithm, we achieve a 55.8\% higher percentage of
distinct triphones -- for samples of similar size -- while the currently
available phonetic-rich corpus, CETUC and TTS-Portuguese, 12.6\% and 12.3\% in
comparison to a non-phonetically rich dataset. | cs.CL | Phonetically rich corpus construction for a low-resourced language |
2024-02-08T16:23:29Z | http://arxiv.org/pdf/2402.05785v3 | http://arxiv.org/abs/2402.05785v3 | Jonathan Thomm, Aleksandar Terzic, Giacomo Camposampiero, Michael Hersche, Bernhard Schölkopf, Abbas Rahimi | 2024-05-25T11:09:28Z | cs.LG, cs.AI, cs.CL | We analyze the capabilities of Transformer language models in learning
compositional discrete tasks. To this end, we evaluate training LLaMA models
and prompting GPT-4 and Gemini on four tasks demanding to learn a composition
of several discrete sub-tasks. On both training LLaMA models from scratch and
prompting on GPT-4 and Gemini, we measure how well these models can reuse
primitives observable in the sub-tasks to learn the composition task. Our
results indicate that compositional learning in state-of-the-art Transformer
language models is highly sample inefficient: LLaMA requires more data samples
than relearning all sub-tasks from scratch to learn the compositional task;
in-context prompting with few samples is unreliable and fails at executing the
sub-tasks or correcting the errors in multi-round code generation. Further, by
leveraging complexity theory, we support these findings with a theoretical
analysis focused on the sample inefficiency of gradient descent in memorizing
feedforward models. | cs.LG | Limits of Transformer Language Models on Learning to Compose Algorithms |
2024-02-08T16:17:24Z | http://arxiv.org/pdf/2402.05783v2 | http://arxiv.org/abs/2402.05783v2 | Fenia Christopoulou, Guchun Zhang, Gerasimos Lampouras | 2024-02-12T10:44:18Z | cs.CL | Large pre-trained language models have recently been expanded and applied to
programming language tasks with great success, often through further
pre-training of a strictly-natural language model--where training sequences
typically contain both natural and (linearised) programming language. Such
approaches effectively map both modalities of the sequence into the same
embedding space. However, programming language keywords (e.g. "while") often
have very strictly defined semantics. As such, transfer learning from their
natural language usage may not necessarily be beneficial to their code
application and vise versa. Assuming an already pre-trained language model, in
this work we investigate how sequence tokens can be adapted and represented
differently, depending on which modality they belong to, and to the ultimate
benefit of the downstream task. We experiment with separating embedding spaces
between modalities during further model pre-training with modality-relative
training objectives. We focus on text-to-code generation and observe consistent
improvements across two backbone models and two test sets, measuring pass@$k$
and a novel incremental variation. | cs.CL | Text-to-Code Generation with Modality-relative Pre-training |
2024-02-08T16:11:23Z | http://arxiv.org/pdf/2402.05779v1 | http://arxiv.org/abs/2402.05779v1 | Kathleen C. Fraser, Svetlana Kiritchenko | 2024-02-08T16:11:23Z | cs.CY, cs.CL, cs.CV | Following on recent advances in large language models (LLMs) and subsequent
chat models, a new wave of large vision-language models (LVLMs) has emerged.
Such models can incorporate images as input in addition to text, and perform
tasks such as visual question answering, image captioning, story generation,
etc. Here, we examine potential gender and racial biases in such systems, based
on the perceived characteristics of the people in the input images. To
accomplish this, we present a new dataset PAIRS (PArallel Images for eveRyday
Scenarios). The PAIRS dataset contains sets of AI-generated images of people,
such that the images are highly similar in terms of background and visual
content, but differ along the dimensions of gender (man, woman) and race
(Black, white). By querying the LVLMs with such images, we observe significant
differences in the responses according to the perceived gender or race of the
person depicted. | cs.CY | Examining Gender and Racial Bias in Large Vision-Language Models Using a Novel Dataset of Parallel Images |
2024-02-08T15:40:35Z | http://arxiv.org/pdf/2402.07939v5 | http://arxiv.org/abs/2402.07939v5 | Chaoyun Zhang, Liqun Li, Shilin He, Xu Zhang, Bo Qiao, Si Qin, Minghua Ma, Yu Kang, Qingwei Lin, Saravan Rajmohan, Dongmei Zhang, Qi Zhang | 2024-05-23T05:18:58Z | cs.HC, cs.AI, cs.CL | We introduce UFO, an innovative UI-Focused agent to fulfill user requests
tailored to applications on Windows OS, harnessing the capabilities of
GPT-Vision. UFO employs a dual-agent framework to meticulously observe and
analyze the graphical user interface (GUI) and control information of Windows
applications. This enables the agent to seamlessly navigate and operate within
individual applications and across them to fulfill user requests, even when
spanning multiple applications. The framework incorporates a control
interaction module, facilitating action grounding without human intervention
and enabling fully automated execution. Consequently, UFO transforms arduous
and time-consuming processes into simple tasks achievable solely through
natural language commands. We conducted testing of UFO across 9 popular Windows
applications, encompassing a variety of scenarios reflective of users' daily
usage. The results, derived from both quantitative metrics and real-case
studies, underscore the superior effectiveness of UFO in fulfilling user
requests. To the best of our knowledge, UFO stands as the first UI agent
specifically tailored for task completion within the Windows OS environment.
The open-source code for UFO is available on https://github.com/microsoft/UFO. | cs.HC | UFO: A UI-Focused Agent for Windows OS Interaction |
2024-02-08T15:39:32Z | http://arxiv.org/pdf/2402.05755v1 | http://arxiv.org/abs/2402.05755v1 | Tu Anh Nguyen, Benjamin Muller, Bokai Yu, Marta R. Costa-jussa, Maha Elbayad, Sravya Popuri, Paul-Ambroise Duquenne, Robin Algayres, Ruslan Mavlyutov, Itai Gat, Gabriel Synnaeve, Juan Pino, Benoit Sagot, Emmanuel Dupoux | 2024-02-08T15:39:32Z | cs.CL, cs.SD, eess.AS | We introduce SPIRIT-LM, a foundation multimodal language model that freely
mixes text and speech. Our model is based on a pretrained text language model
that we extend to the speech modality by continuously training it on text and
speech units. Speech and text sequences are concatenated as a single set of
tokens, and trained with a word-level interleaving method using a small
automatically-curated speech-text parallel corpus. SPIRIT-LM comes in two
versions: a BASE version that uses speech semantic units and an EXPRESSIVE
version that models expressivity using pitch and style units in addition to the
semantic units. For both versions, the text is encoded with subword BPE tokens.
The resulting model displays both the semantic abilities of text models and the
expressive abilities of speech models. Additionally, we demonstrate that
SPIRIT-LM is able to learn new tasks in a few-shot fashion across modalities
(i.e. ASR, TTS, Speech Classification). | cs.CL | SpiRit-LM: Interleaved Spoken and Written Language Model |
2024-02-08T15:08:57Z | http://arxiv.org/pdf/2402.05733v1 | http://arxiv.org/abs/2402.05733v1 | Yikai Zhang, Siyu Yuan, Caiyu Hu, Kyle Richardson, Yanghua Xiao, Jiangjie Chen | 2024-02-08T15:08:57Z | cs.CL | Despite remarkable advancements in emulating human-like behavior through
Large Language Models (LLMs), current textual simulations do not adequately
address the notion of time. To this end, we introduce TimeArena, a novel
textual simulated environment that incorporates complex temporal dynamics and
constraints that better reflect real-life planning scenarios. In TimeArena,
agents are asked to complete multiple tasks as soon as possible, allowing for
parallel processing to save time. We implement the dependency between actions,
the time duration for each action, and the occupancy of the agent and the
objects in the environment. TimeArena grounds to 30 real-world tasks in
cooking, household activities, and laboratory work. We conduct extensive
experiments with various state-of-the-art LLMs using TimeArena. Our findings
reveal that even the most powerful models, e.g., GPT-4, still lag behind humans
in effective multitasking, underscoring the need for enhanced temporal
awareness in the development of language agents. | cs.CL | TimeArena: Shaping Efficient Multitasking Language Agents in a Time-Aware Simulation |
2024-02-08T14:35:09Z | http://arxiv.org/pdf/2402.05706v1 | http://arxiv.org/abs/2402.05706v1 | Heeseung Kim, Soonshin Seo, Kyeongseok Jeong, Ohsung Kwon, Jungwhan Kim, Jaehong Lee, Eunwoo Song, Myungwoo Oh, Sungroh Yoon, Kang Min Yoo | 2024-02-08T14:35:09Z | cs.CL, cs.SD, eess.AS | While recent work shows promising results in expanding the capabilities of
large language models (LLM) to directly understand and synthesize speech, an
LLM-based strategy for modeling spoken dialogs remains elusive and calls for
further investigation. This work proposes an extensive speech-text LLM
framework, named the Unified Spoken Dialog Model (USDM), to generate coherent
spoken responses with organic prosodic features relevant to the given input
speech without relying on automatic speech recognition (ASR) or text-to-speech
(TTS) solutions. Our approach employs a multi-step speech-text inference scheme
that leverages chain-of-reasoning capabilities exhibited by the underlying LLM.
We also propose a generalized speech-text pretraining scheme that helps with
capturing cross-modal semantics. Automatic and human evaluations show that the
proposed approach is effective in generating natural-sounding spoken responses,
outperforming both prior and cascaded baselines. Detailed comparative studies
reveal that, despite the cascaded approach being stronger in individual
components, the joint speech-text modeling improves robustness against
recognition errors and speech quality. Demo is available at
https://unifiedsdm.github.io. | cs.CL | Unified Speech-Text Pretraining for Spoken Dialog Modeling |
2024-02-08T14:21:03Z | http://arxiv.org/pdf/2402.05699v3 | http://arxiv.org/abs/2402.05699v3 | Xianghe Pang, Shuo Tang, Rui Ye, Yuxin Xiong, Bolun Zhang, Yanfeng Wang, Siheng Chen | 2024-06-08T06:13:55Z | cs.CL, cs.AI, cs.CY | Aligning large language models (LLMs) with human values is imperative to
mitigate potential adverse effects resulting from their misuse. Drawing from
the sociological insight that acknowledging all parties' concerns is a key
factor in shaping human values, this paper proposes a novel direction to align
LLMs by themselves: social scene simulation. To achieve this, we present
MATRIX, a novel social scene simulator that emulates realistic scenes around a
user's input query, enabling the LLM to take social consequences into account
before responding. MATRIX serves as a virtual rehearsal space, akin to a
Monopolylogue, where the LLM performs diverse roles related to the query and
practice by itself. To inject this alignment, we fine-tune the LLM with
MATRIX-simulated data, ensuring adherence to human values without compromising
inference speed. We theoretically show that the LLM with MATRIX outperforms
Constitutional AI under mild assumptions. Finally, extensive experiments
validate that our method outperforms over 10 baselines across 4 benchmarks. As
evidenced by 875 user ratings, our tuned 13B-size LLM exceeds GPT-4 in aligning
with human values. See our project page at
https://shuotang123.github.io/MATRIX. | cs.CL | Self-Alignment of Large Language Models via Monopolylogue-based Social Scene Simulation |
2024-02-08T13:47:50Z | http://arxiv.org/pdf/2402.05672v1 | http://arxiv.org/abs/2402.05672v1 | Liang Wang, Nan Yang, Xiaolong Huang, Linjun Yang, Rangan Majumder, Furu Wei | 2024-02-08T13:47:50Z | cs.CL, cs.IR | This technical report presents the training methodology and evaluation
results of the open-source multilingual E5 text embedding models, released in
mid-2023. Three embedding models of different sizes (small / base / large) are
provided, offering a balance between the inference efficiency and embedding
quality. The training procedure adheres to the English E5 model recipe,
involving contrastive pre-training on 1 billion multilingual text pairs,
followed by fine-tuning on a combination of labeled datasets. Additionally, we
introduce a new instruction-tuned embedding model, whose performance is on par
with state-of-the-art, English-only models of similar sizes. Information
regarding the model release can be found at
https://github.com/microsoft/unilm/tree/master/e5 . | cs.CL | Multilingual E5 Text Embeddings: A Technical Report |
2024-02-08T13:42:50Z | http://arxiv.org/pdf/2402.05668v1 | http://arxiv.org/abs/2402.05668v1 | Junjie Chu, Yugeng Liu, Ziqing Yang, Xinyue Shen, Michael Backes, Yang Zhang | 2024-02-08T13:42:50Z | cs.CR, cs.AI, cs.CL, cs.LG | Misuse of the Large Language Models (LLMs) has raised widespread concern. To
address this issue, safeguards have been taken to ensure that LLMs align with
social ethics. However, recent findings have revealed an unsettling
vulnerability bypassing the safeguards of LLMs, known as jailbreak attacks. By
applying techniques, such as employing role-playing scenarios, adversarial
examples, or subtle subversion of safety objectives as a prompt, LLMs can
produce an inappropriate or even harmful response. While researchers have
studied several categories of jailbreak attacks, they have done so in
isolation. To fill this gap, we present the first large-scale measurement of
various jailbreak attack methods. We concentrate on 13 cutting-edge jailbreak
methods from four categories, 160 questions from 16 violation categories, and
six popular LLMs. Our extensive experimental results demonstrate that the
optimized jailbreak prompts consistently achieve the highest attack success
rates, as well as exhibit robustness across different LLMs. Some jailbreak
prompt datasets, available from the Internet, can also achieve high attack
success rates on many LLMs, such as ChatGLM3, GPT-3.5, and PaLM2. Despite the
claims from many organizations regarding the coverage of violation categories
in their policies, the attack success rates from these categories remain high,
indicating the challenges of effectively aligning LLM policies and the ability
to counter jailbreak attacks. We also discuss the trade-off between the attack
performance and efficiency, as well as show that the transferability of the
jailbreak prompts is still viable, becoming an option for black-box models.
