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
|
|---|---|---|---|---|---|---|---|---|
2024-04-05T02:35:43Z | http://arxiv.org/pdf/2404.03865v1 | http://arxiv.org/abs/2404.03865v1 | Ajay Jaiswal, Bodun Hu, Lu Yin, Yeonju Ro, Shiwei Liu, Tianlong Chen, Aditya Akella | 2024-04-05T02:35:43Z | cs.CL, cs.LG | Autoregressive Large Language Models (e.g., LLaMa, GPTs) are omnipresent
achieving remarkable success in language understanding and generation. However,
such impressive capability typically comes with a substantial model size, which
presents significant challenges for autoregressive token-by-token generation.
To mitigate computation overload incurred during generation, several early-exit
and layer-dropping strategies have been proposed. Despite some promising
success due to the redundancy across LLMs layers on metrics like Rough-L/BLUE,
our careful knowledge-intensive evaluation unveils issues such as generation
collapse, hallucination of wrong facts, and noticeable performance drop even at
the trivial exit ratio of 10-15% of layers. We attribute these errors primarily
to ineffective handling of the KV cache through state copying during
early-exit. In this work, we observed the saturation of computationally
expensive feed-forward blocks of LLM layers and proposed FFN-SkipLLM, which is
a novel fine-grained skip strategy of autoregressive LLMs. More specifically,
FFN-SkipLLM is an input-adaptive feed-forward skipping strategy that can skip
25-30% of FFN blocks of LLMs with marginal change in performance on
knowledge-intensive generation tasks without any requirement to handle KV
cache. Our extensive experiments and ablation across benchmarks like MT-Bench,
Factoid-QA, and variable-length text summarization illustrate how our simple
and ease-at-use method can facilitate faster autoregressive decoding. | cs.CL | FFN-SkipLLM: A Hidden Gem for Autoregressive Decoding with Adaptive Feed Forward Skipping |
2024-04-05T02:27:09Z | http://arxiv.org/pdf/2404.03862v1 | http://arxiv.org/abs/2404.03862v1 | Jingyu Zhang, Marc Marone, Tianjian Li, Benjamin Van Durme, Daniel Khashabi | 2024-04-05T02:27:09Z | cs.CL | For humans to trust the fluent generations of large language models (LLMs),
they must be able to verify their correctness against trusted, external
sources. Recent efforts aim to increase verifiability through citations of
retrieved documents or post-hoc provenance. However, such citations are prone
to mistakes that further complicate their verifiability. To address these
limitations, we tackle the verifiability goal with a different philosophy: we
trivialize the verification process by developing models that quote verbatim
statements from trusted sources in pre-training data. We propose Quote-Tuning,
which demonstrates the feasibility of aligning LLMs to leverage memorized
information and quote from pre-training data. Quote-Tuning quantifies quoting
against large corpora with efficient membership inference tools, and uses the
amount of quotes as an implicit reward signal to construct a synthetic
preference dataset for quoting, without any human annotation. Next, the target
model is aligned to quote using preference optimization algorithms.
Experimental results show that Quote-Tuning significantly increases the
percentage of LLM generation quoted verbatim from high-quality pre-training
documents by 55% to 130% relative to untuned models while maintaining response
quality. Further experiments demonstrate that Quote-Tuning generalizes quoting
to out-of-domain data, is applicable in different tasks, and provides
additional benefits to truthfulness. Quote-Tuning not only serves as a
hassle-free method to increase quoting but also opens up avenues for improving
LLM trustworthiness through better verifiability. | cs.CL | Verifiable by Design: Aligning Language Models to Quote from Pre-Training Data |
2024-04-05T01:17:25Z | http://arxiv.org/pdf/2404.03854v2 | http://arxiv.org/abs/2404.03854v2 | Zitao Shuai, Liyue Shen | 2024-05-24T15:08:38Z | cs.LG, cs.CL, cs.CV | Vision-language pre-training (VLP) has arised as an efficient scheme for
multimodal representation learning, but it requires large-scale multimodal data
for pre-training, making it an obstacle especially for medical applications. To
overcome the data limitation, federated learning (FL) can be a promising
strategy to scale up the dataset for medical VLP while protecting data privacy.
However, client data are often heterogeneous in real-world scenarios, and we
observe that local training on heterogeneous client data would distort the
multimodal representation learning and lead to biased cross-modal alignment. To
address this challenge, we propose a Federated Align as IDeal (FedAID)
framework for federated VLP with robustness to data heterogeneity, to bind
local clients with an ideal crossmodal alignment. Specifically, to reduce
distortions on global-aggregated features while learning diverse semantics from
client datasets during local training, we propose to bind the cross-model
aligned representation space learned by local models with an unbiased one via
guidance-based regularization. Moreover, we employ a distribution-based min-max
optimization to learn the unbiased cross-modal alignment at each communication
turn of federated pre-training. The experiments on real-world datasets
demonstrate our method successfully promotes efficient federated multimodal
learning for medical VLP with data heterogeneity. | cs.LG | Align as Ideal: Cross-Modal Alignment Binding for Federated Medical Vision-Language Pre-training |
2024-04-04T22:52:41Z | http://arxiv.org/pdf/2404.03823v1 | http://arxiv.org/abs/2404.03823v1 | Zahra Mousavi, Chadni Islam, Kristen Moore, Alsharif Abuadbba, Muhammad Ali Babar | 2024-04-04T22:52:41Z | cs.CR, cs.CL, cs.CY | The increasing trend of using Large Language Models (LLMs) for code
generation raises the question of their capability to generate trustworthy
code. While many researchers are exploring the utility of code generation for
uncovering software vulnerabilities, one crucial but often overlooked aspect is
the security Application Programming Interfaces (APIs). APIs play an integral
role in upholding software security, yet effectively integrating security APIs
presents substantial challenges. This leads to inadvertent misuse by
developers, thereby exposing software to vulnerabilities. To overcome these
challenges, developers may seek assistance from LLMs. In this paper, we
systematically assess ChatGPT's trustworthiness in code generation for security
API use cases in Java. To conduct a thorough evaluation, we compile an
extensive collection of 48 programming tasks for 5 widely used security APIs.
We employ both automated and manual approaches to effectively detect security
API misuse in the code generated by ChatGPT for these tasks. Our findings are
concerning: around 70% of the code instances across 30 attempts per task
contain security API misuse, with 20 distinct misuse types identified.
Moreover, for roughly half of the tasks, this rate reaches 100%, indicating
that there is a long way to go before developers can rely on ChatGPT to
securely implement security API code. | cs.CR | An Investigation into Misuse of Java Security APIs by Large Language Models |
2024-04-04T22:31:58Z | http://arxiv.org/pdf/2404.03820v2 | http://arxiv.org/abs/2404.03820v2 | Makesh Narsimhan Sreedhar, Traian Rebedea, Shaona Ghosh, Jiaqi Zeng, Christopher Parisien | 2024-06-21T13:57:11Z | cs.CL | Recent advancements in instruction-tuning datasets have predominantly focused
on specific tasks like mathematical or logical reasoning. There has been a
notable gap in data designed for aligning language models to maintain topic
relevance in conversations - a critical aspect for deploying chatbots to
production. We introduce the CantTalkAboutThis dataset to help language models
remain focused on the subject at hand during task-oriented interactions. It
consists of synthetic dialogues on a wide range of conversation topics from
different domains. These dialogues are interspersed with distractor turns that
intentionally divert the chatbot from the predefined topic. Fine-tuning
language models on this dataset helps make them resilient to deviating from the
role assigned and improves their ability to maintain topical coherence compared
to general-purpose instruction-tuned LLMs like GPT-4-turbo and
Mixtral-Instruct. Additionally, preliminary observations suggest that training
models on this dataset also enhance their performance on fine-grained
instruction following tasks, including safety alignment. | cs.CL | CantTalkAboutThis: Aligning Language Models to Stay on Topic in Dialogues |
2024-04-04T21:57:11Z | http://arxiv.org/pdf/2404.03818v1 | http://arxiv.org/abs/2404.03818v1 | Zhangdie Yuan, Chenxi Whitehouse, Eric Chamoun, Rami Aly, Andreas Vlachos | 2024-04-04T21:57:11Z | cs.CL | This paper introduces PRobELM (Plausibility Ranking Evaluation for Language
Models), a benchmark designed to assess language models' ability to discern
more plausible from less plausible scenarios through their parametric
knowledge. While benchmarks such as TruthfulQA emphasise factual accuracy or
truthfulness, and others such as COPA explore plausible scenarios without
explicitly incorporating world knowledge, PRobELM seeks to bridge this gap by
evaluating models' capabilities to prioritise plausible scenarios that leverage
world knowledge over less plausible alternatives. This design allows us to
assess the potential of language models for downstream use cases such as
literature-based discovery where the focus is on identifying information that
is likely but not yet known. Our benchmark is constructed from a dataset
curated from Wikidata edit histories, tailored to align the temporal bounds of
the training data for the evaluated models. PRobELM facilitates the evaluation
of language models across multiple prompting types, including statement, text
completion, and question-answering. Experiments with 10 models of various sizes
and architectures on the relationship between model scales, training recency,
and plausibility performance, reveal that factual accuracy does not directly
correlate with plausibility performance and that up-to-date training data
enhances plausibility assessment across different model architectures. | cs.CL | PRobELM: Plausibility Ranking Evaluation for Language Models |
2024-04-04T21:47:43Z | http://arxiv.org/pdf/2404.04302v1 | http://arxiv.org/abs/2404.04302v1 | Nirmalie Wiratunga, Ramitha Abeyratne, Lasal Jayawardena, Kyle Martin, Stewart Massie, Ikechukwu Nkisi-Orji, Ruvan Weerasinghe, Anne Liret, Bruno Fleisch | 2024-04-04T21:47:43Z | cs.CL, cs.AI | Retrieval-Augmented Generation (RAG) enhances Large Language Model (LLM)
output by providing prior knowledge as context to input. This is beneficial for
knowledge-intensive and expert reliant tasks, including legal
question-answering, which require evidence to validate generated text outputs.
We highlight that Case-Based Reasoning (CBR) presents key opportunities to
structure retrieval as part of the RAG process in an LLM. We introduce CBR-RAG,
where CBR cycle's initial retrieval stage, its indexing vocabulary, and
similarity knowledge containers are used to enhance LLM queries with
contextually relevant cases. This integration augments the original LLM query,
providing a richer prompt. We present an evaluation of CBR-RAG, and examine
different representations (i.e. general and domain-specific embeddings) and
methods of comparison (i.e. inter, intra and hybrid similarity) on the task of
legal question-answering. Our results indicate that the context provided by
CBR's case reuse enforces similarity between relevant components of the
questions and the evidence base leading to significant improvements in the
quality of generated answers. | cs.CL | CBR-RAG: Case-Based Reasoning for Retrieval Augmented Generation in LLMs for Legal Question Answering |
2024-04-04T20:27:37Z | http://arxiv.org/pdf/2404.04298v1 | http://arxiv.org/abs/2404.04298v1 | Dongwei Jiang, Jingyu Zhang, Orion Weller, Nathaniel Weir, Benjamin Van Durme, Daniel Khashabi | 2024-04-04T20:27:37Z | cs.AI, cs.CL, cs.LG | Can LLMs continually improve their previous outputs for better results? An
affirmative answer would require LLMs to be better at discriminating among
previously-generated alternatives, than generating initial responses. We
explore the validity of this hypothesis in practice. We first introduce a
unified framework that allows us to compare the generative and discriminative
capability of any model on any task. Then, in our resulting experimental
analysis of several LLMs, we do not observe the performance of those models on
discrimination to be reliably better than generation. We hope these findings
inform the growing literature on self-improvement AI systems. | cs.AI | SELF-[IN]CORRECT: LLMs Struggle with Refining Self-Generated Responses |
2024-04-04T18:35:25Z | http://arxiv.org/pdf/2404.03746v1 | http://arxiv.org/abs/2404.03746v1 | Kaustubh Dhole, Eugene Agichtein | 2024-04-04T18:35:25Z | cs.IR, cs.AI, cs.CL | Query Reformulation(QR) is a set of techniques used to transform a user's
original search query to a text that better aligns with the user's intent and
improves their search experience. Recently, zero-shot QR has been shown to be a
promising approach due to its ability to exploit knowledge inherent in large
language models. By taking inspiration from the success of ensemble prompting
strategies which have benefited many tasks, we investigate if they can help
improve query reformulation. In this context, we propose an ensemble based
prompting technique, GenQREnsemble which leverages paraphrases of a zero-shot
instruction to generate multiple sets of keywords ultimately improving
retrieval performance. We further introduce its post-retrieval variant,
GenQREnsembleRF to incorporate pseudo relevant feedback. On evaluations over
four IR benchmarks, we find that GenQREnsemble generates better reformulations
with relative nDCG@10 improvements up to 18% and MAP improvements upto 24% over
the previous zero-shot state-of-art. On the MSMarco Passage Ranking task,
GenQREnsembleRF shows relative gains of 5% MRR using pseudo-relevance feedback,
and 9% nDCG@10 using relevant feedback documents. | cs.IR | GenQREnsemble: Zero-Shot LLM Ensemble Prompting for Generative Query Reformulation |
2024-04-04T18:34:32Z | http://arxiv.org/pdf/2404.03745v1 | http://arxiv.org/abs/2404.03745v1 | Mahjabin Nahar, Haeseung Seo, Eun-Ju Lee, Aiping Xiong, Dongwon Lee | 2024-04-04T18:34:32Z | cs.HC, cs.AI, cs.CL | The widespread adoption and transformative effects of large language models
(LLMs) have sparked concerns regarding their capacity to produce inaccurate and
fictitious content, referred to as `hallucinations'. Given the potential risks
associated with hallucinations, humans should be able to identify them. This
research aims to understand the human perception of LLM hallucinations by
systematically varying the degree of hallucination (genuine, minor
hallucination, major hallucination) and examining its interaction with warning
(i.e., a warning of potential inaccuracies: absent vs. present). Participants
(N=419) from Prolific rated the perceived accuracy and engaged with content
(e.g., like, dislike, share) in a Q/A format. Results indicate that humans rank
content as truthful in the order genuine > minor hallucination > major
hallucination and user engagement behaviors mirror this pattern. More
importantly, we observed that warning improves hallucination detection without
significantly affecting the perceived truthfulness of genuine content. We
conclude by offering insights for future tools to aid human detection of
hallucinations. | cs.HC | Fakes of Varying Shades: How Warning Affects Human Perception and Engagement Regarding LLM Hallucinations |
2024-04-04T18:01:21Z | http://arxiv.org/pdf/2404.03732v1 | http://arxiv.org/abs/2404.03732v1 | Bradley P. Allen, Fina Polat, Paul Groth | 2024-04-04T18:01:21Z | cs.CL, cs.AI | We describe the University of Amsterdam Intelligent Data Engineering Lab
team's entry for the SemEval-2024 Task 6 competition. The SHROOM-INDElab system
builds on previous work on using prompt programming and in-context learning
with large language models (LLMs) to build classifiers for hallucination
detection, and extends that work through the incorporation of context-specific
definition of task, role, and target concept, and automated generation of
examples for use in a few-shot prompting approach. The resulting system
achieved fourth-best and sixth-best performance in the model-agnostic track and
model-aware tracks for Task 6, respectively, and evaluation using the
validation sets showed that the system's classification decisions were
consistent with those of the crowd-sourced human labellers. We further found
that a zero-shot approach provided better accuracy than a few-shot approach
using automatically generated examples. Code for the system described in this
paper is available on Github. | cs.CL | SHROOM-INDElab at SemEval-2024 Task 6: Zero- and Few-Shot LLM-Based Classification for Hallucination Detection |
2024-04-04T17:59:46Z | http://arxiv.org/pdf/2404.03653v2 | http://arxiv.org/abs/2404.03653v2 | Dongzhi Jiang, Guanglu Song, Xiaoshi Wu, Renrui Zhang, Dazhong Shen, Zhuofan Zong, Yu Liu, Hongsheng Li | 2024-06-03T06:02:34Z | cs.CV, cs.AI, cs.CL | Diffusion models have demonstrated great success in the field of
text-to-image generation. However, alleviating the misalignment between the
text prompts and images is still challenging. The root reason behind the
misalignment has not been extensively investigated. We observe that the
misalignment is caused by inadequate token attention activation. We further
attribute this phenomenon to the diffusion model's insufficient condition
utilization, which is caused by its training paradigm. To address the issue, we
propose CoMat, an end-to-end diffusion model fine-tuning strategy with an
image-to-text concept matching mechanism. We leverage an image captioning model
to measure image-to-text alignment and guide the diffusion model to revisit
ignored tokens. A novel attribute concentration module is also proposed to
address the attribute binding problem. Without any image or human preference
data, we use only 20K text prompts to fine-tune SDXL to obtain CoMat-SDXL.