Overall, our research highlights the necessity of evaluating different
jailbreak methods. We hope our study can provide insights for future research
on jailbreak attacks and serve as a benchmark tool for evaluating them for
practitioners. | cs.CR | Comprehensive Assessment of Jailbreak Attacks Against LLMs |
2024-02-08T13:21:44Z | http://arxiv.org/pdf/2402.14579v1 | http://arxiv.org/abs/2402.14579v1 | Hye Jin Kim, Nicolas Lell, Ansgar Scherp | 2024-02-08T13:21:44Z | cs.CV, cs.CL, cs.LG | Text role classification involves classifying the semantic role of textual
elements within scientific charts. For this task, we propose to finetune two
pretrained multimodal document layout analysis models, LayoutLMv3 and UDOP, on
chart datasets. The transformers utilize the three modalities of text, image,
and layout as input. We further investigate whether data augmentation and
balancing methods help the performance of the models. The models are evaluated
on various chart datasets, and results show that LayoutLMv3 outperforms UDOP in
all experiments. LayoutLMv3 achieves the highest F1-macro score of 82.87 on the
ICPR22 test dataset, beating the best-performing model from the ICPR22
CHART-Infographics challenge. Moreover, the robustness of the models is tested
on a synthetic noisy dataset ICPR22-N. Finally, the generalizability of the
models is evaluated on three chart datasets, CHIME-R, DeGruyter, and EconBiz,
for which we added labels for the text roles. Findings indicate that even in
cases where there is limited training data, transformers can be used with the
help of data augmentation and balancing methods. The source code and datasets
are available on GitHub under
https://github.com/hjkimk/text-role-classification | cs.CV | Text Role Classification in Scientific Charts Using Multimodal Transformers |
2024-02-08T12:51:36Z | http://arxiv.org/pdf/2402.10943v1 | http://arxiv.org/abs/2402.10943v1 | Benjamin Kiessling, Gennady Kurin, Matthew Thomas Miller, Kader Smail | 2024-02-08T12:51:36Z | cs.CL, cs.CV | This work presents an accuracy study of the open source OCR engine, Kraken,
on the leading Arabic scholarly journal, al-Abhath. In contrast with other
commercially available OCR engines, Kraken is shown to be capable of producing
highly accurate Arabic-script OCR. The study also assesses the relative
accuracy of typeface-specific and generalized models on the al-Abhath data and
provides a microanalysis of the ``error instances'' and the contextual features
that may have contributed to OCR misrecognition. Building on this analysis, the
paper argues that Arabic-script OCR can be significantly improved through (1) a
more systematic approach to training data production, and (2) the development
of key technological components, especially multi-language models and improved
line segmentation and layout analysis.
Cet article pr{\'e}sente une {\'e}tude d'exactitude du moteur ROC open
source, Krakan, sur la revue acad{\'e}mique arabe de premier rang, al-Abhath.
Contrairement {\`a} d'autres moteurs ROC disponibles sur le march{\'e}, Kraken
se r{\'e}v{\`e}le {\^e}tre capable de produire de la ROC extr{\^e}mement exacte
de l'{\'e}criture arabe. L'{\'e}tude {\'e}value aussi l'exactitude relative des
mod{\`e}les sp{\'e}cifiquement configur{\'e}s {\`a} des polices et celle des
mod{\`e}les g{\'e}n{\'e}ralis{\'e}s sur les donn{\'e}es d'al-Abhath et fournit
une microanalyse des "occurrences d'erreurs", ainsi qu'une microanalyse des
{\'e}l{\'e}ments contextuels qui pourraient avoir contribu{\'e} {\`a} la
m{\'e}reconnaissance ROC. S'appuyant sur cette analyse, cet article fait valoir
que la ROC de l'{\'e}criture arabe peut {\^e}tre consid{\'e}rablement
am{\'e}lior{\'e}e gr{\^a}ce {\`a} (1) une approche plus syst{\'e}matique
d'entra{\^i}nement de la production de donn{\'e}es et (2) gr{\^a}ce au
d{\'e}veloppement de composants technologiques fondamentaux,
notammentl'am{\'e}lioration des mod{\`e}les multilingues, de la segmentation de
ligne et de l'analyse de la mise en page. | cs.CL | Advances and Limitations in Open Source Arabic-Script OCR: A Case Study |
2024-02-08T12:36:29Z | http://arxiv.org/pdf/2402.05629v4 | http://arxiv.org/abs/2402.05629v4 | Cheng-Han Chiang, Hung-yi Lee | 2024-06-07T02:28:40Z | cs.CL | Long-form generations from large language models (LLMs) contain a mix of
factual and non-factual claims, making evaluating factuality difficult. Prior
works evaluate the factuality of a long paragraph by decomposing it into
multiple facts, verifying those facts independently, and aggregating the
results. Such methods assume that combining factual claims forms a factual
paragraph. The above assumption can be violated: we show that strong
open-source models like Llama-chat can generate paragraphs that contain
verifiable facts, but the facts are combined into a non-factual paragraph due
to entity ambiguity. We further reveal that existing factuality metrics,
including FActScore and citation recall, cannot properly evaluate these
non-factual paragraphs and overestimate their factuality. To address this, we
introduce an enhanced metric, D-FActScore, specifically designed for content
with ambiguous entities. We evaluate the D-FActScores of people biographies
generated by retrieval-augmented LLMs. We show that D-FActScore can better
assess the factuality of paragraphs with entity ambiguity than FActScore. We
also find that four widely used open-source LLMs tend to mix information of
distinct entities to form non-factual paragraphs, making their D-FActScore much
lower than FActScore by over 10%. | cs.CL | Merging Facts, Crafting Fallacies: Evaluating the Contradictory Nature of Aggregated Factual Claims in Long-Form Generations |
2024-02-08T12:28:18Z | http://arxiv.org/pdf/2402.05624v1 | http://arxiv.org/abs/2402.05624v1 | Christoph Tillmann, Aashka Trivedi, Bishwaranjan Bhattacharjee | 2024-02-08T12:28:18Z | cs.CL, cs.AI, cs.HC | Large Language Models (LLMs) are the cornerstone for many Natural Language
Processing (NLP) tasks like sentiment analysis, document classification, named
entity recognition, question answering, summarization, etc. LLMs are often
trained on data which originates from the web. This data is prone to having
content with Hate, Abuse and Profanity (HAP). For a detailed definition of HAP,
please refer to the Appendix. Due to the LLMs being exposed to HAP content
during training, the models learn it and may then generate hateful or profane
content. For example, when the open-source RoBERTa model (specifically, the
RoBERTA base model) from the HuggingFace (HF) Transformers library is prompted
to replace the mask token in `I do not know that Persian people are that MASK`
it returns the word `stupid` with the highest score. This is unacceptable in
civil discourse.The detection of Hate, Abuse and Profanity in text is a vital
component of creating civil and unbiased LLMs, which is needed not only for
English, but for all languages. In this article, we briefly describe the
creation of HAP detectors and various ways of using them to make models civil
and acceptable in the output they generate. | cs.CL | Efficient Models for the Detection of Hate, Abuse and Profanity |
2024-02-08T12:20:28Z | http://arxiv.org/pdf/2402.05617v1 | http://arxiv.org/abs/2402.05617v1 | Elena Senger, Mike Zhang, Rob van der Goot, Barbara Plank | 2024-02-08T12:20:28Z | cs.CL | Recent years have brought significant advances to Natural Language Processing
(NLP), which enabled fast progress in the field of computational job market
analysis. Core tasks in this application domain are skill extraction and
classification from job postings. Because of its quick growth and its
interdisciplinary nature, there is no exhaustive assessment of this emerging
field. This survey aims to fill this gap by providing a comprehensive overview
of deep learning methodologies, datasets, and terminologies specific to
NLP-driven skill extraction and classification. Our comprehensive cataloging of
publicly available datasets addresses the lack of consolidated information on
dataset creation and characteristics. Finally, the focus on terminology
addresses the current lack of consistent definitions for important concepts,
such as hard and soft skills, and terms relating to skill extraction and
classification. | cs.CL | Deep Learning-based Computational Job Market Analysis: A Survey on Skill Extraction and Classification from Job Postings |
2024-02-08T12:19:32Z | http://arxiv.org/pdf/2402.05616v1 | http://arxiv.org/abs/2402.05616v1 | Ben Fauber | 2024-02-08T12:19:32Z | cs.CL, cs.AI, cs.LG | We propose that small pretrained foundational generative language models with
millions of parameters can be utilized as a general learning framework for
sequence-based tasks. Our proposal overcomes the computational resource, skill
set, and timeline challenges associated with training neural networks and
language models from scratch. Further, our approach focuses on creating small
and highly specialized models that can accurately execute a challenging task of
which the base model is incapable of performing. We demonstrate that 125M,
350M, and 1.3B parameter pretrained foundational language models can be
instruction fine-tuned with 10,000-to-1,000,000 instruction examples to achieve
near state-of-the-art results on challenging cheminformatics tasks. We also
demonstrate the role of successive language model fine-tuning epochs on
improved outcomes, as well as the importance of both data formatting and
pretrained foundational language model selection for instruction fine-tuning
success. | cs.CL | Pretrained Generative Language Models as General Learning Frameworks for Sequence-Based Tasks |
2024-02-08T12:01:24Z | http://arxiv.org/pdf/2402.05602v2 | http://arxiv.org/abs/2402.05602v2 | Reduan Achtibat, Sayed Mohammad Vakilzadeh Hatefi, Maximilian Dreyer, Aakriti Jain, Thomas Wiegand, Sebastian Lapuschkin, Wojciech Samek | 2024-06-10T09:58:55Z | cs.CL, cs.AI, cs.CV, cs.LG | Large Language Models are prone to biased predictions and hallucinations,
underlining the paramount importance of understanding their model-internal
reasoning process. However, achieving faithful attributions for the entirety of
a black-box transformer model and maintaining computational efficiency is an
unsolved challenge. By extending the Layer-wise Relevance Propagation
attribution method to handle attention layers, we address these challenges
effectively. While partial solutions exist, our method is the first to
faithfully and holistically attribute not only input but also latent
representations of transformer models with the computational efficiency similar
to a single backward pass. Through extensive evaluations against existing
methods on LLaMa 2, Mixtral 8x7b, Flan-T5 and vision transformer architectures,
we demonstrate that our proposed approach surpasses alternative methods in
terms of faithfulness and enables the understanding of latent representations,
opening up the door for concept-based explanations. We provide an LRP library
at https://github.com/rachtibat/LRP-eXplains-Transformers. | cs.CL | AttnLRP: Attention-Aware Layer-Wise Relevance Propagation for Transformers |
2024-02-08T11:44:25Z | http://arxiv.org/pdf/2402.05591v1 | http://arxiv.org/abs/2402.05591v1 | Juhwan Choi, Kyohoon Jin, Junho Lee, Sangmin Song, Youngbin Kim | 2024-02-08T11:44:25Z | cs.CL, cs.AI | Rule-based text data augmentation is widely used for NLP tasks due to its
simplicity. However, this method can potentially damage the original meaning of
the text, ultimately hurting the performance of the model. To overcome this
limitation, we propose a straightforward technique for applying soft labels to
augmented data. We conducted experiments across seven different classification
tasks and empirically demonstrated the effectiveness of our proposed approach.
We have publicly opened our source code for reproducibility. | cs.CL | SoftEDA: Rethinking Rule-Based Data Augmentation with Soft Labels |
2024-02-08T11:36:23Z | http://arxiv.org/pdf/2402.05584v1 | http://arxiv.org/abs/2402.05584v1 | Juhwan Choi, Kyohoon Jin, Junho Lee, Sangmin Song, Youngbin Kim | 2024-02-08T11:36:23Z | cs.CL, cs.AI | Text data augmentation is a complex problem due to the discrete nature of
sentences. Although rule-based augmentation methods are widely adopted in
real-world applications because of their simplicity, they suffer from potential
semantic damage. Previous researchers have suggested easy data augmentation
with soft labels (softEDA), employing label smoothing to mitigate this problem.
However, finding the best factor for each model and dataset is challenging;
therefore, using softEDA in real-world applications is still difficult. In this
paper, we propose adapting AutoAugment to solve this problem. The experimental
results suggest that the proposed method can boost existing augmentation
methods and that rule-based methods can enhance cutting-edge pre-trained
language models. We offer the source code. | cs.CL | AutoAugment Is What You Need: Enhancing Rule-based Augmentation Methods in Low-resource Regimes |
2024-02-08T11:31:23Z | http://arxiv.org/pdf/2402.05581v1 | http://arxiv.org/abs/2402.05581v1 | Maxime Fily, Guillaume Wisniewski, Severine Guillaume, Gilles Adda, Alexis Michaud | 2024-02-08T11:31:23Z | cs.CL, cs.SD, eess.AS | In the highly constrained context of low-resource language studies, we
explore vector representations of speech from a pretrained model to determine
their level of abstraction with regard to the audio signal. We propose a new
unsupervised method using ABX tests on audio recordings with carefully curated
metadata to shed light on the type of information present in the
representations. ABX tests determine whether the representations computed by a
multilingual speech model encode a given characteristic. Three experiments are
devised: one on room acoustics aspects, one on linguistic genre, and one on
phonetic aspects. The results confirm that the representations extracted from
recordings with different linguistic/extra-linguistic characteristics differ
along the same lines. Embedding more audio signal in one vector better
discriminates extra-linguistic characteristics, whereas shorter snippets are
better to distinguish segmental information. The method is fully unsupervised,
potentially opening new research avenues for comparative work on
under-documented languages. | cs.CL | Establishing degrees of closeness between audio recordings along different dimensions using large-scale cross-lingual models |
2024-02-08T11:16:13Z | http://arxiv.org/pdf/2402.05571v1 | http://arxiv.org/abs/2402.05571v1 | José Alberto Benítez-Andrades, José-Manuel Alija-Pérez, Maria-Esther Vidal, Rafael Pastor-Vargas, María Teresa García-Ordás | 2024-02-08T11:16:13Z | cs.CL, cs.LG | Background: Eating disorders are increasingly prevalent, and social networks
offer valuable information.
Objective: Our goal was to identify efficient machine learning models for
categorizing tweets related to eating disorders.
Methods: Over three months, we collected tweets about eating disorders. A
2,000-tweet subset was labeled for: (1) being written by individuals with
eating disorders, (2) promoting eating disorders, (3) informativeness, and (4)
scientific content. Both traditional machine learning and deep learning models
were employed for classification, assessing accuracy, F1 score, and
computational time.
Results: From 1,058,957 collected tweets, transformer-based bidirectional
encoder representations achieved the highest F1 scores (71.1%-86.4%) across all
four categories.
Conclusions: Transformer-based models outperform traditional techniques in
classifying eating disorder-related tweets, though they require more
computational resources. | cs.CL | Traditional Machine Learning Models and Bidirectional Encoder Representations From Transformer (BERT)-Based Automatic Classification of Tweets About Eating Disorders: Algorithm Development and Validation Study |
2024-02-08T10:32:06Z | http://arxiv.org/pdf/2402.05547v2 | http://arxiv.org/abs/2402.05547v2 | Hengguan Huang, Songtao Wang, Hongfu Liu, Hao Wang, Ye Wang | 2024-06-08T16:36:56Z | cs.CL, cs.AI | Traditional applications of natural language processing (NLP) in healthcare
have predominantly focused on patient-centered services, enhancing patient
interactions and care delivery, such as through medical dialogue systems.
However, the potential of NLP to benefit inexperienced doctors, particularly in
areas such as communicative medical coaching, remains largely unexplored. We
introduce "ChatCoach", a human-AI cooperative framework designed to assist
medical learners in practicing their communication skills during patient
consultations. ChatCoach (Our data and code are available online:
https://github.com/zerowst/Chatcoach)differentiates itself from conventional
dialogue systems by offering a simulated environment where medical learners can
practice dialogues with a patient agent, while a coach agent provides
immediate, structured feedback. This is facilitated by our proposed Generalized
Chain-of-Thought (GCoT) approach, which fosters the generation of structured
feedback and enhances the utilization of external knowledge sources.
Additionally, we have developed a dataset specifically for evaluating Large
Language Models (LLMs) within the ChatCoach framework on communicative medical
coaching tasks. Our empirical results validate the effectiveness of ChatCoach. | cs.CL | Benchmarking Large Language Models on Communicative Medical Coaching: a Novel System and Dataset |
2024-02-08T10:29:11Z | http://arxiv.org/pdf/2402.05545v1 | http://arxiv.org/abs/2402.05545v1 | Bibiána Lajčinová, Patrik Valábek, Michal Spišiak | 2024-02-08T10:29:11Z | cs.CL, I.2.7 | This paper introduces an approach for building a Named Entity Recognition
(NER) model built upon a Bidirectional Encoder Representations from
Transformers (BERT) architecture, specifically utilizing the SlovakBERT model.
This NER model extracts address parts from data acquired from speech-to-text
transcriptions. Due to scarcity of real data, a synthetic dataset using GPT API
was generated. The importance of mimicking spoken language variability in this
artificial data is emphasized. The performance of our NER model, trained solely
on synthetic data, is evaluated using small real test dataset. | cs.CL | Named Entity Recognition for Address Extraction in Speech-to-Text Transcriptions Using Synthetic Data |
2024-02-08T10:15:41Z | http://arxiv.org/pdf/2402.05536v1 | http://arxiv.org/abs/2402.05536v1 | José Alberto Benítez-Andrades, María Teresa García-Ordás, Mayra Russo, Ahmad Sakor, Luis Daniel Fernandes Rotger, Maria-Esther Vidal | 2024-02-08T10:15:41Z | cs.LG, cs.CL | Social networks are vital for information sharing, especially in the health
sector for discussing diseases and treatments. These platforms, however, often
feature posts as brief texts, posing challenges for Artificial Intelligence
(AI) in understanding context. We introduce a novel hybrid approach combining
community-maintained knowledge graphs (like Wikidata) with deep learning to
enhance the categorization of social media posts. This method uses advanced
entity recognizers and linkers (like Falcon 2.0) to connect short post entities
to knowledge graphs. Knowledge graph embeddings (KGEs) and contextualized word
embeddings (like BERT) are then employed to create rich, context-based
representations of these posts.
Our focus is on the health domain, particularly in identifying posts related
to eating disorders (e.g., anorexia, bulimia) to aid healthcare providers in
early diagnosis. We tested our approach on a dataset of 2,000 tweets about
eating disorders, finding that merging word embeddings with knowledge graph
information enhances the predictive models' reliability. This methodology aims
to assist health experts in spotting patterns indicative of mental disorders,
thereby improving early detection and accurate diagnosis for personalized
medicine. | cs.LG | Empowering machine learning models with contextual knowledge for enhancing the detection of eating disorders in social media posts |
2024-02-08T09:48:02Z | http://arxiv.org/pdf/2402.05515v2 | http://arxiv.org/abs/2402.05515v2 | Yufeng Zhao, Yoshihiro Sakai, Naoya Inoue | 2024-02-15T15:25:47Z | cs.CL, cs.AI | In-Context Learning (ICL) is suffering from unsatisfactory performance and
under-calibration due to high prior bias and unfaithful confidence. Some
previous works fine-tuned language models for better ICL performance with
enormous datasets and computing costs. In this paper, we propose NoisyICL,
simply perturbing the model parameters by random noises to strive for better
performance and calibration. Our experiments on two models and 12 downstream
datasets show that NoisyICL can help ICL produce more accurate predictions. Our
further analysis indicates that NoisyICL enables the model to provide more fair
predictions, and also with more faithful confidence. Therefore, we believe that
NoisyICL is an effective calibration of ICL. Our experimental code is uploaded
to Github. | cs.CL | NoisyICL: A Little Noise in Model Parameters Calibrates In-context Learning |
2024-02-08T09:44:02Z | http://arxiv.org/pdf/2402.05512v1 | http://arxiv.org/abs/2402.05512v1 | Juhwan Choi, Eunju Lee, Kyohoon Jin, YoungBin Kim | 2024-02-08T09:44:02Z | cs.CL, cs.AI | Data annotation is an essential step for constructing new datasets. However,
the conventional approach of data annotation through crowdsourcing is both
time-consuming and expensive. In addition, the complexity of this process
increases when dealing with low-resource languages owing to the difference in
the language pool of crowdworkers. To address these issues, this study proposes
an autonomous annotation method by utilizing large language models, which have
been recently demonstrated to exhibit remarkable performance. Through our
experiments, we demonstrate that the proposed method is not just cost-efficient
but also applicable for low-resource language annotation. Additionally, we
constructed an image captioning dataset using our approach and are committed to
open this dataset for future study. We have opened our source code for further
study and reproducibility. | cs.CL | GPTs Are Multilingual Annotators for Sequence Generation Tasks |
2024-02-08T07:56:49Z | http://arxiv.org/pdf/2402.05467v1 | http://arxiv.org/abs/2402.05467v1 | Guangyu Shen, Siyuan Cheng, Kaiyuan Zhang, Guanhong Tao, Shengwei An, Lu Yan, Zhuo Zhang, Shiqing Ma, Xiangyu Zhang | 2024-02-08T07:56:49Z | cs.AI, cs.CL, cs.CR | Large Language Models (LLMs) have become prevalent across diverse sectors,
transforming human life with their extraordinary reasoning and comprehension
abilities. As they find increased use in sensitive tasks, safety concerns have
gained widespread attention. Extensive efforts have been dedicated to aligning
LLMs with human moral principles to ensure their safe deployment. Despite their
potential, recent research indicates aligned LLMs are prone to specialized
jailbreaking prompts that bypass safety measures to elicit violent and harmful
content. The intrinsic discrete nature and substantial scale of contemporary
LLMs pose significant challenges in automatically generating diverse,
efficient, and potent jailbreaking prompts, representing a continuous obstacle.