Extensive experiments show that CoMat-SDXL significantly outperforms the
baseline model SDXL in two text-to-image alignment benchmarks and achieves
start-of-the-art performance. | cs.CV | CoMat: Aligning Text-to-Image Diffusion Model with Image-to-Text Concept Matching |
2024-04-04T17:58:40Z | http://arxiv.org/pdf/2404.03648v1 | http://arxiv.org/abs/2404.03648v1 | Hanyu Lai, Xiao Liu, Iat Long Iong, Shuntian Yao, Yuxuan Chen, Pengbo Shen, Hao Yu, Hanchen Zhang, Xiaohan Zhang, Yuxiao Dong, Jie Tang | 2024-04-04T17:58:40Z | cs.CL | Large language models (LLMs) have fueled many intelligent agent tasks, such
as web navigation -- but most existing agents perform far from satisfying in
real-world webpages due to three factors: (1) the versatility of actions on
webpages, (2) HTML text exceeding model processing capacity, and (3) the
complexity of decision-making due to the open-domain nature of web. In light of
the challenge, we develop AutoWebGLM, a GPT-4-outperforming automated web
navigation agent built upon ChatGLM3-6B. Inspired by human browsing patterns,
we design an HTML simplification algorithm to represent webpages, preserving
vital information succinctly. We employ a hybrid human-AI method to build web
browsing data for curriculum training. Then, we bootstrap the model by
reinforcement learning and rejection sampling to further facilitate webpage
comprehension, browser operations, and efficient task decomposition by itself.
For testing, we establish a bilingual benchmark -- AutoWebBench -- for
real-world web browsing tasks. We evaluate AutoWebGLM across diverse web
navigation benchmarks, revealing its improvements but also underlying
challenges to tackle real environments. Related code, model, and data will be
released at \url{https://github.com/THUDM/AutoWebGLM}. | cs.CL | AutoWebGLM: Bootstrap And Reinforce A Large Language Model-based Web Navigating Agent |
2024-04-04T17:58:31Z | http://arxiv.org/pdf/2404.03646v1 | http://arxiv.org/abs/2404.03646v1 | Arnab Sen Sharma, David Atkinson, David Bau | 2024-04-04T17:58:31Z | cs.CL | We investigate the mechanisms of factual recall in the Mamba state space
model. Our work is inspired by previous findings in autoregressive transformer
language models suggesting that their knowledge recall is localized to
particular modules at specific token locations; we therefore ask whether
factual recall in Mamba can be similarly localized. To investigate this, we
conduct four lines of experiments on Mamba. First, we apply causal tracing or
interchange interventions to localize key components inside Mamba that are
responsible for recalling facts, revealing that specific components within
middle layers show strong causal effects at the last token of the subject,
while the causal effect of intervening on later layers is most pronounced at
the last token of the prompt, matching previous findings on autoregressive
transformers. Second, we show that rank-one model editing methods can
successfully insert facts at specific locations, again resembling findings on
transformer models. Third, we examine the linearity of Mamba's representations
of factual relations. Finally we adapt attention-knockout techniques to Mamba
to dissect information flow during factual recall. We compare Mamba directly to
a similar-sized transformer and conclude that despite significant differences
in architectural approach, when it comes to factual recall, the two
architectures share many similarities. | cs.CL | Locating and Editing Factual Associations in Mamba |
2024-04-04T17:58:02Z | http://arxiv.org/pdf/2404.04125v2 | http://arxiv.org/abs/2404.04125v2 | Vishaal Udandarao, Ameya Prabhu, Adhiraj Ghosh, Yash Sharma, Philip H. S. Torr, Adel Bibi, Samuel Albanie, Matthias Bethge | 2024-04-08T21:14:43Z | cs.CV, cs.CL, cs.LG | Web-crawled pretraining datasets underlie the impressive "zero-shot"
evaluation performance of multimodal models, such as CLIP for
classification/retrieval and Stable-Diffusion for image generation. However, it
is unclear how meaningful the notion of "zero-shot" generalization is for such
multimodal models, as it is not known to what extent their pretraining datasets
encompass the downstream concepts targeted for during "zero-shot" evaluation.
In this work, we ask: How is the performance of multimodal models on downstream
concepts influenced by the frequency of these concepts in their pretraining
datasets? We comprehensively investigate this question across 34 models and
five standard pretraining datasets (CC-3M, CC-12M, YFCC-15M, LAION-400M,
LAION-Aesthetics), generating over 300GB of data artifacts. We consistently
find that, far from exhibiting "zero-shot" generalization, multimodal models
require exponentially more data to achieve linear improvements in downstream
"zero-shot" performance, following a sample inefficient log-linear scaling
trend. This trend persists even when controlling for sample-level similarity
between pretraining and downstream datasets, and testing on purely synthetic
data distributions. Furthermore, upon benchmarking models on long-tailed data
sampled based on our analysis, we demonstrate that multimodal models across the
board perform poorly. We contribute this long-tail test set as the "Let it
Wag!" benchmark to further research in this direction. Taken together, our
study reveals an exponential need for training data which implies that the key
to "zero-shot" generalization capabilities under large-scale training paradigms
remains to be found. | cs.CV | No "Zero-Shot" Without Exponential Data: Pretraining Concept Frequency Determines Multimodal Model Performance |
2024-04-04T17:56:41Z | http://arxiv.org/pdf/2404.03715v1 | http://arxiv.org/abs/2404.03715v1 | Corby Rosset, Ching-An Cheng, Arindam Mitra, Michael Santacroce, Ahmed Awadallah, Tengyang Xie | 2024-04-04T17:56:41Z | cs.LG, cs.AI, cs.CL | This paper studies post-training large language models (LLMs) using
preference feedback from a powerful oracle to help a model iteratively improve
over itself. The typical approach for post-training LLMs involves Reinforcement
Learning from Human Feedback (RLHF), which traditionally separates reward
learning and subsequent policy optimization. However, such a reward
maximization approach is limited by the nature of "point-wise" rewards (such as
Bradley-Terry model), which fails to express complex intransitive or cyclic
preference relations. While advances on RLHF show reward learning and policy
optimization can be merged into a single contrastive objective for stability,
they yet still remain tethered to the reward maximization framework. Recently,
a new wave of research sidesteps the reward maximization presumptions in favor
of directly optimizing over "pair-wise" or general preferences. In this paper,
we introduce Direct Nash Optimization (DNO), a provable and scalable algorithm
that marries the simplicity and stability of contrastive learning with
theoretical generality from optimizing general preferences. Because DNO is a
batched on-policy algorithm using a regression-based objective, its
implementation is straightforward and efficient. Moreover, DNO enjoys monotonic
improvement across iterations that help it improve even over a strong teacher
(such as GPT-4). In our experiments, a resulting 7B parameter Orca-2.5 model
aligned by DNO achieves the state-of-the-art win-rate against GPT-4-Turbo of
33% on AlpacaEval 2.0 (even after controlling for response length), an absolute
gain of 26% (7% to 33%) over the initializing model. It outperforms models with
far more parameters, including Mistral Large, Self-Rewarding LM (70B
parameters), and older versions of GPT-4. | cs.LG | Direct Nash Optimization: Teaching Language Models to Self-Improve with General Preferences |
2024-04-04T17:54:33Z | http://arxiv.org/pdf/2404.03635v4 | http://arxiv.org/abs/2404.03635v4 | Ziyao Zeng, Daniel Wang, Fengyu Yang, Hyoungseob Park, Yangchao Wu, Stefano Soatto, Byung-Woo Hong, Dong Lao, Alex Wong | 2024-06-02T04:56:32Z | cs.CV, cs.AI, cs.CL, cs.LG, cs.MM | Three-dimensional (3D) reconstruction from a single image is an ill-posed
problem with inherent ambiguities, i.e. scale. Predicting a 3D scene from text
description(s) is similarly ill-posed, i.e. spatial arrangements of objects
described. We investigate the question of whether two inherently ambiguous
modalities can be used in conjunction to produce metric-scaled reconstructions.
To test this, we focus on monocular depth estimation, the problem of predicting
a dense depth map from a single image, but with an additional text caption
describing the scene. To this end, we begin by encoding the text caption as a
mean and standard deviation; using a variational framework, we learn the
distribution of the plausible metric reconstructions of 3D scenes corresponding
to the text captions as a prior. To "select" a specific reconstruction or depth
map, we encode the given image through a conditional sampler that samples from
the latent space of the variational text encoder, which is then decoded to the
output depth map. Our approach is trained alternatingly between the text and
image branches: in one optimization step, we predict the mean and standard
deviation from the text description and sample from a standard Gaussian, and in
the other, we sample using a (image) conditional sampler. Once trained, we
directly predict depth from the encoded text using the conditional sampler. We
demonstrate our approach on indoor (NYUv2) and outdoor (KITTI) scenarios, where
we show that language can consistently improve performance in both. | cs.CV | WorDepth: Variational Language Prior for Monocular Depth Estimation |
2024-04-04T17:48:28Z | http://arxiv.org/pdf/2404.03626v1 | http://arxiv.org/abs/2404.03626v1 | Brian Lester, Jaehoon Lee, Alex Alemi, Jeffrey Pennington, Adam Roberts, Jascha Sohl-Dickstein, Noah Constant | 2024-04-04T17:48:28Z | cs.CL, cs.LG | In this paper, we explore the idea of training large language models (LLMs)
over highly compressed text. While standard subword tokenizers compress text by
a small factor, neural text compressors can achieve much higher rates of
compression. If it were possible to train LLMs directly over neurally
compressed text, this would confer advantages in training and serving
efficiency, as well as easier handling of long text spans. The main obstacle to
this goal is that strong compression tends to produce opaque outputs that are
not well-suited for learning. In particular, we find that text na\"ively
compressed via Arithmetic Coding is not readily learnable by LLMs. To overcome
this, we propose Equal-Info Windows, a novel compression technique whereby text
is segmented into blocks that each compress to the same bit length. Using this
method, we demonstrate effective learning over neurally compressed text that
improves with scale, and outperforms byte-level baselines by a wide margin on
perplexity and inference speed benchmarks. While our method delivers worse
perplexity than subword tokenizers for models trained with the same parameter
count, it has the benefit of shorter sequence lengths. Shorter sequence lengths
require fewer autoregressive generation steps, and reduce latency. Finally, we
provide extensive analysis of the properties that contribute to learnability,
and offer concrete suggestions for how to further improve the performance of
high-compression tokenizers. | cs.CL | Training LLMs over Neurally Compressed Text |
2024-04-04T17:45:59Z | http://arxiv.org/pdf/2404.03623v1 | http://arxiv.org/abs/2404.03623v1 | Marco Bronzini, Carlo Nicolini, Bruno Lepri, Jacopo Staiano, Andrea Passerini | 2024-04-04T17:45:59Z | cs.CL, cs.AI, cs.CY | Large Language Models (LLMs) demonstrate an impressive capacity to recall a
vast range of common factual knowledge information. However, unravelling the
underlying reasoning of LLMs and explaining their internal mechanisms of
exploiting this factual knowledge remain active areas of investigation. Our
work analyzes the factual knowledge encoded in the latent representation of
LLMs when prompted to assess the truthfulness of factual claims. We propose an
end-to-end framework that jointly decodes the factual knowledge embedded in the
latent space of LLMs from a vector space to a set of ground predicates and
represents its evolution across the layers using a temporal knowledge graph.
Our framework relies on the technique of activation patching which intervenes
in the inference computation of a model by dynamically altering its latent
representations. Consequently, we neither rely on external models nor training
processes. We showcase our framework with local and global interpretability
analyses using two claim verification datasets: FEVER and CLIMATE-FEVER. The
local interpretability analysis exposes different latent errors from
representation to multi-hop reasoning errors. On the other hand, the global
analysis uncovered patterns in the underlying evolution of the model's factual
knowledge (e.g., store-and-seek factual information). By enabling graph-based
analyses of the latent representations, this work represents a step towards the
mechanistic interpretability of LLMs. | cs.CL | Unveiling LLMs: The Evolution of Latent Representations in a Temporal Knowledge Graph |
2024-04-04T17:45:08Z | http://arxiv.org/pdf/2404.03622v2 | http://arxiv.org/abs/2404.03622v2 | Wenshan Wu, Shaoguang Mao, Yadong Zhang, Yan Xia, Li Dong, Lei Cui, Furu Wei | 2024-05-24T04:07:44Z | cs.CL | Large language models (LLMs) have exhibited impressive performance in
language comprehension and various reasoning tasks. However, their abilities in
spatial reasoning, a crucial aspect of human cognition, remain relatively
unexplored. Human possess a remarkable ability to create mental images of
unseen objects and actions through a process known as the Mind's Eye, enabling
the imagination of the unseen world. Inspired by this cognitive capacity, we
propose Visualization-of-Thought (VoT) prompting. VoT aims to elicit spatial
reasoning of LLMs by visualizing their reasoning traces, thereby guiding
subsequent reasoning steps. We employed VoT for multi-hop spatial reasoning
tasks, including natural language navigation, visual navigation, and visual
tiling in 2D grid worlds. Experimental results demonstrated that VoT
significantly enhances the spatial reasoning abilities of LLMs. Notably, VoT
outperformed existing multimodal large language models (MLLMs) in these tasks.
While VoT works surprisingly well on LLMs, the ability to generate mental
images to facilitate spatial reasoning resembles the mind's eye process,
suggesting its potential viability in MLLMs. | cs.CL | Mind's Eye of LLMs: Visualization-of-Thought Elicits Spatial Reasoning in Large Language Models |
2024-04-04T17:31:32Z | http://arxiv.org/pdf/2404.03608v1 | http://arxiv.org/abs/2404.03608v1 | Longxu Dou, Qian Liu, Guangtao Zeng, Jia Guo, Jiahui Zhou, Wei Lu, Min Lin | 2024-04-04T17:31:32Z | cs.CL, cs.AI | We present Sailor, a family of open language models ranging from 0.5B to 7B
parameters, tailored for South-East Asian (SEA) languages. These models are
continually pre-trained from Qwen1.5, a great language model for multilingual
use cases. From Qwen1.5, Sailor models accept 200B to 400B tokens, primarily
covering the languages of English, Chinese, Vietnamese, Thai, Indonesian,
Malay, and Lao. The training leverages several techniques, including BPE
dropout for improving the model robustness, aggressive data cleaning and
deduplication, and small proxy models to optimize data mixture. Experimental
results on four typical tasks indicate that Sailor models demonstrate strong
performance across different benchmarks, including commonsense reasoning,
question answering, reading comprehension and examination. Embracing the
open-source spirit, we share our insights through this report to spark a wider
interest in developing large language models for multilingual use cases. | cs.CL | Sailor: Open Language Models for South-East Asia |
2024-04-04T17:25:30Z | http://arxiv.org/pdf/2404.03605v1 | http://arxiv.org/abs/2404.03605v1 | Aniruddha Nrusimha, Mayank Mishra, Naigang Wang, Dan Alistarh, Rameswar Panda, Yoon Kim | 2024-04-04T17:25:30Z | cs.LG, cs.CL | We consider the problem of accurate quantization for language models, where
both the weights and activations are uniformly quantized to 4 bits per
parameter, the lowest bitwidth format natively supported by GPU hardware. In
this context, the key challenge is activation quantization: it is known that
language models contain outlier channels whose values on average are orders of
magnitude higher than than other channels, which prevents accurate low-bitwidth
quantization with known techniques. We systematically study this phenomena and
find that these outlier channels emerge early in training, and that they occur
more frequently in layers with residual streams. We then propose a simple
strategy which regularizes a layer's inputs via quantization-aware training
(QAT) and its outputs via activation kurtosis regularization. We show that
regularizing both the inputs and outputs is crucial for preventing a model's
"migrating" the difficulty in input quantization to the weights, which makes
post-training quantization (PTQ) of weights more difficult. When combined with
weight PTQ, we show that our approach can obtain a W4A4 model that performs
competitively to the standard-precision W16A16 baseline. | cs.LG | Mitigating the Impact of Outlier Channels for Language Model Quantization with Activation Regularization |
2024-04-04T17:19:47Z | http://arxiv.org/pdf/2404.03602v1 | http://arxiv.org/abs/2404.03602v1 | Ryo Kamoi, Sarkar Snigdha Sarathi Das, Renze Lou, Jihyun Janice Ahn, Yilun Zhao, Xiaoxin Lu, Nan Zhang, Yusen Zhang, Ranran Haoran Zhang, Sujeeth Reddy Vummanthala, Salika Dave, Shaobo Qin, Arman Cohan, Wenpeng Yin, Rui Zhang | 2024-04-04T17:19:47Z | cs.CL | With Large Language Models (LLMs) being widely used across various tasks,
detecting errors in their responses is increasingly crucial. However, little
research has been conducted on error detection of LLM responses. Collecting
error annotations on LLM responses is challenging due to the subjective nature
of many NLP tasks, and thus previous research focuses on tasks of little
practical value (e.g., word sorting) or limited error types (e.g., faithfulness
in summarization). This work introduces ReaLMistake, the first error detection
benchmark consisting of objective, realistic, and diverse errors made by LLMs.