In this paper, we introduce RIPPLE (Rapid Optimization via Subconscious
Exploitation and Echopraxia), a novel optimization-based method inspired by two
psychological concepts: subconsciousness and echopraxia, which describe the
processes of the mind that occur without conscious awareness and the
involuntary mimicry of actions, respectively. Evaluations across 6 open-source
LLMs and 4 commercial LLM APIs show RIPPLE achieves an average Attack Success
Rate of 91.5\%, outperforming five current methods by up to 47.0\% with an 8x
reduction in overhead. Furthermore, it displays significant transferability and
stealth, successfully evading established detection mechanisms. The code of our
work is available at
\url{https://github.com/SolidShen/RIPPLE_official/tree/official} | cs.AI | Rapid Optimization for Jailbreaking LLMs via Subconscious Exploitation and Echopraxia |
2024-02-08T07:21:45Z | http://arxiv.org/pdf/2402.05457v1 | http://arxiv.org/abs/2402.05457v1 | Chen Chen, Ruizhe Li, Yuchen Hu, Sabato Marco Siniscalchi, Pin-Yu Chen, Ensiong Chng, Chao-Han Huck Yang | 2024-02-08T07:21:45Z | cs.CL, cs.AI, cs.MM, cs.SD, eess.AS | Recent studies have successfully shown that large language models (LLMs) can
be successfully used for generative error correction (GER) on top of the
automatic speech recognition (ASR) output. Specifically, an LLM is utilized to
carry out a direct mapping from the N-best hypotheses list generated by an ASR
system to the predicted output transcription. However, despite its
effectiveness, GER introduces extra data uncertainty since the LLM is trained
without taking into account acoustic information available in the speech
signal. In this work, we aim to overcome such a limitation by infusing acoustic
information before generating the predicted transcription through a novel late
fusion solution termed Uncertainty-Aware Dynamic Fusion (UADF). UADF is a
multimodal fusion approach implemented into an auto-regressive decoding process
and works in two stages: (i) It first analyzes and calibrates the token-level
LLM decision, and (ii) it then dynamically assimilates the information from the
acoustic modality. Experimental evidence collected from various ASR tasks shows
that UADF surpasses existing fusion mechanisms in several ways. It yields
significant improvements in word error rate (WER) while mitigating data
uncertainty issues in LLM and addressing the poor generalization relied with
sole modality during fusion. We also demonstrate that UADF seamlessly adapts to
audio-visual speech recognition. | cs.CL | It's Never Too Late: Fusing Acoustic Information into Large Language Models for Automatic Speech Recognition |
2024-02-08T07:20:02Z | http://arxiv.org/pdf/2402.05455v1 | http://arxiv.org/abs/2402.05455v1 | Samuel Joseph Amouyal, Aya Meltzer-Asscher, Jonathan Berant | 2024-02-08T07:20:02Z | cs.CL | In psycholinguistics, the creation of controlled materials is crucial to
ensure that research outcomes are solely attributed to the intended
manipulations and not influenced by extraneous factors. To achieve this,
psycholinguists typically pretest linguistic materials, where a common pretest
is to solicit plausibility judgments from human evaluators on specific
sentences. In this work, we investigate whether Language Models (LMs) can be
used to generate these plausibility judgements. We investigate a wide range of
LMs across multiple linguistic structures and evaluate whether their
plausibility judgements correlate with human judgements. We find that GPT-4
plausibility judgements highly correlate with human judgements across the
structures we examine, whereas other LMs correlate well with humans on commonly
used syntactic structures. We then test whether this correlation implies that
LMs can be used instead of humans for pretesting. We find that when
coarse-grained plausibility judgements are needed, this works well, but when
fine-grained judgements are necessary, even GPT-4 does not provide satisfactory
discriminative power. | cs.CL | Large Language Models for Psycholinguistic Plausibility Pretesting |
2024-02-08T06:53:31Z | http://arxiv.org/pdf/2402.05445v2 | http://arxiv.org/abs/2402.05445v2 | Haotong Qin, Xudong Ma, Xingyu Zheng, Xiaoyang Li, Yang Zhang, Shouda Liu, Jie Luo, Xianglong Liu, Michele Magno | 2024-05-27T09:20:35Z | cs.LG, cs.CL | The LoRA-finetuning quantization of LLMs has been extensively studied to
obtain accurate yet compact LLMs for deployment on resource-constrained
hardware. However, existing methods cause the quantized LLM to severely degrade
and even fail to benefit from the finetuning of LoRA. This paper proposes a
novel IR-QLoRA for pushing quantized LLMs with LoRA to be highly accurate
through information retention. The proposed IR-QLoRA mainly relies on two
technologies derived from the perspective of unified information: (1)
statistics-based Information Calibration Quantization allows the quantized
parameters of LLM to retain original information accurately; (2)
finetuning-based Information Elastic Connection makes LoRA utilizes elastic
representation transformation with diverse information. Comprehensive
experiments show that IR-QLoRA can significantly improve accuracy across LLaMA
and LLaMA2 families under 2-4 bit-widths, e.g., 4- bit LLaMA-7B achieves 1.4%
improvement on MMLU compared with the state-of-the-art methods. The significant
performance gain requires only a tiny 0.31% additional time consumption,
revealing the satisfactory efficiency of our IR-QLoRA. We highlight that
IR-QLoRA enjoys excellent versatility, compatible with various frameworks
(e.g., NormalFloat and Integer quantization) and brings general accuracy gains.
The code is available at https://github.com/htqin/ir-qlora. | cs.LG | Accurate LoRA-Finetuning Quantization of LLMs via Information Retention |
2024-02-08T06:34:11Z | http://arxiv.org/pdf/2402.05440v1 | http://arxiv.org/abs/2402.05440v1 | Jack Zhang | 2024-02-08T06:34:11Z | cs.CL | Enhancing AI systems with efficient communication skills for effective human
assistance necessitates proactive initiatives from the system side to discern
specific circumstances and interact aptly. This research focuses on a
collective building assignment in the Minecraft dataset, employing language
modeling to enhance task understanding through state-of-the-art methods. These
models focus on grounding multi-modal understanding and task-oriented dialogue
comprehension tasks, providing insights into their interpretative and
responsive capabilities. Our experimental results showcase a substantial
improvement over existing methods, indicating a promising direction for future
research in this domain. | cs.CL | Improving Agent Interactions in Virtual Environments with Language Models |
2024-02-08T06:20:01Z | http://arxiv.org/pdf/2402.05435v1 | http://arxiv.org/abs/2402.05435v1 | Christopher J. Lynch, Erik Jensen, Madison H. Munro, Virginia Zamponi, Joseph Martinez, Kevin O'Brien, Brandon Feldhaus, Katherine Smith, Ann Marie Reinhold, Ross Gore | 2024-02-08T06:20:01Z | cs.CL, cs.AI, cs.LG, I.2.7; I.6.4 | Large Language Models (LLMs) play a pivotal role in generating vast arrays of
narratives, facilitating a systematic exploration of their effectiveness for
communicating life events in narrative form. In this study, we employ a
zero-shot structured narrative prompt to generate 24,000 narratives using
OpenAI's GPT-4. From this dataset, we manually classify 2,880 narratives and
evaluate their validity in conveying birth, death, hiring, and firing events.
Remarkably, 87.43% of the narratives sufficiently convey the intention of the
structured prompt. To automate the identification of valid and invalid
narratives, we train and validate nine Machine Learning models on the
classified datasets. Leveraging these models, we extend our analysis to predict
the classifications of the remaining 21,120 narratives. All the ML models
excelled at classifying valid narratives as valid, but experienced challenges
at simultaneously classifying invalid narratives as invalid. Our findings not
only advance the study of LLM capabilities, limitations, and validity but also
offer practical insights for narrative generation and natural language
processing applications. | cs.CL | GPT-4 Generated Narratives of Life Events using a Structured Narrative Prompt: A Validation Study |
2024-02-08T04:48:26Z | http://arxiv.org/pdf/2402.05406v2 | http://arxiv.org/abs/2402.05406v2 | Lucio Dery, Steven Kolawole, Jean-François Kagy, Virginia Smith, Graham Neubig, Ameet Talwalkar | 2024-02-09T19:53:56Z | cs.LG, cs.CL | Given the generational gap in available hardware between lay practitioners
and the most endowed institutions, LLMs are becoming increasingly inaccessible
as they grow in size. Whilst many approaches have been proposed to compress
LLMs to make their resource consumption manageable, these methods themselves
tend to be resource intensive, putting them out of the reach of the very user
groups they target. In this work, we explore the problem of structured pruning
of LLMs using only forward passes. We seek to empower practitioners to prune
models so large that their available hardware has just enough memory to run
inference. We develop Bonsai, a gradient-free, perturbative pruning method
capable of delivering small, fast, and accurate pruned models.
We observe that Bonsai outputs pruned models that (i) outperform those
generated by more expensive gradient-based structured pruning methods, and (ii)
are twice as fast (with comparable accuracy) as those generated by
semi-structured pruning methods requiring comparable resources as Bonsai. We
also leverage Bonsai to produce a new sub-2B model using a single A6000 that
yields state-of-the-art performance on 4/6 tasks on the Huggingface Open LLM
leaderboard. | cs.LG | Everybody Prune Now: Structured Pruning of LLMs with only Forward Passes |
2024-02-08T04:42:29Z | http://arxiv.org/pdf/2402.05403v2 | http://arxiv.org/abs/2402.05403v2 | Tianjun Zhang, Aman Madaan, Luyu Gao, Steven Zheng, Swaroop Mishra, Yiming Yang, Niket Tandon, Uri Alon | 2024-02-09T19:09:52Z | cs.CL, cs.AI | In-context learning (ICL, also known as few-shot prompting) has been the
standard method of adapting LLMs to downstream tasks, by learning from a few
input-output examples. Nonetheless, all ICL-based approaches only learn from
correct input-output pairs. In this paper, we revisit this paradigm, by
learning more from the few given input-output examples. We introduce Learning
Principles (LEAP): First, we intentionally induce the model to make mistakes on
these few examples; then we reflect on these mistakes, and learn explicit
task-specific "principles" from them, which help solve similar problems and
avoid common mistakes; finally, we prompt the model to answer unseen test
questions using the original few-shot examples and these learned general
principles. We evaluate LEAP on a wide range of benchmarks, including multi-hop
question answering (Hotpot QA), textual QA (DROP), Big-Bench Hard reasoning,
and math problems (GSM8K and MATH); in all these benchmarks, LEAP improves the
strongest available LLMs such as GPT-3.5-turbo, GPT-4, GPT-4 turbo and
Claude-2.1. For example, LEAP improves over the standard few-shot prompting
using GPT-4 by 7.5% in DROP, and by 3.3% in HotpotQA. Importantly, LEAP does
not require any more input or examples than the standard few-shot prompting
settings. | cs.CL | In-Context Principle Learning from Mistakes |
2024-02-08T03:41:39Z | http://arxiv.org/pdf/2402.10941v1 | http://arxiv.org/abs/2402.10941v1 | Shiyu Wang, Yihao Feng, Tian Lan, Ning Yu, Yu Bai, Ran Xu, Huan Wang, Caiming Xiong, Silvio Savarese | 2024-02-08T03:41:39Z | cs.CL, cs.AI, cs.LG | Natural language serves as a common and straightforward control signal for
humans to interact seamlessly with machines. Recognizing the importance of this
interface, the machine learning community is investing considerable effort in
generating data that is semantically coherent with textual instructions. While
strides have been made in text-to-data generation spanning image editing, audio
synthesis, video creation, and beyond, low-resource areas characterized by
expensive annotations or complex data structures, such as molecules, motion
dynamics, and time series, often lack textual labels. This deficiency impedes
supervised learning, thereby constraining the application of advanced
generative models for text-to-data tasks. In response to these challenges in
the low-resource scenario, we propose Text2Data, a novel approach that utilizes
unlabeled data to understand the underlying data distribution through an
unsupervised diffusion model. Subsequently, it undergoes controllable
finetuning via a novel constraint optimization-based learning objective that
ensures controllability and effectively counteracts catastrophic forgetting.
Comprehensive experiments demonstrate that Text2Data is able to achieve
enhanced performance regarding controllability across various modalities,
including molecules, motions and time series, when compared to existing
baselines. | cs.CL | Text2Data: Low-Resource Data Generation with Textual Control |
2024-02-08T03:17:38Z | http://arxiv.org/pdf/2402.05376v1 | http://arxiv.org/abs/2402.05376v1 | Feihu Jin, Yifan Liu, Ying Tan | 2024-02-08T03:17:38Z | cs.CL | Large Language Models (LLMs) have demonstrated remarkable performance across
diverse tasks and exhibited impressive reasoning abilities by applying
zero-shot Chain-of-Thought (CoT) prompting. However, due to the evolving nature
of sentence prefixes during the pre-training phase, existing zero-shot CoT
prompting methods that employ identical CoT prompting across all task instances
may not be optimal. In this paper, we introduce a novel zero-shot prompting
method that leverages evolutionary algorithms to generate diverse promptings
for LLMs dynamically. Our approach involves initializing two CoT promptings,
performing evolutionary operations based on LLMs to create a varied set, and
utilizing the LLMs to select a suitable CoT prompting for a given problem.