ReaLMistake contains three challenging and meaningful tasks that introduce
objectively assessable errors in four categories (reasoning correctness,
instruction-following, context-faithfulness, and parameterized knowledge),
eliciting naturally observed and diverse errors in responses of GPT-4 and Llama
2 70B annotated by experts. We use ReaLMistake to evaluate error detectors
based on 12 LLMs. Our findings show: 1) Top LLMs like GPT-4 and Claude 3 detect
errors made by LLMs at very low recall, and all LLM-based error detectors
perform much worse than humans. 2) Explanations by LLM-based error detectors
lack reliability. 3) LLMs-based error detection is sensitive to small changes
in prompts but remains challenging to improve. 4) Popular approaches to
improving LLMs, including self-consistency and majority vote, do not improve
the error detection performance. Our benchmark and code are provided at
https://github.com/psunlpgroup/ReaLMistake. | cs.CL | Evaluating LLMs at Detecting Errors in LLM Responses |
2024-04-04T17:09:52Z | http://arxiv.org/pdf/2404.03598v1 | http://arxiv.org/abs/2404.03598v1 | Ankan Mullick, Mukur Gupta, Pawan Goyal | 2024-04-04T17:09:52Z | cs.CL | Biomedical queries have become increasingly prevalent in web searches,
reflecting the growing interest in accessing biomedical literature. Despite
recent research on large-language models (LLMs) motivated by endeavours to
attain generalized intelligence, their efficacy in replacing task and
domain-specific natural language understanding approaches remains questionable.
In this paper, we address this question by conducting a comprehensive empirical
evaluation of intent detection and named entity recognition (NER) tasks from
biomedical text. We show that Supervised Fine Tuned approaches are still
relevant and more effective than general-purpose LLMs. Biomedical transformer
models such as PubMedBERT can surpass ChatGPT on NER task with only 5
supervised examples. | cs.CL | Intent Detection and Entity Extraction from BioMedical Literature |
2024-04-04T17:03:47Z | http://arxiv.org/pdf/2404.04295v1 | http://arxiv.org/abs/2404.04295v1 | Hainan Xu, Zhehuai Chen, Fei Jia, Boris Ginsburg | 2024-04-04T17:03:47Z | cs.CL, cs.LG, cs.SD, eess.AS | This paper proposes Transducers with Pronunciation-aware Embeddings (PET).
Unlike conventional Transducers where the decoder embeddings for different
tokens are trained independently, the PET model's decoder embedding
incorporates shared components for text tokens with the same or similar
pronunciations. With experiments conducted in multiple datasets in Mandarin
Chinese and Korean, we show that PET models consistently improve speech
recognition accuracy compared to conventional Transducers. Our investigation
also uncovers a phenomenon that we call error chain reactions. Instead of
recognition errors being evenly spread throughout an utterance, they tend to
group together, with subsequent errors often following earlier ones. Our
analysis shows that PET models effectively mitigate this issue by substantially
reducing the likelihood of the model generating additional errors following a
prior one. Our implementation will be open-sourced with the NeMo toolkit. | cs.CL | Transducers with Pronunciation-aware Embeddings for Automatic Speech Recognition |
2024-04-04T17:00:37Z | http://arxiv.org/pdf/2404.03592v3 | http://arxiv.org/abs/2404.03592v3 | Zhengxuan Wu, Aryaman Arora, Zheng Wang, Atticus Geiger, Dan Jurafsky, Christopher D. Manning, Christopher Potts | 2024-05-22T17:52:31Z | cs.CL, cs.AI, cs.LG | Parameter-efficient finetuning (PEFT) methods seek to adapt large neural
models via updates to a small number of weights. However, much prior
interpretability work has shown that representations encode rich semantic
information, suggesting that editing representations might be a more powerful
alternative. We pursue this hypothesis by developing a family of Representation
Finetuning (ReFT) methods. ReFT methods operate on a frozen base model and
learn task-specific interventions on hidden representations. We define a strong
instance of the ReFT family, Low-rank Linear Subspace ReFT (LoReFT), and we
identify an ablation of this method that trades some performance for increased
efficiency. Both are drop-in replacements for existing PEFTs and learn
interventions that are 15x--65x more parameter-efficient than LoRA. We showcase
LoReFT on eight commonsense reasoning tasks, four arithmetic reasoning tasks,
instruction-tuning, and GLUE. In all these evaluations, our ReFTs deliver the
best balance of efficiency and performance, and almost always outperform
state-of-the-art PEFTs. We release a generic ReFT training library publicly at
https://github.com/stanfordnlp/pyreft. | cs.CL | ReFT: Representation Finetuning for Language Models |
2024-04-04T16:40:11Z | http://arxiv.org/pdf/2404.03577v1 | http://arxiv.org/abs/2404.03577v1 | Yantao Liu, Zijun Yao, Xin Lv, Yuchen Fan, Shulin Cao, Jifan Yu, Lei Hou, Juanzi Li | 2024-04-04T16:40:11Z | cs.CL | Providing knowledge documents for large language models (LLMs) has emerged as
a promising solution to update the static knowledge inherent in their
parameters. However, knowledge in the document may conflict with the memory of
LLMs due to outdated or incorrect knowledge in the LLMs' parameters. This leads
to the necessity of examining the capability of LLMs to assimilate supplemental
external knowledge that conflicts with their memory. While previous studies
have explained to what extent LLMs extract conflicting knowledge from the
provided text, they neglect the necessity to reason with conflicting knowledge.
Furthermore, there lack a detailed analysis on strategies to enable LLMs to
resolve conflicting knowledge via prompting, decoding strategy, and supervised
fine-tuning. To address these limitations, we construct a new dataset, dubbed
KNOT, for knowledge conflict resolution examination in the form of question
answering. KNOT facilitates in-depth analysis by dividing reasoning with
conflicting knowledge into three levels: (1) Direct Extraction, which directly
extracts conflicting knowledge to answer questions. (2) Explicit Reasoning,
which reasons with conflicting knowledge when the reasoning path is explicitly
provided in the question. (3) Implicit Reasoning, where reasoning with
conflicting knowledge requires LLMs to infer the reasoning path independently
to answer questions. We also conduct extensive experiments on KNOT to establish
empirical guidelines for LLMs to utilize conflicting knowledge in complex
circumstances. Dataset and associated codes can be accessed at
https://github.com/THU-KEG/KNOT . | cs.CL | Untangle the KNOT: Interweaving Conflicting Knowledge and Reasoning Skills in Large Language Models |
2024-04-04T16:20:34Z | http://arxiv.org/pdf/2404.03565v2 | http://arxiv.org/abs/2404.03565v2 | Kai Zhang, Lizhi Qing, Yangyang Kang, Xiaozhong Liu | 2024-06-11T10:47:02Z | cs.CL | Large Language Models (LLMs) have exhibited remarkable proficiency in
comprehending and generating natural language. On the other hand, personalized
LLM response generation holds the potential to offer substantial benefits for
individuals in critical areas such as medical. Existing research has explored
memory-augmented methods to prompt the LLM with pre-stored user-specific
knowledge for personalized response generation in terms of new queries. We
contend that such paradigm is unable to perceive fine-granularity information.
In this study, we propose a novel \textbf{M}emory-\textbf{i}njected approach
using parameter-efficient fine-tuning (PEFT) and along with a Bayesian
Optimisation searching strategy to achieve \textbf{L}LM
\textbf{P}ersonalization(\textbf{MiLP}). | cs.CL | Personalized LLM Response Generation with Parameterized Memory Injection |
2024-04-04T16:18:37Z | http://arxiv.org/pdf/2404.03563v2 | http://arxiv.org/abs/2404.03563v2 | Regina Stodden | 2024-04-05T08:55:20Z | cs.CL | In this work, we propose EASSE-multi, a framework for easier automatic
sentence evaluation for languages other than English. Compared to the original
EASSE framework, EASSE-multi does not focus only on English. It contains
tokenizers and versions of text simplification evaluation metrics which are
suitable for multiple languages. In this paper, we exemplify the usage of
EASSE-multi for German TS, resulting in EASSE-DE. Further, we compare text
simplification results when evaluating with different language or tokenization
settings of the metrics. Based on this, we formulate recommendations on how to
make the evaluation of (German) TS models more transparent and better
comparable. The code of EASSE-multi and its German specialisation (EASSE-DE)
can be found at https://github.com/rstodden/easse-de. | cs.CL | EASSE-DE: Easier Automatic Sentence Simplification Evaluation for German |
2024-04-04T16:16:53Z | http://arxiv.org/pdf/2404.03561v1 | http://arxiv.org/abs/2404.03561v1 | Rohit Saxena, Frank Keller | 2024-04-04T16:16:53Z | cs.CL | Abstractive summarization for long-form narrative texts such as movie scripts
is challenging due to the computational and memory constraints of current
language models. A movie script typically comprises a large number of scenes;
however, only a fraction of these scenes are salient, i.e., important for
understanding the overall narrative. The salience of a scene can be
operationalized by considering it as salient if it is mentioned in the summary.
Automatically identifying salient scenes is difficult due to the lack of
suitable datasets. In this work, we introduce a scene saliency dataset that
consists of human-annotated salient scenes for 100 movies. We propose a
two-stage abstractive summarization approach which first identifies the salient
scenes in script and then generates a summary using only those scenes. Using
QA-based evaluation, we show that our model outperforms previous
state-of-the-art summarization methods and reflects the information content of
a movie more accurately than a model that takes the whole movie script as
input. | cs.CL | Select and Summarize: Scene Saliency for Movie Script Summarization |
2024-04-04T16:15:23Z | http://arxiv.org/pdf/2404.03558v1 | http://arxiv.org/abs/2404.03558v1 | Harmon Bhasin, Timothy Ossowski, Yiqiao Zhong, Junjie Hu | 2024-04-04T16:15:23Z | cs.CL, cs.LG | Large language models (LLM) have recently shown the extraordinary ability to
perform unseen tasks based on few-shot examples provided as text, also known as
in-context learning (ICL). While recent works have attempted to understand the
mechanisms driving ICL, few have explored training strategies that incentivize
these models to generalize to multiple tasks. Multi-task learning (MTL) for
generalist models is a promising direction that offers transfer learning
potential, enabling large parameterized models to be trained from simpler,
related tasks. In this work, we investigate the combination of MTL with ICL to
build models that efficiently learn tasks while being robust to
out-of-distribution examples. We propose several effective curriculum learning
strategies that allow ICL models to achieve higher data efficiency and more
stable convergence. Our experiments reveal that ICL models can effectively
learn difficult tasks by training on progressively harder tasks while mixing in
prior tasks, denoted as mixed curriculum in this work. Our code and models are
available at https://github.com/harmonbhasin/curriculum_learning_icl . | cs.CL | How does Multi-Task Training Affect Transformer In-Context Capabilities? Investigations with Function Classes |
2024-04-04T16:07:06Z | http://arxiv.org/pdf/2404.03555v2 | http://arxiv.org/abs/2404.03555v2 | Botond Barta, Dorina Lakatos, Attila Nagy, Milán Konor Nyist, Judit Ács | 2024-04-12T08:05:13Z | cs.CL | Training summarization models requires substantial amounts of training data.
However for less resourceful languages like Hungarian, openly available models
and datasets are notably scarce. To address this gap our paper introduces
HunSum-2 an open-source Hungarian corpus suitable for training abstractive and
extractive summarization models. The dataset is assembled from segments of the
Common Crawl corpus undergoing thorough cleaning, preprocessing and
deduplication. In addition to abstractive summarization we generate
sentence-level labels for extractive summarization using sentence similarity.
We train baseline models for both extractive and abstractive summarization
using the collected dataset. To demonstrate the effectiveness of the trained
models, we perform both quantitative and qualitative evaluation. Our dataset,
models and code are publicly available, encouraging replication, further
research, and real-world applications across various domains. | cs.CL | From News to Summaries: Building a Hungarian Corpus for Extractive and Abstractive Summarization |
2024-04-04T15:49:49Z | http://arxiv.org/pdf/2404.03543v2 | http://arxiv.org/abs/2404.03543v2 | Jiawei Guo, Ziming Li, Xueling Liu, Kaijing Ma, Tianyu Zheng, Zhouliang Yu, Ding Pan, Yizhi LI, Ruibo Liu, Yue Wang, Shuyue Guo, Xingwei Qu, Xiang Yue, Ge Zhang, Wenhu Chen, Jie Fu | 2024-04-06T04:29:25Z | cs.SE, cs.AI, cs.CL, cs.LG | Large Language Models (LLMs) for code are rapidly evolving, with code editing
emerging as a critical capability. We introduce CodeEditorBench, an evaluation
framework designed to rigorously assess the performance of LLMs in code editing
tasks, including debugging, translating, polishing, and requirement switching.
Unlike existing benchmarks focusing solely on code generation, CodeEditorBench
emphasizes real-world scenarios and practical aspects of software development.