Additionally, a rewriting operation, guided by the selected CoT prompting,
enhances the understanding of the LLMs about the problem. Extensive experiments
conducted across ten reasoning datasets demonstrate the superior performance of
our proposed method compared to current zero-shot CoT prompting methods on
GPT-3.5-turbo and GPT-4. Moreover, in-depth analytical experiments underscore
the adaptability and effectiveness of our method in various reasoning tasks. | cs.CL | Zero-Shot Chain-of-Thought Reasoning Guided by Evolutionary Algorithms in Large Language Models |
2024-02-08T03:12:25Z | http://arxiv.org/pdf/2402.05374v2 | http://arxiv.org/abs/2402.05374v2 | Youngsik Yun, Jihie Kim | 2024-05-02T02:41:50Z | cs.CV, cs.AI, cs.CL | Image Captioning generates descriptive sentences from images using
Vision-Language Pre-trained models (VLPs) such as BLIP, which has improved
greatly. However, current methods lack the generation of detailed descriptive
captions for the cultural elements depicted in the images, such as the
traditional clothing worn by people from Asian cultural groups. In this paper,
we propose a new framework, \textbf{Culturally-aware Image Captioning (CIC)},
that generates captions and describes cultural elements extracted from cultural
visual elements in images representing cultures. Inspired by methods combining
visual modality and Large Language Models (LLMs) through appropriate prompts,
our framework (1) generates questions based on cultural categories from images,
(2) extracts cultural visual elements from Visual Question Answering (VQA)
using generated questions, and (3) generates culturally-aware captions using
LLMs with the prompts. Our human evaluation conducted on 45 participants from 4
different cultural groups with a high understanding of the corresponding
culture shows that our proposed framework generates more culturally descriptive
captions when compared to the image captioning baseline based on VLPs. Our code
and dataset will be made publicly available upon acceptance. | cs.CV | CIC: A framework for Culturally-aware Image Captioning |
2024-02-08T02:58:47Z | http://arxiv.org/pdf/2402.05369v2 | http://arxiv.org/abs/2402.05369v2 | Huayu Chen, Guande He, Lifan Yuan, Ganqu Cui, Hang Su, Jun Zhu | 2024-07-03T13:53:06Z | cs.LG, cs.CL | User intentions are typically formalized as evaluation rewards to be
maximized when fine-tuning language models (LMs). Existing alignment methods,
such as Direct Preference Optimization (DPO), are mainly tailored for pairwise
preference data where rewards are implicitly defined rather than explicitly
given. In this paper, we introduce a general framework for LM alignment,
leveraging Noise Contrastive Estimation (NCE) to bridge the gap in handling
reward datasets explicitly annotated with scalar evaluations. Our framework
comprises two parallel algorithms, NCA and InfoNCA, both enabling the direct
extraction of an LM policy from reward data as well as preference data.
Notably, we show that the DPO loss is a special case of our proposed InfoNCA
objective under pairwise preference settings, thereby integrating and extending
current alignment theories. By comparing NCA and InfoNCA, we demonstrate that
the well-observed decreasing-likelihood trend of DPO/InfoNCA is caused by their
focus on adjusting relative likelihood across different responses. In contrast,
NCA optimizes the absolute likelihood for each response, thereby effectively
preventing the chosen likelihood from decreasing. We evaluate our methods in
both reward and preference settings with Mistral-8*7B and 7B models.
Experiments suggest that InfoNCA/NCA surpasses various preference baselines
when reward datasets are available. We also find NCA significantly outperforms
DPO in complex reasoning tasks like math and coding. | cs.LG | Noise Contrastive Alignment of Language Models with Explicit Rewards |
2024-02-08T02:37:30Z | http://arxiv.org/pdf/2402.05359v6 | http://arxiv.org/abs/2402.05359v6 | Yizhou Zhang, Lun Du, Defu Cao, Qiang Fu, Yan Liu | 2024-07-02T18:18:18Z | cs.AI, cs.CL, cs.LG | Foundation models, such as Large language Models (LLMs), have attracted
significant amount of interest due to their large number of applications.
However, when handling tasks involving repetitive sub-tasks and/or deceptive
contents, such as arithmetic calculation and article-level fake news detection,
simple instructional prompts suffer from inaccurate responses. Existing works
show that more complicated prompting strategies, such as Chain-of-Thoughts and
Least-to-Most, can unlock LLM's powerful capacity in diverse areas. Recent
researches reveal that simple divide-and-conquer prompting strategy, i.e.
simply dividing the input sequence to multiple sub-inputs, can also
substantially improve LLM's performance in some specific tasks such as
misinformation detection. In this paper, we aim at examining the utility of
divide-and-conquer prompting strategy and answer on which kind of tasks this
strategy gets advantages. Specifically, we provide a theoretic analysis to
divide-and-conquer prompting strategy and help us identify the specific tasks
where DaC prompting can bring performance boost with theoretic guarantee. We
then present two cases (large integer arithmetic and fact verification) where
experimental results aligns with our theoretic analysis. | cs.AI | An Examination on the Effectiveness of Divide-and-Conquer Prompting in Large Language Models |
2024-02-07T23:47:02Z | http://arxiv.org/pdf/2402.14590v1 | http://arxiv.org/abs/2402.14590v1 | Wei Qiao, Tushar Dogra, Otilia Stretcu, Yu-Han Lyu, Tiantian Fang, Dongjin Kwon, Chun-Ta Lu, Enming Luo, Yuan Wang, Chih-Chun Chia, Ariel Fuxman, Fangzhou Wang, Ranjay Krishna, Mehmet Tek | 2024-02-07T23:47:02Z | cs.IR, cs.CL, cs.LG | Large language models (LLMs) are powerful tools for content moderation, but
their inference costs and latency make them prohibitive for casual use on large
datasets, such as the Google Ads repository. This study proposes a method for
scaling up LLM reviews for content moderation in Google Ads. First, we use
heuristics to select candidates via filtering and duplicate removal, and create
clusters of ads for which we select one representative ad per cluster. We then
use LLMs to review only the representative ads. Finally, we propagate the LLM
decisions for the representative ads back to their clusters. This method
reduces the number of reviews by more than 3 orders of magnitude while
achieving a 2x recall compared to a baseline non-LLM model. The success of this
approach is a strong function of the representations used in clustering and
label propagation; we found that cross-modal similarity representations yield
better results than uni-modal representations. | cs.IR | Scaling Up LLM Reviews for Google Ads Content Moderation |
2024-02-07T23:39:40Z | http://arxiv.org/pdf/2402.05318v1 | http://arxiv.org/abs/2402.05318v1 | Dipankar Sarkar | 2024-02-07T23:39:40Z | cs.IR, cs.CL, cs.LG | Information retrieval is a rapidly evolving field of information retrieval,
which is characterized by a continuous refinement of techniques and
technologies, from basic hyperlink-based navigation to sophisticated
algorithm-driven search engines. This paper aims to provide a comprehensive
overview of the evolution of Information Retrieval Technology, with a
particular focus on the role of Large Language Models (LLMs) in bridging the
gap between traditional search methods and the emerging paradigm of answer
retrieval. The integration of LLMs in the realms of response retrieval and
indexing signifies a paradigm shift in how users interact with information
systems. This paradigm shift is driven by the integration of large language
models (LLMs) like GPT-4, which are capable of understanding and generating
human-like text, thus enabling them to provide more direct and contextually
relevant answers to user queries. Through this exploration, we seek to
illuminate the technological milestones that have shaped this journey and the
potential future directions in this rapidly changing field. | cs.IR | Navigating the Knowledge Sea: Planet-scale answer retrieval using LLMs |
2024-02-07T22:16:53Z | http://arxiv.org/pdf/2402.05294v1 | http://arxiv.org/abs/2402.05294v1 | Pramit Saha, Divyanshu Mishra, Felix Wagner, Konstantinos Kamnitsas, J. Alison Noble | 2024-02-07T22:16:53Z | cs.LG, cs.AI, cs.CL, cs.CV | Multimodal Federated Learning (MMFL) utilizes multiple modalities in each
client to build a more powerful Federated Learning (FL) model than its unimodal
counterpart. However, the impact of missing modality in different clients, also
called modality incongruity, has been greatly overlooked. This paper, for the
first time, analyses the impact of modality incongruity and reveals its
connection with data heterogeneity across participating clients. We
particularly inspect whether incongruent MMFL with unimodal and multimodal
clients is more beneficial than unimodal FL. Furthermore, we examine three
potential routes of addressing this issue. Firstly, we study the effectiveness
of various self-attention mechanisms towards incongruity-agnostic information
fusion in MMFL. Secondly, we introduce a modality imputation network (MIN)
pre-trained in a multimodal client for modality translation in unimodal clients
and investigate its potential towards mitigating the missing modality problem.
Thirdly, we assess the capability of client-level and server-level
regularization techniques towards mitigating modality incongruity effects.
Experiments are conducted under several MMFL settings on two publicly available
real-world datasets, MIMIC-CXR and Open-I, with Chest X-Ray and radiology
reports. | cs.LG | Examining Modality Incongruity in Multimodal Federated Learning for Medical Vision and Language-based Disease Detection |
2024-02-07T21:54:53Z | http://arxiv.org/pdf/2402.05282v1 | http://arxiv.org/abs/2402.05282v1 | Ran Zmigrod, Zhiqiang Ma, Armineh Nourbakhsh, Sameena Shah | 2024-02-07T21:54:53Z | cs.CL | Visually Rich Form Understanding (VRFU) poses a complex research problem due
to the documents' highly structured nature and yet highly variable style and
content. Current annotation schemes decompose form understanding and omit key
hierarchical structure, making development and evaluation of end-to-end models
difficult. In this paper, we propose a novel F1 metric to evaluate form parsers
and describe a new content-agnostic, tree-based annotation scheme for VRFU:
TreeForm. We provide methods to convert previous annotation schemes into
TreeForm structures and evaluate TreeForm predictions using a modified version
of the normalized tree-edit distance. We present initial baselines for our
end-to-end performance metric and the TreeForm edit distance, averaged over the
FUNSD and XFUND datasets, of 61.5 and 26.4 respectively. We hope that TreeForm
encourages deeper research in annotating, modeling, and evaluating the
complexities of form-like documents. | cs.CL | TreeForm: End-to-end Annotation and Evaluation for Form Document Parsing |
2024-02-07T21:08:49Z | http://arxiv.org/pdf/2402.07938v2 | http://arxiv.org/abs/2402.07938v2 | Syed Mekael Wasti, Ken Q. Pu, Ali Neshati | 2024-04-16T07:39:05Z | cs.HC, cs.AI, cs.CL, cs.LG, I.2.7; I.2.1 | The evolution of Large Language Models (LLMs) has showcased remarkable
capacities for logical reasoning and natural language comprehension. These
capabilities can be leveraged in solutions that semantically and textually
model complex problems. In this paper, we present our efforts toward
constructing a framework that can serve as an intermediary between a user and
their user interface (UI), enabling dynamic and real-time interactions. We
employ a system that stands upon textual semantic mappings of UI components, in
the form of annotations. These mappings are stored, parsed, and scaled in a
custom data structure, supplementary to an agent-based prompting backend
engine. Employing textual semantic mappings allows each component to not only
explain its role to the engine but also provide expectations. By comprehending
the needs of both the user and the components, our LLM engine can classify the
most appropriate application, extract relevant parameters, and subsequently
execute precise predictions of the user's expected actions. Such an integration
evolves static user interfaces into highly dynamic and adaptable solutions,
introducing a new frontier of intelligent and responsive user experiences. | cs.HC | Large Language User Interfaces: Voice Interactive User Interfaces powered by LLMs |
2024-02-07T20:11:56Z | http://arxiv.org/pdf/2402.10940v1 | http://arxiv.org/abs/2402.10940v1 | Pei-Hung Chung, Shuhan He, Norawit Kijpaisalratana, Abdel-badih el Ariss, Byung-Jun Yoon | 2024-02-07T20:11:56Z | cs.CL, cs.AI, cs.LG | A Clinical Decision Support System (CDSS) is designed to enhance clinician
decision-making by combining system-generated recommendations with medical
expertise. Given the high costs, intensive labor, and time-sensitive nature of
medical treatments, there is a pressing need for efficient decision support,
especially in complex emergency scenarios. In these scenarios, where
information can be limited, an advanced CDSS framework that leverages AI
(artificial intelligence) models to effectively reduce diagnostic uncertainty
has utility. Such an AI-enabled CDSS framework with quantified uncertainty
promises to be practical and beneficial in the demanding context of real-world
medical care. In this study, we introduce the concept of Medical Entropy,
quantifying uncertainties in patient outcomes predicted by neural machine
translation based on the ICD-9 code of procedures. Our experimental results not
only show strong correlations between procedure and diagnosis sequences based
on the simple ICD-9 code but also demonstrate the promising capacity to model
trends of uncertainties during hospitalizations through a data-driven approach. | cs.CL | Neural machine translation of clinical procedure codes for medical diagnosis and uncertainty quantification |
2024-02-07T20:02:09Z | http://arxiv.org/pdf/2402.05224v2 | http://arxiv.org/abs/2402.05224v2 | Berk Atil, Mahsa Sheikhi Karizaki, Rebecca J. Passonneau | 2024-04-25T16:16:36Z | cs.CL, cs.AI, cs.LG | With an increasing focus in STEM education on critical thinking skills,
science writing plays an ever more important role in curricula that stress
inquiry skills. A recently published dataset of two sets of college level lab
reports from an inquiry-based physics curriculum relies on analytic assessment
rubrics that utilize multiple dimensions, specifying subject matter knowledge
and general components of good explanations. Each analytic dimension is
assessed on a 6-point scale, to provide detailed feedback to students that can
help them improve their science writing skills. Manual assessment can be slow,
and difficult to calibrate for consistency across all students in large
classes. While much work exists on automated assessment of open-ended questions
in STEM subjects, there has been far less work on long-form writing such as lab
reports. We present an end-to-end neural architecture that has separate
verifier and assessment modules, inspired by approaches to Open Domain Question
Answering (OpenQA). VerAs first verifies whether a report contains any content
relevant to a given rubric dimension, and if so, assesses the relevant
sentences. On the lab reports, VerAs outperforms multiple baselines based on
OpenQA systems or Automated Essay Scoring (AES). VerAs also performs well on an
analytic rubric for middle school physics essays. | cs.CL | VerAs: Verify then Assess STEM Lab Reports |
2024-02-07T19:11:23Z | http://arxiv.org/pdf/2402.05201v2 | http://arxiv.org/abs/2402.05201v2 | Matthew Renze, Erhan Guven | 2024-06-14T18:41:51Z | cs.CL, cs.AI | In this research study, we empirically investigate the effect of sampling
temperature on the performance of Large Language Models (LLMs) on various
problem-solving tasks. We created a multiple-choice question-and-answer (MCQA)
exam by randomly sampling problems from standard LLM benchmarks. Then, we used
nine popular LLMs with five prompt-engineering techniques to solve the MCQA
problems while increasing the sampling temperature from 0.0 to 1.6. Despite
anecdotal reports to the contrary, our empirical results indicate that changes
in temperature from 0.0 to 1.0 do not have a statistically significant impact
on LLM performance for problem-solving tasks. In addition, these results appear
to generalize across LLMs, prompt-engineering techniques, and problem domains.
All code, data, and supplemental materials are available on GitHub at:
https://github.com/matthewrenze/jhu-llm-temperature | cs.CL | The Effect of Sampling Temperature on Problem Solving in Large Language Models |
2024-02-07T19:10:36Z | http://arxiv.org/pdf/2402.05200v1 | http://arxiv.org/abs/2402.05200v1 | Santiago Miret, N M Anoop Krishnan | 2024-02-07T19:10:36Z | cond-mat.mtrl-sci, cs.AI, cs.CL, cs.LG | Large Language Models (LLMs) create exciting possibilities for powerful
language processing tools to accelerate research in materials science. While
LLMs have great potential to accelerate materials understanding and discovery,
they currently fall short in being practical materials science tools. In this
position paper, we show relevant failure cases of LLMs in materials science
that reveal current limitations of LLMs related to comprehending and reasoning
over complex, interconnected materials science knowledge. Given those
shortcomings, we outline a framework for developing Materials Science LLMs
(MatSci-LLMs) that are grounded in materials science knowledge and hypothesis
generation followed by hypothesis testing. The path to attaining performant
MatSci-LLMs rests in large part on building high-quality, multi-modal datasets
sourced from scientific literature where various information extraction
challenges persist. As such, we describe key materials science information
extraction challenges which need to be overcome in order to build large-scale,
multi-modal datasets that capture valuable materials science knowledge.
Finally, we outline a roadmap for applying future MatSci-LLMs for real-world
materials discovery via: 1. Automated Knowledge Base Generation; 2. Automated
In-Silico Material Design; and 3. MatSci-LLM Integrated Self-Driving Materials
Laboratories. | cond-mat.mtrl-sci | Are LLMs Ready for Real-World Materials Discovery? |
2024-02-07T19:07:10Z | http://arxiv.org/pdf/2402.05195v2 | http://arxiv.org/abs/2402.05195v2 | Maitreya Patel, Sangmin Jung, Chitta Baral, Yezhou Yang | 2024-04-09T22:14:37Z | cs.CV, cs.CL | Despite the recent advances in personalized text-to-image (P-T2I) generative
models, it remains challenging to perform finetuning-free multi-subject-driven
T2I in a resource-efficient manner. Predominantly, contemporary approaches,
involving the training of Hypernetworks and Multimodal Large Language Models
(MLLMs), require heavy computing resources that range from 600 to 12300 GPU
hours of training. These subject-driven T2I methods hinge on Latent Diffusion
Models (LDMs), which facilitate T2I mapping through cross-attention layers.
While LDMs offer distinct advantages, P-T2I methods' reliance on the latent
space of these diffusion models significantly escalates resource demands,
leading to inconsistent results and necessitating numerous iterations for a
single desired image. In this paper, we present $\lambda$-ECLIPSE, an
alternative prior-training strategy that works in the latent space of a
pre-trained CLIP model without relying on the diffusion UNet models.
$\lambda$-ECLIPSE leverages the image-text interleaved pre-training for fast
and effective multi-subject-driven P-T2I. Through extensive experiments, we
establish that $\lambda$-ECLIPSE surpasses existing baselines in composition
alignment while preserving concept alignment performance, even with
significantly lower resource utilization. $\lambda$-ECLIPSE performs
multi-subject driven P-T2I with just 34M parameters and is trained on a mere 74
GPU hours. Additionally, $\lambda$-ECLIPSE demonstrates the unique ability to
perform multi-concept interpolations. | cs.CV | $λ$-ECLIPSE: Multi-Concept Personalized Text-to-Image Diffusion Models by Leveraging CLIP Latent Space |
2024-02-07T19:01:11Z | http://arxiv.org/pdf/2402.05188v1 | http://arxiv.org/abs/2402.05188v1 | Jiaqiang Ye Zhu, Carla Gomez Cano, David Vazquez Bermudez, Michal Drozdzal | 2024-02-07T19:01:11Z | cs.RO, cs.AI, cs.CL | One of the challenges in robotics is to enable robotic units with the
reasoning capability that would be robust enough to execute complex tasks in
dynamic environments. Recent advances in LLMs have positioned them as go-to
tools for simple reasoning tasks, motivating the pioneering work of Liang et
al. [35] that uses an LLM to translate natural language commands into low-level
static execution plans for robotic units. Using LLMs inside robotics systems
brings their generalization to a new level, enabling zero-shot generalization
to new tasks. This paper extends this prior work to dynamic environments. We
propose InCoRo, a system that uses a classical robotic feedback loop composed
of an LLM controller, a scene understanding unit, and a robot. Our system
continuously analyzes the state of the environment and provides adapted
execution commands, enabling the robot to adjust to changing environmental
conditions and correcting for controller errors. Our system does not require
any iterative optimization to learn to accomplish a task as it leverages
in-context learning with an off-the-shelf LLM model. Through an extensive
validation process involving two standardized industrial robotic units -- SCARA
and DELTA types -- we contribute knowledge about these robots, not popular in
the community, thereby enriching it. We highlight the generalization
capabilities of our system and show that (1) in-context learning in combination
with the current state-of-the-art LLMs is an effective way to implement a
robotic controller; (2) in static environments, InCoRo surpasses the prior art
in terms of the success rate; (3) in dynamic environments, we establish new
state-of-the-art for the SCARA and DELTA units, respectively. This research
paves the way towards building reliable, efficient, intelligent autonomous
systems that adapt to dynamic environments. | cs.RO | InCoRo: In-Context Learning for Robotics Control with Feedback Loops |
2024-02-07T18:59:31Z | http://arxiv.org/pdf/2402.05111v1 | http://arxiv.org/abs/2402.05111v1 | Rose E. Wang, Dorottya Demszky | 2024-02-07T18:59:31Z | cs.CL, cs.AI | We introduce Edu-ConvoKit, an open-source library designed to handle
pre-processing, annotation and analysis of conversation data in education.
Resources for analyzing education conversation data are scarce, making the
research challenging to perform and therefore hard to access. We address these
challenges with Edu-ConvoKit. Edu-ConvoKit is open-source
(https://github.com/stanfordnlp/edu-convokit ), pip-installable
(https://pypi.org/project/edu-convokit/ ), with comprehensive documentation
(https://edu-convokit.readthedocs.io/en/latest/ ). Our demo video is available
at: https://youtu.be/zdcI839vAko?si=h9qlnl76ucSuXb8- . We include additional
resources, such as Colab applications of Edu-ConvoKit to three diverse
education datasets and a repository of Edu-ConvoKit related papers, that can be
found in our GitHub repository. | cs.CL | Edu-ConvoKit: An Open-Source Library for Education Conversation Data |
2024-02-07T18:57:37Z | http://arxiv.org/pdf/2402.05106v1 | http://arxiv.org/abs/2402.05106v1 | Rafael Silva de Alencar, William Alberto Cruz Castañeda, Marcellus Amadeus | 2024-02-07T18:57:37Z | cs.CV, cs.AI, cs.CL | This work presents the early development of a model of image captioning for
the Brazilian Portuguese language. We used the GRIT (Grid - and Region-based
Image captioning Transformer) model to accomplish this work. GRIT is a
Transformer-only neural architecture that effectively utilizes two visual
features to generate better captions. The GRIT method emerged as a proposal to
be a more efficient way to generate image captioning. In this work, we adapt
the GRIT model to be trained in a Brazilian Portuguese dataset to have an image
captioning method for the Brazilian Portuguese Language. | cs.CV | Image captioning for Brazilian Portuguese using GRIT model |
2024-02-07T18:34:38Z | http://arxiv.org/pdf/2402.05162v3 | http://arxiv.org/abs/2402.05162v3 | Boyi Wei, Kaixuan Huang, Yangsibo Huang, Tinghao Xie, Xiangyu Qi, Mengzhou Xia, Prateek Mittal, Mengdi Wang, Peter Henderson | 2024-07-01T07:11:17Z | cs.LG, cs.AI, cs.CL | Large language models (LLMs) show inherent brittleness in their safety
mechanisms, as evidenced by their susceptibility to jailbreaking and even
non-malicious fine-tuning. This study explores this brittleness of safety
alignment by leveraging pruning and low-rank modifications. We develop methods
to identify critical regions that are vital for safety guardrails, and that are
disentangled from utility-relevant regions at both the neuron and rank levels.