We curate diverse coding challenges and scenarios from five sources, covering
various programming languages, complexity levels, and editing tasks. Evaluation
of 19 LLMs reveals that closed-source models (particularly Gemini-Ultra and
GPT-4), outperform open-source models in CodeEditorBench, highlighting
differences in model performance based on problem types and prompt
sensitivities. CodeEditorBench aims to catalyze advancements in LLMs by
providing a robust platform for assessing code editing capabilities. We will
release all prompts and datasets to enable the community to expand the dataset
and benchmark emerging LLMs. By introducing CodeEditorBench, we contribute to
the advancement of LLMs in code editing and provide a valuable resource for
researchers and practitioners. | cs.SE | CodeEditorBench: Evaluating Code Editing Capability of Large Language Models |
2024-04-04T15:36:53Z | http://arxiv.org/pdf/2404.03532v1 | http://arxiv.org/abs/2404.03532v1 | Yuchen Fan, Yantao Liu, Zijun Yao, Jifan Yu, Lei Hou, Juanzi Li | 2024-04-04T15:36:53Z | cs.CL | Modern Large Language Models (LLMs) have showcased remarkable prowess in
various tasks necessitating sophisticated cognitive behaviors. Nevertheless, a
paradoxical performance discrepancy is observed, where these models
underperform in seemingly elementary tasks like relation extraction and event
extraction due to two issues in conventional evaluation. (1) The imprecision of
existing evaluation metrics that struggle to effectively gauge semantic
consistency between model outputs and ground truth, and (2) The inherent
incompleteness of evaluation benchmarks, primarily due to restrictive human
annotation schemas, resulting in underestimated LLM performances. Inspired by
the principles in subjective question correction, we propose a new evaluation
method, SQC-Score. This method innovatively utilizes LLMs, fine-tuned through
subjective question correction data, to refine matching between model outputs
and golden labels. Additionally, by incorporating a Natural Language Inference
(NLI) model, SQC-Score enriches golden labels, addressing benchmark
incompleteness by acknowledging correct yet previously omitted answers. Results
on three information extraction tasks show that SQC-Score is more preferred by
human annotators than the baseline metrics. Utilizing SQC-Score, we conduct a
comprehensive evaluation of the state-of-the-art LLMs and provide insights for
future research for information extraction. Dataset and associated codes can be
accessed at https://github.com/THU-KEG/SQC-Score. | cs.CL | Evaluating Generative Language Models in Information Extraction as Subjective Question Correction |
2024-04-04T15:31:21Z | http://arxiv.org/pdf/2404.03528v3 | http://arxiv.org/abs/2404.03528v3 | Azmine Toushik Wasi, Taki Hasan Rafi, Raima Islam, Dong-Kyu Chae | 2024-06-05T13:39:56Z | cs.CL, cs.IR, cs.LG, cs.NE, cs.SI | Knowledge Graphs (KGs) have proven essential in information processing and
reasoning applications because they link related entities and give context-rich
information, supporting efficient information retrieval and knowledge
discovery; presenting information flow in a very effective manner. Despite
being widely used globally, Bangla is relatively underrepresented in KGs due to
a lack of comprehensive datasets, encoders, NER (named entity recognition)
models, POS (part-of-speech) taggers, and lemmatizers, hindering efficient
information processing and reasoning applications in the language. Addressing
the KG scarcity in Bengali, we propose BanglaAutoKG, a pioneering framework
that is able to automatically construct Bengali KGs from any Bangla text. We
utilize multilingual LLMs to understand various languages and correlate
entities and relations universally. By employing a translation dictionary to
identify English equivalents and extracting word features from pre-trained BERT
models, we construct the foundational KG. To reduce noise and align word
embeddings with our goal, we employ graph-based polynomial filters. Lastly, we
implement a GNN-based semantic filter, which elevates contextual understanding
and trims unnecessary edges, culminating in the formation of the definitive KG.
Empirical findings and case studies demonstrate the universal effectiveness of
our model, capable of autonomously constructing semantically enriched KGs from
any text. | cs.CL | BanglaAutoKG: Automatic Bangla Knowledge Graph Construction with Semantic Neural Graph Filtering |
2024-04-04T15:21:22Z | http://arxiv.org/pdf/2404.03514v1 | http://arxiv.org/abs/2404.03514v1 | Chengkai Huang, Rui Wang, Kaige Xie, Tong Yu, Lina Yao | 2024-04-04T15:21:22Z | cs.CL, cs.AI | Retrieval-augmented large language models (LLMs) have been remarkably
competent in various NLP tasks. Despite their great success, the knowledge
provided by the retrieval process is not always useful for improving the model
prediction, since in some samples LLMs may already be quite knowledgeable and
thus be able to answer the question correctly without retrieval. Aiming to save
the cost of retrieval, previous work has proposed to determine when to do/skip
the retrieval in a data-aware manner by analyzing the LLMs' pretraining data.
However, these data-aware methods pose privacy risks and memory limitations,
especially when requiring access to sensitive or extensive pretraining data.
Moreover, these methods offer limited adaptability under fine-tuning or
continual learning settings. We hypothesize that token embeddings are able to
capture the model's intrinsic knowledge, which offers a safer and more
straightforward way to judge the need for retrieval without the privacy risks
associated with accessing pre-training data. Moreover, it alleviates the need
to retain all the data utilized during model pre-training, necessitating only
the upkeep of the token embeddings. Extensive experiments and in-depth analyses
demonstrate the superiority of our model-aware approach. | cs.CL | Learn When (not) to Trust Language Models: A Privacy-Centric Adaptive Model-Aware Approach |
2024-04-04T14:45:26Z | http://arxiv.org/pdf/2404.03491v1 | http://arxiv.org/abs/2404.03491v1 | Jifan Yu, Xiaohan Zhang, Yifan Xu, Xuanyu Lei, Zijun Yao, Jing Zhang, Lei Hou, Juanzi Li | 2024-04-04T14:45:26Z | cs.CL, cs.AI | Empowered by the large-scale pretrained language models, existing dialogue
systems have demonstrated impressive performance conducting fluent and
natural-sounding conversations. However, they are still plagued by the
hallucination problem, causing unpredictable factual errors in the generated
responses. Recently, knowledge-grounded dialogue generation models, that
intentionally invoke external knowledge resources to more informative
responses, are also proven to be effective in reducing hallucination. Following
the idea of getting high-quality knowledge, a few efforts have achieved pretty
good performance on this issue. As some inevitable knowledge noises may also
lead to hallucinations, it is emergent to investigate the reason and future
directions for building noise-tolerant methods in KGD tasks. In this paper, we
analyze the causal story behind this problem with counterfactual reasoning
methods. Based on the causal effect analysis, we propose a possible solution
for alleviating the hallucination in KGD by exploiting the dialogue-knowledge
interaction. Experimental results of our example implementation show that this
method can reduce hallucination without disrupting other dialogue performance,
while keeping adaptive to different generation models. We hope our efforts can
support and call for more attention to developing lightweight techniques
towards robust and trusty dialogue systems. | cs.CL | A Cause-Effect Look at Alleviating Hallucination of Knowledge-grounded Dialogue Generation |
2024-04-04T14:40:07Z | http://arxiv.org/pdf/2404.03486v1 | http://arxiv.org/abs/2404.03486v1 | Jenny Pettersson, Elias Hult, Tim Eriksson, Tosin Adewumi | 2024-04-04T14:40:07Z | cs.CL | We present insightful results of a survey on the adoption of generative
artificial intelligence (GenAI) by university teachers in their teaching
activities. The transformation of education by GenAI, particularly large
language models (LLMs), has been presenting both opportunities and challenges,
including cheating by students. We prepared the online survey according to best
practices and the questions were created by the authors, who have pedagogy
experience. The survey contained 12 questions and a pilot study was first
conducted. The survey was then sent to all teachers in multiple departments
across different campuses of the university of interest in Sweden: Lule{\aa}
University of Technology. The survey was available in both Swedish and English.
The results show that 35 teachers (more than half) use GenAI out of 67
respondents. Preparation is the teaching activity with the most frequency that
GenAI is used for and ChatGPT is the most commonly used GenAI. 59% say it has
impacted their teaching, however, 55% say there should be legislation around
the use of GenAI, especially as inaccuracies and cheating are the biggest
concerns. | cs.CL | Generative AI and Teachers -- For Us or Against Us? A Case Study |
2024-04-04T14:24:06Z | http://arxiv.org/pdf/2404.03471v2 | http://arxiv.org/abs/2404.03471v2 | Farnaz Kohankhaki, Jacob-Junqi Tian, David Emerson, Laleh Seyyed-Kalantari, Faiza Khan Khattak | 2024-04-07T21:55:38Z | cs.CL, cs.CY, cs.LG | Large language models (LLMs), trained on vast datasets, can carry biases that
manifest in various forms, from overt discrimination to implicit stereotypes.
One facet of bias is performance disparities in LLMs, often harming
underprivileged groups, such as racial minorities. A common approach to
quantifying bias is to use template-based bias probes, which explicitly state
group membership (e.g. White) and evaluate if the outcome of a task, sentiment
analysis for instance, is invariant to the change of group membership (e.g.
change White race to Black). This approach is widely used in bias
quantification. However, in this work, we find evidence of an unexpectedly
overlooked consequence of using template-based probes for LLM bias
quantification. We find that in doing so, text examples associated with White
ethnicities appear to be classified as exhibiting negative sentiment at
elevated rates. We hypothesize that the scenario arises artificially through a
mismatch between the pre-training text of LLMs and the templates used to
measure bias through reporting bias, unstated norms that imply group membership
without explicit statement. Our finding highlights the potential misleading
impact of varying group membership through explicit mention in bias
quantification | cs.CL | The Impact of Unstated Norms in Bias Analysis of Language Models |
2024-04-04T13:39:06Z | http://arxiv.org/pdf/2404.03441v2 | http://arxiv.org/abs/2404.03441v2 | Sean McLeish, Avi Schwarzschild, Tom Goldstein | 2024-04-16T21:32:47Z | cs.AI, cs.CL, cs.LG | We evaluate ChatGPT's ability to solve algorithm problems from the CLRS
benchmark suite that is designed for GNNs. The benchmark requires the use of a
specified classical algorithm to solve a given problem. We find that ChatGPT
outperforms specialist GNN models, using Python to successfully solve these
problems. This raises new points in the discussion about learning algorithms
with neural networks and how we think about what out of distribution testing
looks like with web scale training data. | cs.AI | Benchmarking ChatGPT on Algorithmic Reasoning |
2024-04-04T13:36:01Z | http://arxiv.org/pdf/2404.03437v1 | http://arxiv.org/abs/2404.03437v1 | Tinatin Osmonova, Alexey Tikhonov, Ivan P. Yamshchikov | 2024-04-04T13:36:01Z | cs.CL, cs.SI | With the rise of computational social science, many scholars utilize data
analysis and natural language processing tools to analyze social media, news
articles, and other accessible data sources for examining political and social
discourse. Particularly, the study of the emergence of echo-chambers due to the
dissemination of specific information has become a topic of interest in mixed
methods research areas. In this paper, we analyze data collected from two news
portals, Breitbart News (BN) and New York Times (NYT) to prove the hypothesis
that the formation of echo-chambers can be partially explained on the level of
an individual information consumption rather than a collective topology of
individuals' social networks. Our research findings are presented through
knowledge graphs, utilizing a dataset spanning 11.5 years gathered from BN and
NYT media portals. We demonstrate that the application of knowledge
representation techniques to the aforementioned news streams highlights,
contrary to common assumptions, shows relative "internal" neutrality of both
sources and polarizing attitude towards a small fraction of entities.
Additionally, we argue that such characteristics in information sources lead to
fundamental disparities in audience worldviews, potentially acting as a
catalyst for the formation of echo-chambers. | cs.CL | Knowledge Graph Representation for Political Information Sources |
2024-04-04T13:22:28Z | http://arxiv.org/pdf/2404.03429v1 | http://arxiv.org/abs/2404.03429v1 | Zhengyuan Liu, Stella Xin Yin, Carolyn Lee, Nancy F. Chen | 2024-04-04T13:22:28Z | cs.CL | Intelligent tutoring systems (ITSs) that imitate human tutors and aim to
provide immediate and customized instructions or feedback to learners have
shown their effectiveness in education. With the emergence of generative
artificial intelligence, large language models (LLMs) further entitle the
systems to complex and coherent conversational interactions. These systems
would be of great help in language education as it involves developing skills
in communication, which, however, drew relatively less attention. Additionally,
due to the complicated cognitive development at younger ages, more endeavors
are needed for practical uses. Scaffolding refers to a teaching technique where
teachers provide support and guidance to students for learning and developing
new concepts or skills. It is an effective way to support diverse learning
needs, goals, processes, and outcomes. In this work, we investigate how
pedagogical instructions facilitate the scaffolding in ITSs, by conducting a
case study on guiding children to describe images for language learning. We
construct different types of scaffolding tutoring systems grounded in four
fundamental learning theories: knowledge construction, inquiry-based learning,
dialogic teaching, and zone of proximal development. For qualitative and
quantitative analyses, we build and refine a seven-dimension rubric to evaluate
the scaffolding process. In our experiment on GPT-4V, we observe that LLMs
demonstrate strong potential to follow pedagogical instructions and achieve
self-paced learning in different student groups. Moreover, we extend our
evaluation framework from a manual to an automated approach, paving the way to
benchmark various conversational tutoring systems. | cs.CL | Scaffolding Language Learning via Multi-modal Tutoring Systems with Pedagogical Instructions |
2024-04-04T13:15:28Z | http://arxiv.org/pdf/2404.03428v1 | http://arxiv.org/abs/2404.03428v1 | Marija Šakota, Isaac Johnson, Guosheng Feng, Robert West | 2024-04-04T13:15:28Z | cs.CL | An edit summary is a succinct comment written by a Wikipedia editor
explaining the nature of, and reasons for, an edit to a Wikipedia page. Edit
summaries are crucial for maintaining the encyclopedia: they are the first
thing seen by content moderators and help them decide whether to accept or
reject an edit. Additionally, edit summaries constitute a valuable data source
for researchers. Unfortunately, as we show, for many edits, summaries are
either missing or incomplete. To overcome this problem and help editors write
useful edit summaries, we propose a model for recommending edit summaries
generated by a language model trained to produce good edit summaries given the
representation of an edit diff. This is a challenging task for multiple
reasons, including mixed-quality training data, the need to understand not only
what was changed in the article but also why it was changed, and efficiency
requirements imposed by the scale of Wikipedia. We address these challenges by
curating a mix of human and synthetically generated training data and
fine-tuning a generative language model sufficiently small to be used on
Wikipedia at scale. Our model performs on par with human editors. Commercial
large language models are able to solve this task better than human editors,
but would be too expensive to run on Wikipedia at scale. More broadly, this
paper showcases how language modeling technology can be used to support humans
in maintaining one of the largest and most visible projects on the Web. | cs.CL | Edisum: Summarizing and Explaining Wikipedia Edits at Scale |
2024-04-04T12:46:37Z | http://arxiv.org/pdf/2404.03414v1 | http://arxiv.org/abs/2404.03414v1 | Jooyoung Lee, Fan Yang, Thanh Tran, Qian Hu, Emre Barut, Kai-Wei Chang, Chengwei Su | 2024-04-04T12:46:37Z | cs.CL, cs.AI | We introduce a novel framework, LM-Guided CoT, that leverages a lightweight
(i.e., <1B) language model (LM) for guiding a black-box large (i.e., >10B) LM
in reasoning tasks. Specifically, the lightweight LM first generates a
rationale for each input instance. The Frozen large LM is then prompted to
predict a task output based on the rationale generated by the lightweight LM.
Our approach is resource-efficient in the sense that it only requires training
the lightweight LM. We optimize the model through 1) knowledge distillation and
2) reinforcement learning from rationale-oriented and task-oriented reward
signals. We assess our method with multi-hop extractive question answering (QA)
benchmarks, HotpotQA, and 2WikiMultiHopQA. Experimental results show that our
approach outperforms all baselines regarding answer prediction accuracy. We
also find that reinforcement learning helps the model to produce higher-quality
rationales with improved QA performance. | cs.CL | Can Small Language Models Help Large Language Models Reason Better?: LM-Guided Chain-of-Thought |
2024-04-04T12:38:14Z | http://arxiv.org/pdf/2404.03411v1 | http://arxiv.org/abs/2404.03411v1 | Shuo Chen, Zhen Han, Bailan He, Zifeng Ding, Wenqian Yu, Philip Torr, Volker Tresp, Jindong Gu | 2024-04-04T12:38:14Z | cs.LG, cs.CL, cs.CR | Various jailbreak attacks have been proposed to red-team Large Language
Models (LLMs) and revealed the vulnerable safeguards of LLMs. Besides, some
methods are not limited to the textual modality and extend the jailbreak attack
to Multimodal Large Language Models (MLLMs) by perturbing the visual input.
However, the absence of a universal evaluation benchmark complicates the
performance reproduction and fair comparison. Besides, there is a lack of
comprehensive evaluation of closed-source state-of-the-art (SOTA) models,
especially MLLMs, such as GPT-4V. To address these issues, this work first
builds a comprehensive jailbreak evaluation dataset with 1445 harmful questions
covering 11 different safety policies. Based on this dataset, extensive
red-teaming experiments are conducted on 11 different LLMs and MLLMs, including
both SOTA proprietary models and open-source models. We then conduct a deep
analysis of the evaluated results and find that (1) GPT4 and GPT-4V demonstrate
better robustness against jailbreak attacks compared to open-source LLMs and
MLLMs. (2) Llama2 and Qwen-VL-Chat are more robust compared to other
open-source models. (3) The transferability of visual jailbreak methods is
relatively limited compared to textual jailbreak methods. The dataset and code
can be found here
https://anonymous.4open.science/r/red_teaming_gpt4-C1CE/README.md . | cs.LG | Red Teaming GPT-4V: Are GPT-4V Safe Against Uni/Multi-Modal Jailbreak Attacks? |
2024-04-04T11:32:03Z | http://arxiv.org/pdf/2405.01563v1 | http://arxiv.org/abs/2405.01563v1 | Yasin Abbasi Yadkori, Ilja Kuzborskij, David Stutz, András György, Adam Fisch, Arnaud Doucet, Iuliya Beloshapka, Wei-Hung Weng, Yao-Yuan Yang, Csaba Szepesvári, Ali Taylan Cemgil, Nenad Tomasev | 2024-04-04T11:32:03Z | cs.LG, cs.AI, cs.CL | We develop a principled procedure for determining when a large language model
(LLM) should abstain from responding (e.g., by saying "I don't know") in a
general domain, instead of resorting to possibly "hallucinating" a non-sensical
or incorrect answer. Building on earlier approaches that use self-consistency
as a more reliable measure of model confidence, we propose using the LLM itself
to self-evaluate the similarity between each of its sampled responses for a
given query. We then further leverage conformal prediction techniques to
develop an abstention procedure that benefits from rigorous theoretical
guarantees on the hallucination rate (error rate). Experimentally, our
resulting conformal abstention method reliably bounds the hallucination rate on
various closed-book, open-domain generative question answering datasets, while
also maintaining a significantly less conservative abstention rate on a dataset
with long responses (Temporal Sequences) compared to baselines using
log-probability scores to quantify uncertainty, while achieveing comparable
performance on a dataset with short answers (TriviaQA). To evaluate the
experiments automatically, one needs to determine if two responses are
equivalent given a question. Following standard practice, we use a thresholded
similarity function to determine if two responses match, but also provide a
method for calibrating the threshold based on conformal prediction, with
theoretical guarantees on the accuracy of the match prediction, which might be
of independent interest. | cs.LG | Mitigating LLM Hallucinations via Conformal Abstention |
2024-04-04T11:27:54Z | http://arxiv.org/pdf/2404.03381v2 | http://arxiv.org/abs/2404.03381v2 | Constanza Fierro, Reinald Kim Amplayo, Fantine Huot, Nicola De Cao, Joshua Maynez, Shashi Narayan, Mirella Lapata | 2024-05-13T15:47:57Z | cs.CL | The increasing demand for the deployment of LLMs in information-seeking
scenarios has spurred efforts in creating verifiable systems, which generate
responses to queries along with supporting evidence. In this paper, we explore
the attribution capabilities of plan-based models which have been recently
shown to improve the faithfulness, grounding, and controllability of generated
text. We conceptualize plans as a sequence of questions which serve as
blueprints of the generated content and its organization. We propose two
attribution models that utilize different variants of blueprints, an
abstractive model where questions are generated from scratch, and an extractive
model where questions are copied from the input. Experiments on long-form
question-answering show that planning consistently improves attribution
quality. Moreover, the citations generated by blueprint models are more
accurate compared to those obtained from LLM-based pipelines lacking a planning
component. | cs.CL | Learning to Plan and Generate Text with Citations |
2024-04-04T11:03:33Z | http://arxiv.org/pdf/2404.03361v1 | http://arxiv.org/abs/2404.03361v1 | Nicolay Rusnachenko, Huizhi Liang | 2024-04-04T11:03:33Z | cs.CL | Emotion expression is one of the essential traits of conversations. It may be
self-related or caused by another speaker. The variety of reasons may serve as
a source of the further emotion causes: conversation history, speaker's
emotional state, etc. Inspired by the most recent advances in Chain-of-Thought,
in this work, we exploit the existing three-hop reasoning approach (THOR) to
perform large language model instruction-tuning for answering: emotion states
(THOR-state), and emotion caused by one speaker to the other (THOR-cause). We
equip THOR-cause with the reasoning revision (rr) for devising a reasoning path
in fine-tuning. In particular, we rely on the annotated speaker emotion states
to revise reasoning path. Our final submission, based on Flan-T5-base (250M)
and the rule-based span correction technique, preliminary tuned with THOR-state
and fine-tuned with THOR-cause-rr on competition training data, results in 3rd
and 4th places (F1-proportional) and 5th place (F1-strict) among 15
participating teams. Our THOR implementation fork is publicly available:
https://github.com/nicolay-r/THOR-ECAC | cs.CL | nicolay-r at SemEval-2024 Task 3: Using Flan-T5 for Reasoning Emotion Cause in Conversations with Chain-of-Thought on Emotion States |
2024-04-04T10:45:07Z | http://arxiv.org/pdf/2404.03353v1 | http://arxiv.org/abs/2404.03353v1 | Pol G. Recasens, Yue Zhu, Chen Wang, Eun Kyung Lee, Olivier Tardieu, Alaa Youssef, Jordi Torres, Josep Ll. Berral | 2024-04-04T10:45:07Z | cs.CL | Large language models (LLMs) have revolutionized the state-of-the-art of many
different natural language processing tasks. Although serving LLMs is
computationally and memory demanding, the rise of Small Language Models (SLMs)
offers new opportunities for resource-constrained users, who now are able to
serve small models with cutting-edge performance. In this paper, we present a
set of experiments designed to benchmark SLM inference at performance and
energy levels. Our analysis provides a new perspective in serving, highlighting
that the small memory footprint of SLMs allows for reaching the Pareto-optimal
throughput within the resource capacity of a single accelerator. In this
regard, we present an initial set of findings demonstrating how model
replication can effectively improve resource utilization for serving SLMs. | cs.CL | Towards Pareto Optimal Throughput in Small Language Model Serving |
2024-04-04T10:18:03Z | http://arxiv.org/pdf/2404.03344v2 | http://arxiv.org/abs/2404.03344v2 | Juri Opitz | 2024-05-27T10:33:40Z | cs.CL | The Area Under Curve measure (AUC) seems apt to evaluate and compare diverse
models, possibly without calibration. An important example of AUC application
is the evaluation and benchmarking of models that predict faithfulness of
generated text. But we show that the AUC yields an academic and optimistic
notion of accuracy that can misalign with the actual accuracy observed in
application, yielding significant changes in benchmark rankings. To paint a
more realistic picture of downstream model performance (and prepare a model for
actual application), we explore different calibration modes, testing
calibration data and method. | cs.CL | Schroedinger's Threshold: When the AUC doesn't predict Accuracy |
2024-04-04T10:09:28Z | http://arxiv.org/pdf/2404.15320v2 | http://arxiv.org/abs/2404.15320v2 | Joan Giner-Miguelez, Abel Gómez, Jordi Cabot | 2024-05-24T11:25:49Z | cs.DL, cs.AI, cs.CL, H.4.4 | Recent regulatory initiatives like the European AI Act and relevant voices in
the Machine Learning (ML) community stress the need to describe datasets along
several key dimensions for trustworthy AI, such as the provenance processes and
social concerns. However, this information is typically presented as
unstructured text in accompanying documentation, hampering their automated
analysis and processing. In this work, we explore using large language models
(LLM) and a set of prompting strategies to automatically extract these
dimensions from documents and enrich the dataset description with them. Our
approach could aid data publishers and practitioners in creating
machine-readable documentation to improve the discoverability of their
datasets, assess their compliance with current AI regulations, and improve the
overall quality of ML models trained on them.
In this paper, we evaluate the approach on 12 scientific dataset papers
published in two scientific journals (Nature's Scientific Data and Elsevier's
Data in Brief) using two different LLMs (GPT3.5 and Flan-UL2). Results show
good accuracy with our prompt extraction strategies. Concrete results vary
depending on the dimensions, but overall, GPT3.5 shows slightly better accuracy
(81,21%) than FLAN-UL2 (69,13%) although it is more prone to hallucinations. We
have released an open-source tool implementing our approach and a replication
package, including the experiments' code and results, in an open-source
repository. | cs.DL | Using Large Language Models to Enrich the Documentation of Datasets for Machine Learning |
2024-04-04T09:48:14Z | http://arxiv.org/pdf/2404.03324v1 | http://arxiv.org/abs/2404.03324v1 | Stephen Meisenbacher, Nihildev Nandakumar, Alexandra Klymenko, Florian Matthes | 2024-04-04T09:48:14Z | cs.CL | The application of Differential Privacy to Natural Language Processing
techniques has emerged in relevance in recent years, with an increasing number
of studies published in established NLP outlets. In particular, the adaptation
of Differential Privacy for use in NLP tasks has first focused on the
$\textit{word-level}$, where calibrated noise is added to word embedding
vectors to achieve "noisy" representations. To this end, several
implementations have appeared in the literature, each presenting an alternative
method of achieving word-level Differential Privacy. Although each of these
includes its own evaluation, no comparative analysis has been performed to
investigate the performance of such methods relative to each other. In this
work, we conduct such an analysis, comparing seven different algorithms on two
NLP tasks with varying hyperparameters, including the $\textit{epsilon
($\varepsilon$)}$ parameter, or privacy budget. In addition, we provide an
in-depth analysis of the results with a focus on the privacy-utility trade-off,
as well as open-source our implementation code for further reproduction. As a
result of our analysis, we give insight into the benefits and challenges of
word-level Differential Privacy, and accordingly, we suggest concrete steps
forward for the research field. | cs.CL | A Comparative Analysis of Word-Level Metric Differential Privacy: Benchmarking The Privacy-Utility Trade-off |
2024-04-04T09:17:22Z | http://arxiv.org/pdf/2404.03312v1 | http://arxiv.org/abs/2404.03312v1 | Sayed Muddashir Hossain, Jan Alexandersson, Philipp Müller | 2024-04-04T09:17:22Z | cs.CL, cs.SD, eess.AS | Accurate utterance classification in motivational interviews is crucial to
automatically understand the quality and dynamics of client-therapist
interaction, and it can serve as a key input for systems mediating such
interactions. Motivational interviews exhibit three important characteristics.
First, there are two distinct roles, namely client and therapist. Second, they
are often highly emotionally charged, which can be expressed both in text and
in prosody. Finally, context is of central importance to classify any given
utterance. Previous works did not adequately incorporate all of these
characteristics into utterance classification approaches for mental health
dialogues. In contrast, we present M3TCM, a Multi-modal, Multi-task Context
Model for utterance classification. Our approach for the first time employs
multi-task learning to effectively model both joint and individual components
of therapist and client behaviour. Furthermore, M3TCM integrates information
from the text and speech modality as well as the conversation context. With our
novel approach, we outperform the state of the art for utterance classification
on the recently introduced AnnoMI dataset with a relative improvement of 20%
for the client- and by 15% for therapist utterance classification. In extensive
ablation studies, we quantify the improvement resulting from each contribution. | cs.CL | M3TCM: Multi-modal Multi-task Context Model for Utterance Classification in Motivational Interviews |
2024-04-04T08:56:48Z | http://arxiv.org/pdf/2404.03304v3 | http://arxiv.org/abs/2404.03304v3 | Chen Huang, Peixin Qin, Yang Deng, Wenqiang Lei, Jiancheng Lv, Tat-Seng Chua | 2024-05-06T12:44:34Z | cs.CL, cs.AI | The conversational recommendation system (CRS) has been criticized regarding
its user experience in real-world scenarios, despite recent significant
progress achieved in academia. Existing evaluation protocols for CRS may
prioritize system-centric factors such as effectiveness and fluency in
conversation while neglecting user-centric aspects. Thus, we propose a new and
inclusive evaluation protocol, Concept, which integrates both system- and
user-centric factors. We conceptualise three key characteristics in
representing such factors and further divide them into six primary abilities.
To implement Concept, we adopt a LLM-based user simulator and evaluator with
scoring rubrics that are tailored for each primary ability. Our protocol,
Concept, serves a dual purpose. First, it provides an overview of the pros and
cons in current CRS models. Second, it pinpoints the problem of low usability
in the "omnipotent" ChatGPT and offers a comprehensive reference guide for
evaluating CRS, thereby setting the foundation for CRS improvement. | cs.CL | Concept -- An Evaluation Protocol on Conversational Recommender Systems with System-centric and User-centric Factors |
2024-04-04T08:52:30Z | http://arxiv.org/pdf/2404.03302v1 | http://arxiv.org/abs/2404.03302v1 | Siye Wu, Jian Xie, Jiangjie Chen, Tinghui Zhu, Kai Zhang, Yanghua Xiao | 2024-04-04T08:52:30Z | cs.CL | By leveraging the retrieval of information from external knowledge databases,
Large Language Models (LLMs) exhibit enhanced capabilities for accomplishing
many knowledge-intensive tasks. However, due to the inherent flaws of current
retrieval systems, there might exist irrelevant information within those
retrieving top-ranked passages. In this work, we present a comprehensive
investigation into the robustness of LLMs to different types of irrelevant
information under various conditions. We initially introduce a framework to
construct high-quality irrelevant information that ranges from semantically
unrelated, partially related, and related to questions. Furthermore, our
analysis demonstrates that the constructed irrelevant information not only
scores highly on similarity metrics, being highly retrieved by existing
systems, but also bears semantic connections to the context. Our investigation
reveals that current LLMs still face challenges in discriminating highly
semantically related information and can be easily distracted by these
irrelevant yet misleading contents. Besides, we also find that current
solutions for handling irrelevant information have limitations in improving the
robustness of LLMs to such distractions. Resources are available at
https://github.com/Di-viner/LLM-Robustness-to-Irrelevant-Information. | cs.CL | How Easily do Irrelevant Inputs Skew the Responses of Large Language Models? |
2024-04-04T08:52:25Z | http://arxiv.org/pdf/2404.03301v1 | http://arxiv.org/abs/2404.03301v1 | Fangru Lin, Daniel Altshuler, Janet B. Pierrehumbert | 2024-04-04T08:52:25Z | cs.CL | Scalar adjectives pertain to various domain scales and vary in intensity
within each scale (e.g. certain is more intense than likely on the likelihood
scale). Scalar implicatures arise from the consideration of alternative
statements which could have been made. They can be triggered by scalar
adjectives and require listeners to reason pragmatically about them. Some
scalar adjectives are more likely to trigger scalar implicatures than others.