Surprisingly, the isolated regions we find are sparse, comprising about $3\%$
at the parameter level and $2.5\%$ at the rank level. Removing these regions
compromises safety without significantly impacting utility, corroborating the
inherent brittleness of the model's safety mechanisms. Moreover, we show that
LLMs remain vulnerable to low-cost fine-tuning attacks even when modifications
to the safety-critical regions are restricted. These findings underscore the
urgent need for more robust safety strategies in LLMs. | cs.LG | Assessing the Brittleness of Safety Alignment via Pruning and Low-Rank Modifications |
2024-02-07T18:21:17Z | http://arxiv.org/pdf/2402.05070v1 | http://arxiv.org/abs/2402.05070v1 | Taylor Sorensen, Jared Moore, Jillian Fisher, Mitchell Gordon, Niloofar Mireshghallah, Christopher Michael Rytting, Andre Ye, Liwei Jiang, Ximing Lu, Nouha Dziri, Tim Althoff, Yejin Choi | 2024-02-07T18:21:17Z | cs.AI, cs.CL, cs.IR | With increased power and prevalence of AI systems, it is ever more critical
that AI systems are designed to serve all, i.e., people with diverse values and
perspectives. However, aligning models to serve pluralistic human values
remains an open research question. In this piece, we propose a roadmap to
pluralistic alignment, specifically using language models as a test bed. We
identify and formalize three possible ways to define and operationalize
pluralism in AI systems: 1) Overton pluralistic models that present a spectrum
of reasonable responses; 2) Steerably pluralistic models that can steer to
reflect certain perspectives; and 3) Distributionally pluralistic models that
are well-calibrated to a given population in distribution. We also propose and
formalize three possible classes of pluralistic benchmarks: 1) Multi-objective
benchmarks, 2) Trade-off steerable benchmarks, which incentivize models to
steer to arbitrary trade-offs, and 3) Jury-pluralistic benchmarks which
explicitly model diverse human ratings. We use this framework to argue that
current alignment techniques may be fundamentally limited for pluralistic AI;
indeed, we highlight empirical evidence, both from our own experiments and from
other work, that standard alignment procedures might reduce distributional
pluralism in models, motivating the need for further research on pluralistic
alignment. | cs.AI | A Roadmap to Pluralistic Alignment |
2024-02-07T17:33:54Z | http://arxiv.org/pdf/2402.05044v4 | http://arxiv.org/abs/2402.05044v4 | Lijun Li, Bowen Dong, Ruohui Wang, Xuhao Hu, Wangmeng Zuo, Dahua Lin, Yu Qiao, Jing Shao | 2024-06-07T12:05:46Z | cs.CL, cs.AI, cs.CR, cs.LG | In the rapidly evolving landscape of Large Language Models (LLMs), ensuring
robust safety measures is paramount. To meet this crucial need, we propose
\emph{SALAD-Bench}, a safety benchmark specifically designed for evaluating
LLMs, attack, and defense methods. Distinguished by its breadth, SALAD-Bench
transcends conventional benchmarks through its large scale, rich diversity,
intricate taxonomy spanning three levels, and versatile
functionalities.SALAD-Bench is crafted with a meticulous array of questions,
from standard queries to complex ones enriched with attack, defense
modifications and multiple-choice. To effectively manage the inherent
complexity, we introduce an innovative evaluators: the LLM-based MD-Judge for
QA pairs with a particular focus on attack-enhanced queries, ensuring a
seamless, and reliable evaluation. Above components extend SALAD-Bench from
standard LLM safety evaluation to both LLM attack and defense methods
evaluation, ensuring the joint-purpose utility. Our extensive experiments shed
light on the resilience of LLMs against emerging threats and the efficacy of
contemporary defense tactics. Data and evaluator are released under
https://github.com/OpenSafetyLab/SALAD-BENCH. | cs.CL | SALAD-Bench: A Hierarchical and Comprehensive Safety Benchmark for Large Language Models |
2024-02-07T17:07:53Z | http://arxiv.org/pdf/2402.05034v1 | http://arxiv.org/abs/2402.05034v1 | Miriam Cuscito, Alfio Ferrara, Martin Ruskov | 2024-02-07T17:07:53Z | cs.CL, cs.CY, I.2.7; J.5 | In this paper, we explore the idea of analysing the historical bias of
contextual language models based on BERT by measuring their adequacy with
respect to Early Modern (EME) and Modern (ME) English. In our preliminary
experiments, we perform fill-in-the-blank tests with 60 masked sentences (20
EME-specific, 20 ME-specific and 20 generic) and three different models (i.e.,
BERT Base, MacBERTh, English HLM). We then rate the model predictions according
to a 5-point bipolar scale between the two language varieties and derive a
weighted score to measure the adequacy of each model to EME and ME varieties of
English. | cs.CL | How BERT Speaks Shakespearean English? Evaluating Historical Bias in Contextual Language Models |
2024-02-07T16:15:59Z | http://arxiv.org/pdf/2402.05000v1 | http://arxiv.org/abs/2402.05000v1 | Shashank Sonkar, Kangqi Ni, Sapana Chaudhary, Richard G. Baraniuk | 2024-02-07T16:15:59Z | cs.CL | In this paper, we introduce the novel concept of pedagogically aligned Large
Language Models (LLMs) that signifies a transformative shift in the application
of LLMs within educational contexts. Rather than providing direct responses to
user queries, pedagogically-aligned LLMs function as scaffolding tools,
breaking complex problems into manageable subproblems and guiding students
towards the final answer through constructive feedback and hints. The objective
is to equip learners with problem-solving strategies that deepen their
understanding and internalization of the subject matter. Previous research in
this field has primarily applied the supervised finetuning approach without
framing the objective as an alignment problem, hence not employing
reinforcement learning through human feedback (RLHF) methods. This study
reinterprets the narrative by viewing the task through the lens of alignment
and demonstrates how RLHF methods emerge naturally as a superior alternative
for aligning LLM behaviour. Building on this perspective, we propose a novel
approach for constructing a reward dataset specifically designed for the
pedagogical alignment of LLMs. We apply three state-of-the-art RLHF algorithms
and find that they outperform SFT significantly. Our qualitative analyses
across model differences and hyperparameter sensitivity further validate the
superiority of RLHF over SFT. Also, our study sheds light on the potential of
online feedback for enhancing the performance of pedagogically-aligned LLMs,
thus providing valuable insights for the advancement of these models in
educational settings. | cs.CL | Pedagogical Alignment of Large Language Models |
2024-02-07T15:56:17Z | http://arxiv.org/pdf/2402.04978v2 | http://arxiv.org/abs/2402.04978v2 | Yihao Li, Ru Zhang, Jianyi Liu | 2024-06-12T16:03:31Z | cs.CL, cs.AI | While Large Language Models (LLMs) demonstrate exceptional performance in a
multitude of Natural Language Processing (NLP) tasks, they encounter challenges
in practical applications, including issues with hallucinations, inadequate
knowledge updating, and limited transparency in the reasoning process. To
overcome these limitations, this study innovatively proposes a collaborative
training-free reasoning scheme involving tight cooperation between Knowledge
Graph (KG) and LLMs. This scheme first involves using LLMs to iteratively
explore KG, selectively retrieving a task-relevant knowledge subgraph to
support reasoning. The LLMs are then guided to further combine inherent
implicit knowledge to reason on the subgraph while explicitly elucidating the
reasoning process. Through such a cooperative approach, our scheme achieves
more reliable knowledge-based reasoning and facilitates the tracing of the
reasoning results. Experimental results show that our scheme significantly
progressed across multiple datasets, notably achieving over a 10% improvement
on the QALD10 dataset compared to the best baseline and the fine-tuned
state-of-the-art (SOTA) work. Building on this success, this study hopes to
offer a valuable reference for future research in the fusion of KG and LLMs,
thereby enhancing LLMs' proficiency in solving complex issues. | cs.CL | An Enhanced Prompt-Based LLM Reasoning Scheme via Knowledge Graph-Integrated Collaboration |
2024-02-07T15:44:55Z | http://arxiv.org/pdf/2402.04967v1 | http://arxiv.org/abs/2402.04967v1 | Piush Aggarwal, Jawar Mehrabanian, Weigang Huang, Özge Alacam, Torsten Zesch | 2024-02-07T15:44:55Z | cs.CL, cs.AI, cs.CV | This paper delves into the formidable challenge of cross-domain
generalization in multimodal hate meme detection, presenting compelling
findings. We provide enough pieces of evidence supporting the hypothesis that
only the textual component of hateful memes enables the existing multimodal
classifier to generalize across different domains, while the image component
proves highly sensitive to a specific training dataset. The evidence includes
demonstrations showing that hate-text classifiers perform similarly to
hate-meme classifiers in a zero-shot setting. Simultaneously, the introduction
of captions generated from images of memes to the hate-meme classifier worsens
performance by an average F1 of 0.02. Through blackbox explanations, we
identify a substantial contribution of the text modality (average of 83%),
which diminishes with the introduction of meme's image captions (52%).
Additionally, our evaluation on a newly created confounder dataset reveals
higher performance on text confounders as compared to image confounders with an
average $\Delta$F1 of 0.18. | cs.CL | Text or Image? What is More Important in Cross-Domain Generalization Capabilities of Hate Meme Detection Models? |
2024-02-07T15:40:22Z | http://arxiv.org/pdf/2402.04957v2 | http://arxiv.org/abs/2402.04957v2 | Lihu Chen, Alexandre Perez-Lebel, Fabian M. Suchanek, Gaël Varoquaux | 2024-06-18T20:31:56Z | cs.CL | Large Language Models (LLMs), including ChatGPT and LLaMA, are susceptible to
generating hallucinated answers in a confident tone. While efforts to elicit
and calibrate confidence scores have proven useful, recent findings show that
controlling uncertainty must go beyond calibration: predicted scores may
deviate significantly from the actual posterior probabilities due to the impact
of grouping loss. In this work, we construct a new evaluation dataset derived
from a knowledge base to assess confidence scores given to answers of Mistral
and LLaMA. Experiments show that they tend to be overconfident. Further, we
show that they are more overconfident on some answers than others, \emph{eg}
depending on the nationality of the person in the query. In
uncertainty-quantification theory, this is grouping loss. To address this, we
propose a solution to reconfidence LLMs, canceling not only calibration but
also grouping loss. The LLMs, after the reconfidencing process, indicate
improved confidence alignment with the accuracy of their responses. | cs.CL | Reconfidencing LLMs from the Grouping Loss Perspective |
2024-02-07T14:44:42Z | http://arxiv.org/pdf/2402.04918v1 | http://arxiv.org/abs/2402.04918v1 | Frances Yung, Mansoor Ahmad, Merel Scholman, Vera Demberg | 2024-02-07T14:44:42Z | cs.CL, cs.AI | Pre-trained large language models, such as ChatGPT, archive outstanding
performance in various reasoning tasks without supervised training and were
found to have outperformed crowdsourcing workers. Nonetheless, ChatGPT's
performance in the task of implicit discourse relation classification, prompted
by a standard multiple-choice question, is still far from satisfactory and
considerably inferior to state-of-the-art supervised approaches. This work
investigates several proven prompting techniques to improve ChatGPT's
recognition of discourse relations. In particular, we experimented with
breaking down the classification task that involves numerous abstract labels
into smaller subtasks. Nonetheless, experiment results show that the inference
accuracy hardly changes even with sophisticated prompt engineering, suggesting
that implicit discourse relation classification is not yet resolvable under
zero-shot or few-shot settings. | cs.CL | Prompting Implicit Discourse Relation Annotation |
2024-02-07T14:41:08Z | http://arxiv.org/pdf/2402.04914v1 | http://arxiv.org/abs/2402.04914v1 | Bashar Alhafni, Vivek Kulkarni, Dhruv Kumar, Vipul Raheja | 2024-02-07T14:41:08Z | cs.CL | As the text generation capabilities of large language models become
increasingly prominent, recent studies have focused on controlling particular
aspects of the generated text to make it more personalized. However, most
research on controllable text generation focuses on controlling the content or
modeling specific high-level/coarse-grained attributes that reflect authors'
writing styles, such as formality, domain, or sentiment. In this paper, we
focus on controlling fine-grained attributes spanning multiple linguistic
dimensions, such as lexical and syntactic attributes. We introduce a novel
benchmark to train generative models and evaluate their ability to generate
personalized text based on multiple fine-grained linguistic attributes. We
systematically investigate the performance of various large language models on
our benchmark and draw insights from the factors that impact their performance.
We make our code, data, and pretrained models publicly available. | cs.CL | Personalized Text Generation with Fine-Grained Linguistic Control |
2024-02-07T14:35:05Z | http://arxiv.org/pdf/2402.04902v3 | http://arxiv.org/abs/2402.04902v3 | Hyesung Jeon, Yulhwa Kim, Jae-joon Kim | 2024-05-22T20:23:54Z | cs.LG, cs.CL | Due to the high memory and computational costs associated with Large Language
Models, model compression via quantization and parameter-efficient fine-tuning
(PEFT) methods, such as low-rank adaptation (LoRA), are gaining popularity.
This has led to active research on quantization-aware PEFT techniques, which
aim to create models with high accuracy and low memory overhead. Among
quantization methods, post-training quantization (PTQ) is more commonly used in
previous works than quantization-aware training (QAT), despite QAT's potential
for higher accuracy. This preference is due to PTQ's low training overhead.