This phenomenon is referred to as scalar diversity. In this study, we probe
different families of Large Language Models such as GPT-4 for their knowledge
of the lexical semantics of scalar adjectives and one specific aspect of their
pragmatics, namely scalar diversity. We find that they encode rich
lexical-semantic information about scalar adjectives. However, the rich
lexical-semantic knowledge does not entail a good understanding of scalar
diversity. We also compare current models of different sizes and complexities
and find that larger models are not always better. Finally, we explain our
probing results by leveraging linguistic intuitions and model training
objectives. | cs.CL | Probing Large Language Models for Scalar Adjective Lexical Semantics and Scalar Diversity Pragmatics |
2024-04-04T08:38:03Z | http://arxiv.org/pdf/2404.04293v1 | http://arxiv.org/abs/2404.04293v1 | Yanda Li, Dixuan Wang, Jiaqing Liang, Guochao Jiang, Qianyu He, Yanghua Xiao, Deqing Yang | 2024-04-04T08:38:03Z | cs.CL, cs.AI | Large Language Models (LLMs) have demonstrated good performance in many
reasoning tasks, but they still struggle with some complicated reasoning tasks
including logical reasoning. One non-negligible reason for LLMs' suboptimal
performance on logical reasoning is their overlooking of understanding logical
fallacies correctly. To evaluate LLMs' capability of logical fallacy
understanding (LFU), we propose five concrete tasks from three cognitive
dimensions of WHAT, WHY, and HOW in this paper. Towards these LFU tasks, we
have successfully constructed a new dataset LFUD based on GPT-4 accompanied by
a little human effort. Our extensive experiments justify that our LFUD can be
used not only to evaluate LLMs' LFU capability, but also to fine-tune LLMs to
obtain significantly enhanced performance on logical reasoning. | cs.CL | Reason from Fallacy: Enhancing Large Language Models' Logical Reasoning through Logical Fallacy Understanding |
2024-04-04T08:04:24Z | http://arxiv.org/pdf/2404.03278v1 | http://arxiv.org/abs/2404.03278v1 | Liam Cripwell, Joël Legrand, Claire Gardent | 2024-04-04T08:04:24Z | cs.CL | Text simplification intends to make a text easier to read while preserving
its core meaning. Intuitively and as shown in previous works, these two
dimensions (simplification and meaning preservation) are often-times inversely
correlated. An overly conservative text will fail to simplify sufficiently,
whereas extreme simplification will degrade meaning preservation. Yet, popular
evaluation metrics either aggregate meaning preservation and simplification
into a single score (SARI, LENS), or target meaning preservation alone
(BERTScore, QuestEval). Moreover, these metrics usually require a set of
references and most previous work has only focused on sentence-level
simplification. In this paper, we focus on the evaluation of document-level
text simplification and compare existing models using distinct metrics for
meaning preservation and simplification. We leverage existing metrics from
similar tasks and introduce a reference-less metric variant for simplicity,
showing that models are mostly biased towards either simplification or meaning
preservation, seldom performing well on both dimensions. Making use of the fact
that the metrics we use are all reference-less, we also investigate the
performance of existing models when applied to unseen data (where reference
simplifications are unavailable). | cs.CL | Evaluating Document Simplification: On the Importance of Separately Assessing Simplicity and Meaning Preservation |
2024-04-04T07:31:56Z | http://arxiv.org/pdf/2404.03259v2 | http://arxiv.org/abs/2404.03259v2 | Chen Li, Huidong Tang, Peng Ju, Debo Cheng, Yasuhiko Morimoto | 2024-04-19T07:30:23Z | cs.CL, cs.AI | Aspect-based sentiment analysis aims to predict sentiment polarity with fine
granularity. While Graph Convolutional Networks (GCNs) are widely utilized for
sentimental feature extraction, their naive application for syntactic feature
extraction can compromise information preservation. This study introduces an
innovative edge-enhanced GCN, named SentiSys, to navigate the syntactic graph
while preserving intact feature information, leading to enhanced performance.
Specifically,we first integrate a bidirectional long short-term memory
(Bi-LSTM) network and a self-attention-based transformer. This combination
facilitates effective text encoding, preventing the loss of information and
predicting long dependency text. A bidirectional GCN (Bi-GCN) with message
passing is then employed to encode relationships between entities.
Additionally, unnecessary information is filtered out using an aspect-specific
masking technique. To validate the effectiveness of our proposed model, we
conduct extensive evaluation experiments and ablation studies on four benchmark
datasets. The results consistently demonstrate improved performance in
aspect-based sentiment analysis when employing SentiSys. This approach
successfully addresses the challenges associated with syntactic feature
extraction, highlighting its potential for advancing sentiment analysis
methodologies. | cs.CL | Enhancing the Performance of Aspect-Based Sentiment Analysis Systems |
2024-04-04T06:16:35Z | http://arxiv.org/pdf/2404.04292v5 | http://arxiv.org/abs/2404.04292v5 | Zhoujian Sun, Cheng Luo, Ziyi Liu, Zhengxing Huang | 2024-05-20T00:45:40Z | cs.CL, cs.AI | The development of large language models (LLMs) has brought unprecedented
possibilities for artificial intelligence (AI) based medical diagnosis.
However, the application perspective of LLMs in real diagnostic scenarios is
still unclear because they are not adept at collecting patient data
proactively. This study presents a LLM-based diagnostic system that enhances
planning capabilities by emulating doctors. Our system involves two external
planners to handle planning tasks. The first planner employs a reinforcement
learning approach to formulate disease screening questions and conduct initial
diagnoses. The second planner uses LLMs to parse medical guidelines and conduct
differential diagnoses. By utilizing real patient electronic medical record
data, we constructed simulated dialogues between virtual patients and doctors
and evaluated the diagnostic abilities of our system. We demonstrated that our
system obtained impressive performance in both disease screening and
differential diagnoses tasks. This research represents a step towards more
seamlessly integrating AI into clinical settings, potentially enhancing the
accuracy and accessibility of medical diagnostics. | cs.CL | Conversational Disease Diagnosis via External Planner-Controlled Large Language Models |
2024-04-04T05:15:07Z | http://arxiv.org/pdf/2404.03204v3 | http://arxiv.org/abs/2404.03204v3 | Detai Xin, Xu Tan, Kai Shen, Zeqian Ju, Dongchao Yang, Yuancheng Wang, Shinnosuke Takamichi, Hiroshi Saruwatari, Shujie Liu, Jinyu Li, Sheng Zhao | 2024-05-19T21:34:28Z | eess.AS, cs.AI, cs.CL, cs.LG, cs.SD | We present RALL-E, a robust language modeling method for text-to-speech (TTS)
synthesis. While previous work based on large language models (LLMs) shows
impressive performance on zero-shot TTS, such methods often suffer from poor
robustness, such as unstable prosody (weird pitch and rhythm/duration) and a
high word error rate (WER), due to the autoregressive prediction style of
language models. The core idea behind RALL-E is chain-of-thought (CoT)
prompting, which decomposes the task into simpler steps to enhance the
robustness of LLM-based TTS. To accomplish this idea, RALL-E first predicts
prosody features (pitch and duration) of the input text and uses them as
intermediate conditions to predict speech tokens in a CoT style. Second, RALL-E
utilizes the predicted duration prompt to guide the computing of self-attention
weights in Transformer to enforce the model to focus on the corresponding
phonemes and prosody features when predicting speech tokens. Results of
comprehensive objective and subjective evaluations demonstrate that, compared
to a powerful baseline method VALL-E, RALL-E significantly improves the WER of
zero-shot TTS from $5.6\%$ (without reranking) and $1.7\%$ (with reranking) to
$2.5\%$ and $1.0\%$, respectively. Furthermore, we demonstrate that RALL-E
correctly synthesizes sentences that are hard for VALL-E and reduces the error
rate from $68\%$ to $4\%$. | eess.AS | RALL-E: Robust Codec Language Modeling with Chain-of-Thought Prompting for Text-to-Speech Synthesis |
2024-04-04T04:49:46Z | http://arxiv.org/pdf/2404.03196v1 | http://arxiv.org/abs/2404.03196v1 | Abhijnan Nath, Shadi Manafi, Avyakta Chelle, Nikhil Krishnaswamy | 2024-04-04T04:49:46Z | cs.CL | In NLP, Event Coreference Resolution (ECR) is the task of connecting event
clusters that refer to the same underlying real-life event, usually via neural
systems. In this work, we investigate using abductive free-text rationales
(FTRs) generated by modern autoregressive LLMs as distant supervision of
smaller student models for cross-document coreference (CDCR) of events. We
implement novel rationale-oriented event clustering and knowledge distillation
methods for event coreference scoring that leverage enriched information from
the FTRs for improved CDCR without additional annotation or expensive document
clustering. Our model using coreference specific knowledge distillation
achieves SOTA B3 F1 on the ECB+ and GVC corpora and we establish a new baseline
on the AIDA Phase 1 corpus. Our code can be found at
https://github.com/csu-signal/llama_cdcr | cs.CL | Okay, Let's Do This! Modeling Event Coreference with Generated Rationales and Knowledge Distillation |
2024-04-04T04:23:19Z | http://arxiv.org/pdf/2404.03192v2 | http://arxiv.org/abs/2404.03192v2 | Yuan Wang, Xuyang Wu, Hsin-Tai Wu, Zhiqiang Tao, Yi Fang | 2024-06-25T20:54:16Z | cs.IR, cs.CL | The integration of Large Language Models (LLMs) in information retrieval has
raised a critical reevaluation of fairness in the text-ranking models. LLMs,
such as GPT models and Llama2, have shown effectiveness in natural language
understanding tasks, and prior works (e.g., RankGPT) have also demonstrated
that the LLMs exhibit better performance than the traditional ranking models in
the ranking task. However, their fairness remains largely unexplored. This
paper presents an empirical study evaluating these LLMs using the TREC Fair
Ranking dataset, focusing on the representation of binary protected attributes
such as gender and geographic location, which are historically underrepresented
in search outcomes. Our analysis delves into how these LLMs handle queries and
documents related to these attributes, aiming to uncover biases in their
ranking algorithms. We assess fairness from both user and content perspectives,
contributing an empirical benchmark for evaluating LLMs as the fair ranker. | cs.IR | Do Large Language Models Rank Fairly? An Empirical Study on the Fairness of LLMs as Rankers |
2024-04-04T04:20:04Z | http://arxiv.org/pdf/2404.03189v2 | http://arxiv.org/abs/2404.03189v2 | Noah Y. Siegel, Oana-Maria Camburu, Nicolas Heess, Maria Perez-Ortiz | 2024-06-07T11:54:44Z | cs.CL, cs.AI | In order to oversee advanced AI systems, it is important to understand their
underlying decision-making process. When prompted, large language models (LLMs)
can provide natural language explanations or reasoning traces that sound
plausible and receive high ratings from human annotators. However, it is
unclear to what extent these explanations are faithful, i.e., truly capture the
factors responsible for the model's predictions. In this work, we introduce
Correlational Explanatory Faithfulness (CEF), a metric that can be used in
faithfulness tests based on input interventions. Previous metrics used in such
tests take into account only binary changes in the predictions. Our metric
accounts for the total shift in the model's predicted label distribution, more
accurately reflecting the explanations' faithfulness. We then introduce the
Correlational Counterfactual Test (CCT) by instantiating CEF on the
Counterfactual Test (CT) from Atanasova et al. (2023). We evaluate the
faithfulness of free-text explanations generated by few-shot-prompted LLMs from
the Llama2 family on three NLP tasks. We find that our metric measures aspects
of faithfulness which the CT misses. | cs.CL | The Probabilities Also Matter: A More Faithful Metric for Faithfulness of Free-Text Explanations in Large Language Models |
2024-04-04T03:50:34Z | http://arxiv.org/pdf/2404.03184v1 | http://arxiv.org/abs/2404.03184v1 | Jiawei Li, Yue Zhang | 2024-04-04T03:50:34Z | cs.CL, cs.AI | Machine reading comprehension is an essential natural language processing
task, which takes into a pair of context and query and predicts the
corresponding answer to query. In this project, we developed an end-to-end
question answering model incorporating BERT and additional linguistic features.
We conclude that the BERT base model will be improved by incorporating the
features. The EM score and F1 score are improved 2.17 and 2.14 compared with
BERT(base). Our best single model reaches EM score 76.55 and F1 score 79.97 in
the hidden test set. Our error analysis also shows that the linguistic
architecture can help model understand the context better in that it can locate
answers that BERT only model predicted "No Answer" wrongly. | cs.CL | The Death of Feature Engineering? BERT with Linguistic Features on SQuAD 2.0 |
2024-04-04T02:58:21Z | http://arxiv.org/pdf/2404.04287v1 | http://arxiv.org/abs/2404.04287v1 | Pouria Rouzrokh, Shahriar Faghani, Cooper U. Gamble, Moein Shariatnia, Bradley J. Erickson | 2024-04-04T02:58:21Z | cs.CL, cs.AI | Retrieval-augmented generation (RAG) frameworks enable large language models
(LLMs) to retrieve relevant information from a knowledge base and incorporate
it into the context for generating responses. This mitigates hallucinations and
allows for the updating of knowledge without retraining the LLM. However, RAG
does not guarantee valid responses if retrieval fails to identify the necessary
information as the context for response generation. Also, if there is
contradictory content, the RAG response will likely reflect only one of the two
possible responses. Therefore, quantifying uncertainty in the retrieval process
is crucial for ensuring RAG trustworthiness. In this report, we introduce a
four-step framework for applying conformal prediction to quantify retrieval
uncertainty in RAG frameworks. First, a calibration set of questions answerable
from the knowledge base is constructed. Each question's embedding is compared
against document embeddings to identify the most relevant document chunks
containing the answer and record their similarity scores. Given a
user-specified error rate ({\alpha}), these similarity scores are then analyzed
to determine a similarity score cutoff threshold. During inference, all chunks
with similarity exceeding this threshold are retrieved to provide context to
the LLM, ensuring the true answer is captured in the context with a
(1-{\alpha}) confidence level. We provide a Python package that enables users
to implement the entire workflow proposed in our work, only using LLMs and
without human intervention. | cs.CL | CONFLARE: CONFormal LArge language model REtrieval |
2024-04-04T02:31:05Z | http://arxiv.org/pdf/2404.03163v1 | http://arxiv.org/abs/2404.03163v1 | Xinmeng Huang, Shuo Li, Mengxin Yu, Matteo Sesia, Hamed Hassani, Insup Lee, Osbert Bastani, Edgar Dobriban | 2024-04-04T02:31:05Z | cs.CL, cs.AI, cs.LG, stat.ML | Language Models (LMs) have shown promising performance in natural language
generation. However, as LMs often generate incorrect or hallucinated responses,
it is crucial to correctly quantify their uncertainty in responding to given
inputs. In addition to verbalized confidence elicited via prompting, many
uncertainty measures ($e.g.$, semantic entropy and affinity-graph-based
measures) have been proposed. However, these measures can differ greatly, and
it is unclear how to compare them, partly because they take values over
different ranges ($e.g.$, $[0,\infty)$ or $[0,1]$). In this work, we address
this issue by developing a novel and practical framework, termed
$Rank$-$Calibration$, to assess uncertainty and confidence measures for LMs.
Our key tenet is that higher uncertainty (or lower confidence) should imply
lower generation quality, on average. Rank-calibration quantifies deviations
from this ideal relationship in a principled manner, without requiring ad hoc
binary thresholding of the correctness score ($e.g.$, ROUGE or METEOR). The
broad applicability and the granular interpretability of our methods are
demonstrated empirically. | cs.CL | Uncertainty in Language Models: Assessment through Rank-Calibration |
2024-04-04T02:22:37Z | http://arxiv.org/pdf/2404.03161v1 | http://arxiv.org/abs/2404.03161v1 | Taichi Nishimura, Koki Yamamoto, Yuto Haneji, Keiya Kajimura, Chihiro Nishiwaki, Eriko Daikoku, Natsuko Okuda, Fumihito Ono, Hirotaka Kameko, Shinsuke Mori | 2024-04-04T02:22:37Z | cs.CV, cs.CL, cs.MM | This paper introduces a biochemical vision-and-language dataset, which
consists of 24 egocentric experiment videos, corresponding protocols, and
video-and-language alignments. The key challenge in the wet-lab domain is
detecting equipment, reagents, and containers is difficult because the lab
environment is scattered by filling objects on the table and some objects are
indistinguishable. Therefore, previous studies assume that objects are manually
annotated and given for downstream tasks, but this is costly and
time-consuming. To address this issue, this study focuses on Micro QR Codes to
detect objects automatically. From our preliminary study, we found that
detecting objects only using Micro QR Codes is still difficult because the
researchers manipulate objects, causing blur and occlusion frequently. To
address this, we also propose a novel object labeling method by combining a
Micro QR Code detector and an off-the-shelf hand object detector. As one of the
applications of our dataset, we conduct the task of generating protocols from
experiment videos and find that our approach can generate accurate protocols. | cs.CV | BioVL-QR: Egocentric Biochemical Video-and-Language Dataset Using Micro QR Codes |
2024-04-04T02:01:25Z | http://arxiv.org/pdf/2404.04286v1 | http://arxiv.org/abs/2404.04286v1 | Yi Ren, Shangmin Guo, Linlu Qiu, Bailin Wang, Danica J. Sutherland | 2024-04-04T02:01:25Z | cs.CL, cs.AI, cs.LG | With the widespread adoption of Large Language Models (LLMs), the prevalence
of iterative interactions among these models is anticipated to increase.