However, PTQ-based PEFT methods often utilize high-precision parameters, making
it difficult to fully exploit the efficiency of quantization. Additionally,
they have limited adaptation ability due to a reduced and constrained LoRA
parameter structure. To overcome these challenges, we propose L4Q, which
leverages joint quantization and fine-tuning to reduce QAT's memory overhead
and produce models that consist entirely of quantized weights while achieving
effective adaptation to downstream tasks. By design, L4Q allows quantization
parameters to reflect weight updates, while weight updates reduce quantization
errors. Our experiments demonstrate that this coupled quantization and
fine-tuning approach yields superior accuracy compared to decoupled fine-tuning
schemes in sub-4-bit quantization. Using the LLaMA model families and
instructional datasets, we showcase L4Q's capabilities in language tasks and
few-shot in-context learning. | cs.LG | L4Q: Parameter Efficient Quantization-Aware Fine-Tuning on Large Language Models |
2024-02-07T14:22:51Z | http://arxiv.org/pdf/2402.04889v2 | http://arxiv.org/abs/2402.04889v2 | Sebastian Schmidt, Ines Zelch, Janek Bevendorff, Benno Stein, Matthias Hagen, Martin Potthast | 2024-04-30T09:15:42Z | cs.IR, cs.CL | Conversational search engines such as YouChat and Microsoft Copilot use large
language models (LLMs) to generate responses to queries. It is only a small
step to also let the same technology insert ads within the generated responses
- instead of separately placing ads next to a response. Inserted ads would be
reminiscent of native advertising and product placement, both of which are very
effective forms of subtle and manipulative advertising. Considering the high
computational costs associated with LLMs, for which providers need to develop
sustainable business models, users of conversational search engines may very
well be confronted with generated native ads in the near future. In this paper,
we thus take a first step to investigate whether LLMs can also be used as a
countermeasure, i.e., to block generated native ads. We compile the Webis
Generated Native Ads 2024 dataset of queries and generated responses with
automatically inserted ads, and evaluate whether LLMs or fine-tuned sentence
transformers can detect the ads. In our experiments, the investigated LLMs
struggle with the task but sentence transformers achieve precision and recall
values above 0.9. | cs.IR | Detecting Generated Native Ads in Conversational Search |
2024-02-07T14:16:28Z | http://arxiv.org/pdf/2402.04875v3 | http://arxiv.org/abs/2402.04875v3 | Kartik Ahuja, Amin Mansouri | 2024-06-07T20:25:05Z | cs.LG, cs.CL, stat.ML | Out-of-distribution generalization capabilities of sequence-to-sequence
models can be studied from the lens of two crucial forms of generalization:
length generalization -- the ability to generalize to longer sequences than
ones seen during training, and compositional generalization: the ability to
generalize to token combinations not seen during training. In this work, we
provide first provable guarantees on length and compositional generalization
for common sequence-to-sequence models -- deep sets, transformers, state space
models, and recurrent neural nets -- trained to minimize the prediction error.
Taking a first principles perspective, we study the realizable case, i.e., the
labeling function is realizable on the architecture. We show that limited
capacity versions of these different architectures achieve both length and
compositional generalization. Across different architectures, we also find that
a linear relationship between the learned representation and the representation
in the labeling function is necessary for length and compositional
generalization. | cs.LG | On Provable Length and Compositional Generalization |
2024-02-07T13:55:27Z | http://arxiv.org/pdf/2402.04858v2 | http://arxiv.org/abs/2402.04858v2 | Natasha Butt, Blazej Manczak, Auke Wiggers, Corrado Rainone, David W. Zhang, Michaël Defferrard, Taco Cohen | 2024-07-01T10:03:33Z | cs.AI, cs.CL, cs.LG | Large language models are increasingly solving tasks that are commonly
believed to require human-level reasoning ability. However, these models still
perform very poorly on benchmarks of general intelligence such as the
Abstraction and Reasoning Corpus (ARC). In this paper, we approach ARC as a
programming-by-examples problem, and introduce a novel and scalable method for
language model self-improvement called Code Iteration (CodeIt). Our method
iterates between 1) program sampling and hindsight relabeling, and 2) learning
from prioritized experience replay. By relabeling the goal of an episode (i.e.,
the target program output given input) to the realized output produced by the
sampled program, our method effectively deals with the extreme sparsity of
rewards in program synthesis. Applying CodeIt to the ARC dataset, we
demonstrate that prioritized hindsight replay, along with pre-training and
data-augmentation, leads to successful inter-task generalization. CodeIt is the
first neuro-symbolic approach that scales to the full ARC evaluation dataset.
Our method solves 15% of ARC evaluation tasks, achieving state-of-the-art
performance and outperforming existing neural and symbolic baselines. Our code
is available at https://github.com/Qualcomm-AI-research/codeit . | cs.AI | CodeIt: Self-Improving Language Models with Prioritized Hindsight Replay |
2024-02-07T13:54:06Z | http://arxiv.org/pdf/2402.04854v4 | http://arxiv.org/abs/2402.04854v4 | Jinghong Li, Huy Phan, Wen Gu, Koichi Ota, Shinobu Hasegawa | 2024-07-03T01:34:52Z | cs.DL, cs.CL, cs.LG | Research surveys have always posed a challenge for beginner researchers who
lack of research training. These researchers struggle to understand the
directions within their research topic, and the discovery of new research
findings within a short time. One way to provide intuitive assistance to
beginner researchers is by offering relevant knowledge graphs(KG) and
recommending related academic papers. However, existing navigation knowledge
graphs primarily rely on keywords in the research field and often fail to
present the logical hierarchy among multiple related papers clearly. Moreover,
most recommendation systems for academic papers simply rely on high text
similarity, which can leave researchers confused as to why a particular article
is being recommended. They may lack of grasp important information about the
insight connection between "Issue resolved" and "Issue finding" that they hope
to obtain. To address these issues, this study aims to support research insight
surveys for beginner researchers by establishing a hierarchical tree-structured
knowledge graph that reflects the inheritance insight of research topics and
the relevance insight among the academic papers. | cs.DL | Hierarchical Tree-structured Knowledge Graph For Academic Insight Survey |
2024-02-07T13:41:22Z | http://arxiv.org/pdf/2403.15391v1 | http://arxiv.org/abs/2403.15391v1 | Mohammad Ali Dadgostarnia, Ramin Mousa, Saba Hesaraki | 2024-02-07T13:41:22Z | cs.CL, cs.SI | Along with factors such as cancer, blood pressure, street accidents and
stroke, suicide has been one of Iran main causes of death. One of the main
reasons for suicide is psychological stressors. Identifying psychological
stressors in an at risk population can help in the early prevention of suicidal
and suicidal behaviours. In recent years, the widespread popularity and flow of
real time information sharing of social media have allowed for potential early
intervention in large scale and even small scale populations. However, some
automated approaches to extract psychiatric stressors from Twitter have been
presented, but most of this research has been for non Persian languages. This
study aims to investigate the techniques of detecting psychological stress
related to suicide from Persian tweets using learning based methods. The
proposed capsule based approach achieved a binary classification accuracy of
0.83. | cs.CL | CapsF: Capsule Fusion for Extracting psychiatric stressors for suicide from twitter |
2024-02-07T13:39:38Z | http://arxiv.org/pdf/2402.04838v4 | http://arxiv.org/abs/2402.04838v4 | Jinghui Lu, Ziwei Yang, Yanjie Wang, Xuejing Liu, Brian Mac Namee, Can Huang | 2024-02-15T01:52:16Z | cs.CL, cs.AI | In this study, we aim to reduce generation latency for Named Entity
Recognition (NER) with Large Language Models (LLMs). The main cause of high
latency in LLMs is the sequential decoding process, which autoregressively
generates all labels and mentions for NER, significantly increase the sequence
length. To this end, we introduce Parallel Decoding in LLM for NE}
(PaDeLLM-NER), a approach that integrates seamlessly into existing generative
model frameworks without necessitating additional modules or architectural
modifications. PaDeLLM-NER allows for the simultaneous decoding of all
mentions, thereby reducing generation latency. Experiments reveal that
PaDeLLM-NER significantly increases inference speed that is 1.76 to 10.22 times
faster than the autoregressive approach for both English and Chinese.
Simultaneously it maintains the quality of predictions as evidenced by the
performance that is on par with the state-of-the-art across various datasets. | cs.CL | PaDeLLM-NER: Parallel Decoding in Large Language Models for Named Entity Recognition |
2024-02-07T13:39:24Z | http://arxiv.org/pdf/2402.10938v1 | http://arxiv.org/abs/2402.10938v1 | Arash Amini, Yigit Ege Bayiz, Ashwin Ram, Radu Marculescu, Ufuk Topcu | 2024-02-07T13:39:24Z | cs.CL, cs.SI | In the era of social media platforms, identifying the credibility of online
content is crucial to combat misinformation. We present the CREDiBERT
(CREDibility assessment using Bi-directional Encoder Representations from
Transformers), a source credibility assessment model fine-tuned for Reddit
submissions focusing on political discourse as the main contribution. We adopt
a semi-supervised training approach for CREDiBERT, leveraging Reddit's
community-based structure. By encoding submission content using CREDiBERT and
integrating it into a Siamese neural network, we significantly improve the
binary classification of submission credibility, achieving a 9% increase in F1
score compared to existing methods. Additionally, we introduce a new version of
the post-to-post network in Reddit that efficiently encodes user interactions
to enhance the binary classification task by nearly 8% in F1 score. Finally, we
employ CREDiBERT to evaluate the susceptibility of subreddits with respect to
different topics. | cs.CL | News Source Credibility Assessment: A Reddit Case Study |
2024-02-07T13:36:02Z | http://arxiv.org/pdf/2403.06988v1 | http://arxiv.org/abs/2403.06988v1 | Luca Beurer-Kellner, Marc Fischer, Martin Vechev | 2024-02-07T13:36:02Z | cs.LG, cs.CL | To ensure that text generated by large language models (LLMs) is in an
expected format, constrained decoding proposes to enforce strict formal
language constraints during generation. However, as we show in this work, not
only do such methods incur performance overhead during generation, but many of
them also significantly impair task accuracy, if they do not correctly align
the underlying LLM sub-word vocabularies with external constraints. To address
this, we present a novel decoding algorithm, DOMINO, that can enforce
constraints in a fully subword-aligned fashion, while leveraging
pre-computation and speculative decoding to achieve virtually no overhead and
in some cases even almost 2$\times$ speedup over unconstrained decoding --
thereby outperforming existing approaches by a wide margin. | cs.LG | Guiding LLMs The Right Way: Fast, Non-Invasive Constrained Generation |
2024-02-07T13:32:11Z | http://arxiv.org/pdf/2402.04833v2 | http://arxiv.org/abs/2402.04833v2 | Hao Zhao, Maksym Andriushchenko, Francesco Croce, Nicolas Flammarion | 2024-06-04T17:20:01Z | cs.CL | There is a consensus that instruction fine-tuning of LLMs requires
high-quality data, but what are they? LIMA (NeurIPS 2023) and AlpaGasus (ICLR
2024) are state-of-the-art methods for selecting such high-quality examples,
either via manual curation or using GPT-3.5-Turbo as a quality scorer. We show
that the extremely simple baseline of selecting the 1,000 instructions with
longest responses -- that intuitively contain more learnable information and
are harder to overfit -- from standard datasets can consistently outperform
these sophisticated methods according to GPT-4 and PaLM-2 as judges, while
remaining competitive on the Open LLM benchmarks that test factual knowledge.
We demonstrate this for several LLMs (Llama-2-7B, Llama-2-13B, Mistral-7B-v0.1)
and datasets (Alpaca-52k, Evol-Instruct-70k). In addition, a lightweight
refinement of such long instructions can further improve the abilities of the
fine-tuned LLMs, and allows us to obtain competitive results on MT-Bench and
the 2nd highest-ranked Llama-2-7B-based model on AlpacaEval 2.0, while training
on only 1,000 examples and no extra preference data. We also conduct a thorough
analysis of our models to ensure that their enhanced performance is not simply
due to GPT-4's preference for longer responses. Overall, our findings suggest
that fine-tuning on the longest responses should be the default baseline for
any work on instruction fine-tuning. We provide our code at
https://github.com/tml-epfl/long-is-more-for-alignment. | cs.CL | Long Is More for Alignment: A Simple but Tough-to-Beat Baseline for Instruction Fine-Tuning |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.