Notably, recent advancements in multi-round self-improving methods allow LLMs
to generate new examples for training subsequent models. At the same time,
multi-agent LLM systems, involving automated interactions among agents, are
also increasing in prominence. Thus, in both short and long terms, LLMs may
actively engage in an evolutionary process. We draw parallels between the
behavior of LLMs and the evolution of human culture, as the latter has been
extensively studied by cognitive scientists for decades. Our approach involves
leveraging Iterated Learning (IL), a Bayesian framework that elucidates how
subtle biases are magnified during human cultural evolution, to explain some
behaviors of LLMs. This paper outlines key characteristics of agents' behavior
in the Bayesian-IL framework, including predictions that are supported by
experimental verification with various LLMs. This theoretical framework could
help to more effectively predict and guide the evolution of LLMs in desired
directions. | cs.CL | Language Model Evolution: An Iterated Learning Perspective |
2024-04-04T01:50:20Z | http://arxiv.org/pdf/2404.03150v1 | http://arxiv.org/abs/2404.03150v1 | Anish Pahilajani, Samyak Rajesh Jain, Devasha Trivedi | 2024-04-04T01:50:20Z | cs.CL, cs.AI | This paper presents our submission to the SemEval 2024 Task 5: The Legal
Argument Reasoning Task in Civil Procedure. We present two approaches to
solving the task of legal answer validation, given an introduction to the case,
a question and an answer candidate. Firstly, we fine-tuned pre-trained
BERT-based models and found that models trained on domain knowledge perform
better. Secondly, we performed few-shot prompting on GPT models and found that
reformulating the answer validation task to be a multiple-choice QA task
remarkably improves the performance of the model. Our best submission is a
BERT-based model that achieved the 7th place out of 20. | cs.CL | NLP at UC Santa Cruz at SemEval-2024 Task 5: Legal Answer Validation using Few-Shot Multi-Choice QA |
2024-04-04T01:07:14Z | http://arxiv.org/pdf/2404.03134v2 | http://arxiv.org/abs/2404.03134v2 | Vagrant Gautam, Eileen Bingert, Dawei Zhu, Anne Lauscher, Dietrich Klakow | 2024-05-01T21:35:00Z | cs.CL, cs.CY | Robust, faithful and harm-free pronoun use for individuals is an important
goal for language models as their use increases, but prior work tends to study
only one or two of these characteristics at a time. To measure progress towards
the combined goal, we introduce the task of pronoun fidelity: given a context
introducing a co-referring entity and pronoun, the task is to reuse the correct
pronoun later. We present RUFF, a carefully-designed dataset of over 5 million
instances to measure robust pronoun fidelity in English, and we evaluate 37
popular large language models across architectures (encoder-only, decoder-only
and encoder-decoder) and scales (11M-70B parameters). When an individual is
introduced with a pronoun, models can mostly faithfully reuse this pronoun in
the next sentence, but they are significantly worse with she/her/her, singular
they and neopronouns. Moreover, models are easily distracted by non-adversarial
sentences discussing other people; even one additional sentence with a
distractor pronoun causes accuracy to drop on average by 34%. Our results show
that pronoun fidelity is neither robust, nor due to reasoning, in a simple,
naturalistic setting where humans achieve nearly 100% accuracy. We encourage
researchers to bridge the gaps we find and to carefully evaluate reasoning in
settings where superficial repetition might inflate perceptions of model
performance. | cs.CL | Robust Pronoun Fidelity with English LLMs: Are they Reasoning, Repeating, or Just Biased? |
2024-04-03T23:42:38Z | http://arxiv.org/pdf/2404.04285v1 | http://arxiv.org/abs/2404.04285v1 | Chunyuan Deng, Xiangru Tang, Yilun Zhao, Hanming Wang, Haoran Wang, Wangchunshu Zhou, Arman Cohan, Mark Gerstein | 2024-04-03T23:42:38Z | cs.CL, cs.AI | Recently, large language models (LLMs) have evolved into interactive agents,
proficient in planning, tool use, and task execution across a wide variety of
tasks. However, without specific agent tuning, open-source models like LLaMA
currently struggle to match the efficiency of GPT- 4, particularly given the
scarcity of agent-tuning datasets for fine-tuning. In response, we introduce
\textsc{Mimir}: a streamlined platform offering a customizable pipeline that
enables users to leverage both private knowledge and publicly available,
legally compliant datasets at scale for \textbf{personalized agent tuning}.
Additionally, \textsc{Mimir} supports the generation of general
instruction-tuning datasets from the same input. This dual capability ensures
that language agents developed through the platform possess both specific agent
abilities and general competencies. \textsc{Mimir} integrates these features
into a cohesive end-to-end platform, facilitating everything from the uploading
of personalized files to one-click agent fine-tuning. | cs.CL | MIMIR: A Streamlined Platform for Personalized Agent Tuning in Domain Expertise |
2024-04-03T23:33:56Z | http://arxiv.org/pdf/2404.03114v1 | http://arxiv.org/abs/2404.03114v1 | William Macke, Michael Doyle | 2024-04-03T23:33:56Z | cs.SE, cs.AI, cs.CL | Large Language Models (LLMs) have demonstrated impressive abilities in recent
years with regards to code generation and understanding. However, little work
has investigated how documentation and other code properties affect an LLM's
ability to understand and generate code or documentation. We present an
empirical analysis of how underlying properties of code or documentation can
affect an LLM's capabilities. We show that providing an LLM with "incorrect"
documentation can greatly hinder code understanding, while incomplete or
missing documentation does not seem to significantly affect an LLM's ability to
understand code. | cs.SE | Testing the Effect of Code Documentation on Large Language Model Code Understanding |
2024-04-03T22:39:33Z | http://arxiv.org/pdf/2404.03098v1 | http://arxiv.org/abs/2404.03098v1 | Lucas E. Resck, Marcos M. Raimundo, Jorge Poco | 2024-04-03T22:39:33Z | cs.CL, cs.AI, cs.LG | Saliency post-hoc explainability methods are important tools for
understanding increasingly complex NLP models. While these methods can reflect
the model's reasoning, they may not align with human intuition, making the
explanations not plausible. In this work, we present a methodology for
incorporating rationales, which are text annotations explaining human
decisions, into text classification models. This incorporation enhances the
plausibility of post-hoc explanations while preserving their faithfulness. Our
approach is agnostic to model architectures and explainability methods. We
introduce the rationales during model training by augmenting the standard
cross-entropy loss with a novel loss function inspired by contrastive learning.
By leveraging a multi-objective optimization algorithm, we explore the
trade-off between the two loss functions and generate a Pareto-optimal frontier
of models that balance performance and plausibility. Through extensive
experiments involving diverse models, datasets, and explainability methods, we
demonstrate that our approach significantly enhances the quality of model
explanations without causing substantial (sometimes negligible) degradation in
the original model's performance. | cs.CL | Exploring the Trade-off Between Model Performance and Explanation Plausibility of Text Classifiers Using Human Rationales |
2024-04-03T22:13:04Z | http://arxiv.org/pdf/2404.03092v1 | http://arxiv.org/abs/2404.03092v1 | Catherine Henry, Casey Kennington | 2024-04-03T22:13:04Z | cs.CL, cs.RO | Towards addressing the Symbol Grounding Problem and motivated by early
childhood language development, we leverage a robot which has been equipped
with an approximate model of curiosity with particular focus on bottom-up
building of unsupervised categories grounded in the physical world. That is,
rather than starting with a top-down symbol (e.g., a word referring to an
object) and providing meaning through the application of predetermined samples,
the robot autonomously and gradually breaks up its exploration space into a
series of increasingly specific unlabeled categories at which point an external
expert may optionally provide a symbol association. We extend prior work by
using a robot that can observe the visual world, introducing a higher
dimensional sensory space, and using a more generalizable method of category
building. Our experiments show that the robot learns categories based on
actions and what it visually observes, and that those categories can be
symbolically grounded into.https://info.arxiv.org/help/prep#comments | cs.CL | Unsupervised, Bottom-up Category Discovery for Symbol Grounding with a Curious Robot |
2024-04-03T22:01:26Z | http://arxiv.org/pdf/2404.03086v1 | http://arxiv.org/abs/2404.03086v1 | Johann D. Gaebler, Sharad Goel, Aziz Huq, Prasanna Tambe | 2024-04-03T22:01:26Z | stat.AP, cs.CL | Regulatory efforts to protect against algorithmic bias have taken on
increased urgency with rapid advances in large language models (LLMs), which
are machine learning models that can achieve performance rivaling human experts
on a wide array of tasks. A key theme of these initiatives is algorithmic
"auditing," but current regulations -- as well as the scientific literature --
provide little guidance on how to conduct these assessments. Here we propose
and investigate one approach for auditing algorithms: correspondence
experiments, a widely applied tool for detecting bias in human judgements. In
the employment context, correspondence experiments aim to measure the extent to
which race and gender impact decisions by experimentally manipulating elements
of submitted application materials that suggest an applicant's demographic
traits, such as their listed name. We apply this method to audit candidate
assessments produced by several state-of-the-art LLMs, using a novel corpus of
applications to K-12 teaching positions in a large public school district. We
find evidence of moderate race and gender disparities, a pattern largely robust
to varying the types of application material input to the models, as well as
the framing of the task to the LLMs. We conclude by discussing some important
limitations of correspondence experiments for auditing algorithms. | stat.AP | Auditing the Use of Language Models to Guide Hiring Decisions |
2024-04-03T21:46:14Z | http://arxiv.org/pdf/2404.03080v2 | http://arxiv.org/abs/2404.03080v2 | Yanpeng Ye, Jie Ren, Shaozhou Wang, Yuwei Wan, Haofen Wang, Imran Razzak, Tong Xie, Wenjie Zhang | 2024-06-04T10:04:05Z | cs.CL, cs.AI | Knowledge in materials science is widely dispersed across extensive
scientific literature, posing significant challenges for efficient discovery
and integration of new materials. Traditional methods, often reliant on costly
and time-consuming experimental approaches, further complicate rapid
innovation. Addressing these challenges, the integration of artificial
intelligence with materials science has opened avenues for accelerating the
discovery process, though it also demands precise annotation, data extraction,
and traceability of information. To tackle these issues, this article
introduces the Materials Knowledge Graph (MKG), which utilizes advanced natural
language processing techniques, integrated with large language models to
extract and systematically organize a decade's worth of high-quality research
into structured triples, contains 162,605 nodes and 731,772 edges. MKG
categorizes information into comprehensive labels such as Name, Formula, and
Application, structured around a meticulously designed ontology, thus enhancing
data usability and integration. By implementing network-based algorithms, MKG
not only facilitates efficient link prediction but also significantly reduces
reliance on traditional experimental methods. This structured approach not only
streamlines materials research but also lays the groundwork for more
sophisticated science knowledge graphs. | cs.CL | Construction and Application of Materials Knowledge Graph in Multidisciplinary Materials Science via Large Language Model |
2024-04-03T21:29:40Z | http://arxiv.org/pdf/2404.03073v1 | http://arxiv.org/abs/2404.03073v1 | Kaavya Chaparala, Guido Zarrella, Bruce Torres Fischer, Larry Kimura, Oiwi Parker Jones | 2024-04-03T21:29:40Z | cs.CL, cs.LG, cs.SD, eess.AS | In this paper we address the challenge of improving Automatic Speech
Recognition (ASR) for a low-resource language, Hawaiian, by incorporating large
amounts of independent text data into an ASR foundation model, Whisper. To do
this, we train an external language model (LM) on ~1.5M words of Hawaiian text.
We then use the LM to rescore Whisper and compute word error rates (WERs) on a
manually curated test set of labeled Hawaiian data. As a baseline, we use
Whisper without an external LM. Experimental results reveal a small but
significant improvement in WER when ASR outputs are rescored with a Hawaiian
LM. The results support leveraging all available data in the development of ASR
systems for underrepresented languages. | cs.CL | Mai Ho'omāuna i ka 'Ai: Language Models Improve Automatic Speech Recognition in Hawaiian |
2024-04-03T20:35:36Z | http://arxiv.org/pdf/2404.03052v1 | http://arxiv.org/abs/2404.03052v1 | Ali Pesaranghader, Nikhil Verma, Manasa Bharadwaj | 2024-04-03T20:35:36Z | cs.CL | Harmful and offensive communication or content is detrimental to social
bonding and the mental state of users on social media platforms. Text
detoxification is a crucial task in natural language processing (NLP), where
the goal is removing profanity and toxicity from text while preserving its
content. Supervised and unsupervised learning are common approaches for
designing text detoxification solutions. However, these methods necessitate
fine-tuning, leading to computational overhead. In this paper, we propose
GPT-DETOX as a framework for prompt-based in-context learning for text
detoxification using GPT-3.5 Turbo. We utilize zero-shot and few-shot prompting
techniques for detoxifying input sentences. To generate few-shot prompts, we
propose two methods: word-matching example selection (WMES) and
context-matching example selection (CMES). We additionally take into account
ensemble in-context learning (EICL) where the ensemble is shaped by base
prompts from zero-shot and all few-shot settings. We use ParaDetox and APPDIA
as benchmark detoxification datasets. Our experimental results show that the
zero-shot solution achieves promising performance, while our best few-shot
setting outperforms the state-of-the-art models on ParaDetox and shows
comparable results on APPDIA. Our EICL solutions obtain the greatest
performance, adding at least 10% improvement, against both datasets. | cs.CL | GPT-DETOX: An In-Context Learning-Based Paraphraser for Text Detoxification |
2024-04-03T20:30:38Z | http://arxiv.org/pdf/2404.03049v1 | http://arxiv.org/abs/2404.03049v1 | Johnathan E. Avery | 2024-04-03T20:30:38Z | cs.CL | This paper explores the integration of linguistic inputs within robotic
navigation systems, drawing upon the symbol interdependency hypothesis to
bridge the divide between symbolic and embodied cognition. It examines previous
work incorporating language and semantics into Neural Network (NN) and
Simultaneous Localization and Mapping (SLAM) approaches, highlighting how these
integrations have advanced the field. By contrasting abstract symbol
manipulation with sensory-motor grounding, we propose a unified framework where
language functions both as an abstract communicative system and as a grounded
representation of perceptual experiences. Our review of cognitive models of
distributional semantics and their application to autonomous agents underscores
the transformative potential of language-integrated systems. | cs.CL | Language, Environment, and Robotic Navigation |
2024-04-03T20:29:40Z | http://arxiv.org/pdf/2404.03048v2 | http://arxiv.org/abs/2404.03048v2 | Vibhor Agarwal, Aravindh Raman, Nishanth Sastry, Ahmed M. Abdelmoniem, Gareth Tyson, Ignacio Castro | 2024-04-16T23:59:51Z | cs.CY, cs.CL | The recent development of decentralised and interoperable social networks
(such as the "fediverse") creates new challenges for content moderators. This
is because millions of posts generated on one server can easily "spread" to
another, even if the recipient server has very different moderation policies.
An obvious solution would be to leverage moderation tools to automatically tag
(and filter) posts that contravene moderation policies, e.g. related to toxic
speech. Recent work has exploited the conversational context of a post to
improve this automatic tagging, e.g. using the replies to a post to help
classify if it contains toxic speech. This has shown particular potential in
environments with large training sets that contain complete conversations.
This, however, creates challenges in a decentralised context, as a single
conversation may be fragmented across multiple servers. Thus, each server only
has a partial view of an entire conversation because conversations are often
federated across servers in a non-synchronized fashion. To address this, we
propose a decentralised conversation-aware content moderation approach suitable
for the fediverse. Our approach employs a graph deep learning model (GraphNLI)
trained locally on each server. The model exploits local data to train a model
that combines post and conversational information captured through random walks
to detect toxicity. We evaluate our approach with data from Pleroma, a major
decentralised and interoperable micro-blogging network containing 2 million
conversations. Our model effectively detects toxicity on larger instances,
exclusively trained using their local post information (0.8837 macro-F1). Our
approach has considerable scope to improve moderation in decentralised and
interoperable social networks such as Pleroma or Mastodon. | cs.CY | Decentralised Moderation for Interoperable Social Networks: A Conversation-based Approach for Pleroma and the Fediverse |
2024-04-03T19:47:33Z | http://arxiv.org/pdf/2404.03036v1 | http://arxiv.org/abs/2404.03036v1 | Constanza Fierro, Nicolas Garneau, Emanuele Bugliarello, Yova Kementchedjhieva, Anders Søgaard | 2024-04-03T19:47:33Z | cs.CL | Facts are subject to contingencies and can be true or false in different
circumstances. One such contingency is time, wherein some facts mutate over a
given period, e.g., the president of a country or the winner of a championship.
Trustworthy language models ideally identify mutable facts as such and process
them accordingly. We create MuLan, a benchmark for evaluating the ability of
English language models to anticipate time-contingency, covering both 1:1 and
1:N relations. We hypothesize that mutable facts are encoded differently than
immutable ones, hence being easier to update. In a detailed evaluation of six
popular large language models, we consistently find differences in the LLMs'
confidence, representations, and update behavior, depending on the mutability
of a fact. Our findings should inform future work on the injection of and
induction of time-contingent knowledge to/from LLMs. | cs.CL | MuLan: A Study of Fact Mutability in Language Models |
2024-04-03T19:45:40Z | http://arxiv.org/pdf/2404.04284v1 | http://arxiv.org/abs/2404.04284v1 | Giuliano Lorenzoni, Cristina Tavares, Nathalia Nascimento, Paulo Alencar, Donald Cowan | 2024-04-03T19:45:40Z | cs.CL | Depression has affected millions of people worldwide and has become one of
the most common mental disorders. Early mental disorder detection can reduce
costs for public health agencies and prevent other major comorbidities.
Additionally, the shortage of specialized personnel is very concerning since
Depression diagnosis is highly dependent on expert professionals and is
time-consuming. Recent research has evidenced that machine learning (ML) and
Natural Language Processing (NLP) tools and techniques have significantly bene
ted the diagnosis of depression. However, there are still several challenges in
the assessment of depression detection approaches in which other conditions
such as post-traumatic stress disorder (PTSD) are present. These challenges
include assessing alternatives in terms of data cleaning and pre-processing
techniques, feature selection, and appropriate ML classification algorithms.
This paper tackels such an assessment based on a case study that compares
different ML classifiers, specifically in terms of data cleaning and
pre-processing, feature selection, parameter setting, and model choices. The
case study is based on the Distress Analysis Interview Corpus - Wizard-of-Oz
(DAIC-WOZ) dataset, which is designed to support the diagnosis of mental
disorders such as depression, anxiety, and PTSD. Besides the assessment of
alternative techniques, we were able to build models with accuracy levels
around 84% with Random Forest and XGBoost models, which is significantly higher
than the results from the comparable literature which presented the level of
accuracy of 72% from the SVM model. | cs.CL | Assessing ML Classification Algorithms and NLP Techniques for Depression Detection: An Experimental Case Study |
2024-04-03T19:31:56Z | http://arxiv.org/pdf/2404.03028v2 | http://arxiv.org/abs/2404.03028v2 | Emmy Liu, Graham Neubig, Jacob Andreas | 2024-04-10T19:03:00Z | cs.CL | Modern language models (LMs) can learn to perform new tasks in different
ways: in instruction following, the target task is described explicitly in
natural language; in few-shot prompting, the task is specified implicitly with
a small number of examples; in instruction inference, LMs are presented with
in-context examples and are then prompted to generate a natural language task
description before making predictions. Each of these procedures may be thought
of as invoking a different form of reasoning: instruction following involves
deductive reasoning, few-shot prompting involves inductive reasoning, and
instruction inference involves abductive reasoning. How do these different
capabilities relate? Across four LMs (from the gpt and llama families) and two
learning problems (involving arithmetic functions and machine translation) we
find a strong dissociation between the different types of reasoning: LMs can
sometimes learn effectively from few-shot prompts even when they are unable to
explain their own prediction rules; conversely, they sometimes infer useful
task descriptions while completely failing to learn from human-generated
descriptions of the same task. Our results highlight the non-systematic nature
of reasoning even in some of today's largest LMs, and underscore the fact that
very different learning mechanisms may be invoked by seemingly similar
prompting procedures. | cs.CL | An Incomplete Loop: Deductive, Inductive, and Abductive Learning in Large Language Models |
2024-04-03T19:23:18Z | http://arxiv.org/pdf/2404.03027v2 | http://arxiv.org/abs/2404.03027v2 | Weidi Luo, Siyuan Ma, Xiaogeng Liu, Xiaoyu Guo, Chaowei Xiao | 2024-04-18T17:11:53Z | cs.CR, cs.AI, cs.CL | With the rapid advancements in Multimodal Large Language Models (MLLMs),
securing these models against malicious inputs while aligning them with human
values has emerged as a critical challenge. In this paper, we investigate an
important and unexplored question of whether techniques that successfully
jailbreak Large Language Models (LLMs) can be equally effective in jailbreaking
MLLMs. To explore this issue, we introduce JailBreakV-28K, a pioneering
benchmark designed to assess the transferability of LLM jailbreak techniques to
MLLMs, thereby evaluating the robustness of MLLMs against diverse jailbreak
attacks. Utilizing a dataset of 2, 000 malicious queries that is also proposed
in this paper, we generate 20, 000 text-based jailbreak prompts using advanced
jailbreak attacks on LLMs, alongside 8, 000 image-based jailbreak inputs from
recent MLLMs jailbreak attacks, our comprehensive dataset includes 28, 000 test
cases across a spectrum of adversarial scenarios. Our evaluation of 10
open-source MLLMs reveals a notably high Attack Success Rate (ASR) for attacks
transferred from LLMs, highlighting a critical vulnerability in MLLMs that
stems from their text-processing capabilities. Our findings underscore the
urgent need for future research to address alignment vulnerabilities in MLLMs
from both textual and visual inputs. | cs.CR | JailBreakV-28K: A Benchmark for Assessing the Robustness of MultiModal Large Language Models against Jailbreak Attacks |
2024-04-03T19:17:43Z | http://arxiv.org/pdf/2404.03022v2 | http://arxiv.org/abs/2404.03022v2 | Amirhossein Abaskohi, Amirhossein Dabiriaghdam, Lele Wang, Giuseppe Carenini | 2024-06-11T19:34:19Z | cs.CL, cs.CV, cs.IT, cs.LG, math.IT | Memes, combining text and images, frequently use metaphors to convey
persuasive messages, shaping public opinion. Motivated by this, our team
engaged in SemEval-2024 Task 4, a hierarchical multi-label classification task
designed to identify rhetorical and psychological persuasion techniques
embedded within memes. To tackle this problem, we introduced a caption
generation step to assess the modality gap and the impact of additional
semantic information from images, which improved our result. Our best model
utilizes GPT-4 generated captions alongside meme text to fine-tune RoBERTa as
the text encoder and CLIP as the image encoder. It outperforms the baseline by
a large margin in all 12 subtasks. In particular, it ranked in top-3 across all
languages in Subtask 2a, and top-4 in Subtask 2b, demonstrating quantitatively
strong performance. The improvement achieved by the introduced intermediate
step is likely attributable to the metaphorical essence of images that
challenges visual encoders. This highlights the potential for improving
abstract visual semantics encoding. | cs.CL | BCAmirs at SemEval-2024 Task 4: Beyond Words: A Multimodal and Multilingual Exploration of Persuasion in Memes |
2024-04-03T19:14:45Z | http://arxiv.org/pdf/2404.03021v1 | http://arxiv.org/abs/2404.03021v1 | Alexander Loth, Martin Kappes, Marc-Oliver Pahl | 2024-04-03T19:14:45Z | cs.CL, cs.AI, I.2.7 | Fake news significantly influence our society. They impact consumers, voters,
and many other societal groups. While Fake News exist for a centuries,
Generative AI brings fake news on a new level. It is now possible to automate
the creation of masses of high-quality individually targeted Fake News. On the
other end, Generative AI can also help detecting Fake News. Both fields are
young but developing fast.
This survey provides a comprehensive examination of the research and
practical use of Generative AI for Fake News detection and creation in 2024.
Following the Structured Literature Survey approach, the paper synthesizes
current results in the following topic clusters 1) enabling technologies, 2)
creation of Fake News, 3) case study social media as most relevant distribution
channel, 4) detection of Fake News, and 5) deepfakes as upcoming technology.
The article also identifies current challenges and open issues. | cs.CL | Blessing or curse? A survey on the Impact of Generative AI on Fake News |
2024-04-03T18:09:33Z | http://arxiv.org/pdf/2404.02983v1 | http://arxiv.org/abs/2404.02983v1 | Gaia Carenini, Luca Bischetti, Walter Schaeken, Valentina Bambini | 2024-04-03T18:09:33Z | cs.CL | The Rational Speech Act (RSA) model provides a flexible framework to model
pragmatic reasoning in computational terms. However, state-of-the-art RSA
models are still fairly distant from modern machine learning techniques and
present a number of limitations related to their interpretability and
scalability. Here, we introduce a new RSA framework for metaphor understanding
that addresses these limitations by providing an explicit formula - based on
the mutually shared information between the speaker and the listener - for the
estimation of the communicative goal and by learning the rationality parameter
using gradient-based methods. The model was tested against 24 metaphors, not
limited to the conventional $\textit{John-is-a-shark}$ type. Results suggest an
overall strong positive correlation between the distributions generated by the
model and the interpretations obtained from the human behavioral data, which
increased when the intended meaning capitalized on properties that were
inherent to the vehicle concept. Overall, findings suggest that metaphor
processing is well captured by a typicality-based Bayesian model, even when
more scalable and interpretable, opening up possible applications to other
pragmatic phenomena and novel uses for increasing Large Language Models
interpretability. Yet, results highlight that the more creative nuances of
metaphorical meaning, not strictly encoded in the lexical concepts, are a
challenging aspect for machines. | cs.CL | Towards a Fully Interpretable and More Scalable RSA Model for Metaphor Understanding |
2024-04-03T17:59:36Z | http://arxiv.org/pdf/2404.02904v1 | http://arxiv.org/abs/2404.02904v1 | Suzanne Petryk, David M. Chan, Anish Kachinthaya, Haodi Zou, John Canny, Joseph E. Gonzalez, Trevor Darrell | 2024-04-03T17:59:36Z | cs.CV, cs.AI, cs.CL, cs.LG | Despite recent advances in multimodal pre-training for visual description,
state-of-the-art models still produce captions containing errors, such as
hallucinating objects not present in a scene. The existing prominent metric for
object hallucination, CHAIR, is limited to a fixed set of MS COCO objects and
synonyms. In this work, we propose a modernized open-vocabulary metric, ALOHa,
which leverages large language models (LLMs) to measure object hallucinations.
Specifically, we use an LLM to extract groundable objects from a candidate
caption, measure their semantic similarity to reference objects from captions
and object detections, and use Hungarian matching to produce a final
hallucination score. We show that ALOHa correctly identifies 13.6% more
hallucinated objects than CHAIR on HAT, a new gold-standard subset of MS COCO
Captions annotated for hallucinations, and 30.8% more on nocaps, where objects
extend beyond MS COCO categories. Our code is available at
https://davidmchan.github.io/aloha/. | cs.CV | ALOHa: A New Measure for Hallucination in Captioning Models |
2024-04-03T17:51:18Z | http://arxiv.org/pdf/2404.02893v1 | http://arxiv.org/abs/2404.02893v1 | Yifan Xu, Xiao Liu, Xinghan Liu, Zhenyu Hou, Yueyan Li, Xiaohan Zhang, Zihan Wang, Aohan Zeng, Zhengxiao Du, Wenyi Zhao, Jie Tang, Yuxiao Dong | 2024-04-03T17:51:18Z | cs.CL | Large language models (LLMs) have shown excellent mastering of human
language, but still struggle in real-world applications that require
mathematical problem-solving. While many strategies and datasets to enhance
LLMs' mathematics are developed, it remains a challenge to simultaneously
maintain and improve both language and mathematical capabilities in deployed
LLM systems.In this work, we tailor the Self-Critique pipeline, which addresses
the challenge in the feedback learning stage of LLM alignment. We first train a
general Math-Critique model from the LLM itself to provide feedback signals.
Then, we sequentially employ rejective fine-tuning and direct preference
optimization over the LLM's own generations for data collection. Based on
ChatGLM3-32B, we conduct a series of experiments on both academic and our newly
created challenging dataset, MathUserEval. Results show that our pipeline
significantly enhances the LLM's mathematical problem-solving while still
improving its language ability, outperforming LLMs that could be two times
larger. Related techniques have been deployed to
ChatGLM\footnote{\url{https://chatglm.cn}}, an online serving LLM. Related
evaluation dataset and scripts are released at
\url{https://github.com/THUDM/ChatGLM-Math}. | cs.CL | ChatGLM-Math: Improving Math Problem-Solving in Large Language Models with a Self-Critique Pipeline |
2024-04-03T17:33:21Z | http://arxiv.org/pdf/2404.02882v1 | http://arxiv.org/abs/2404.02882v1 | Weigao Sun, Zhen Qin, Dong Li, Xuyang Shen, Yu Qiao, Yiran Zhong | 2024-04-03T17:33:21Z | cs.LG, cs.CL | Sequence Parallel (SP) serves as a prevalent strategy to handle long
sequences that exceed the memory limit of a single GPU. However, existing SP
methods do not take advantage of linear attention features, resulting in
sub-optimal parallelism efficiency and usability for linear attention-based
language models. In this paper, we introduce Linear Attention Sequence Parallel
(LASP), an efficient SP method tailored to linear attention-based language
models. Specifically, we design an efficient point-to-point communication
mechanism to leverage the right-product kernel trick of linear attention, which
sharply decreases the communication overhead of SP. We also enhance the
practical efficiency of LASP by performing kernel fusion and intermediate state
caching, making the implementation of LASP hardware-friendly on GPU clusters.
Furthermore, we meticulously ensure the compatibility of sequence-level LASP
with all types of batch-level data parallel methods, which is vital for
distributed training on large clusters with long sequences and large batches.
We conduct extensive experiments on two linear attention-based models with
varying sequence lengths and GPU cluster sizes. LASP scales sequence length up
to 4096K using 128 A100 80G GPUs on 1B models, which is 8 times longer than
existing SP methods while being significantly faster. The code is available at
https://github.com/OpenNLPLab/LASP. | cs.LG | Linear Attention Sequence Parallelism |
2024-04-03T16:16:31Z | http://arxiv.org/pdf/2404.02837v1 | http://arxiv.org/abs/2404.02837v1 | Wanyun Cui, Qianle Wang | 2024-04-03T16:16:31Z | cs.CL | This paper reveals the phenomenon of parameter heterogeneity in large
language models (LLMs). We find that a small subset of ``cherry'' parameters
exhibit a disproportionately large influence on model performance, while the
vast majority of parameters have minimal impact. This heterogeneity is found to
be prevalent across different model families, scales, and types. Motivated by
this observation, we propose CherryQ, a novel quantization method that unifies
the optimization of mixed-precision parameters. CherryQ identifies and
preserves the critical cherry parameters in high precision while aggressively
quantizing the remaining parameters to low precision. Extensive experiments
demonstrate the effectiveness of CherryQ. CherryQ outperforms existing
quantization approaches in terms of perplexity and downstream task performance.
Notably, our 3-bit quantized Vicuna-1.5 exhibits competitive performance
compared to their 16-bit counterparts. These findings highlight the potential
of CherryQ for enabling efficient deployment of LLMs by taking advantage of
parameter heterogeneity. | cs.CL | Cherry on Top: Parameter Heterogeneity and Quantization in Large Language Models |
2024-04-03T16:13:29Z | http://arxiv.org/pdf/2404.02835v1 | http://arxiv.org/abs/2404.02835v1 | Maxime Bouthors, Josep Crego, Francois Yvon | 2024-04-03T16:13:29Z | cs.CL | Retrieval-Augmented Neural Machine Translation (RAMT) architectures retrieve
examples from memory to guide the generation process. While most works in this
trend explore new ways to exploit the retrieved examples, the upstream
retrieval step is mostly unexplored. In this paper, we study the effect of
varying retrieval methods for several translation architectures, to better
understand the interplay between these two processes. We conduct experiments in
two language pairs in a multi-domain setting and consider several downstream
architectures based on a standard autoregressive model, an edit-based model,
and a large language model with in-context learning. Our experiments show that
the choice of the retrieval technique impacts the translation scores, with
variance across architectures. We also discuss the effects of increasing the
number and diversity of examples, which are mostly positive across the board. | cs.CL | Retrieving Examples from Memory for Retrieval Augmented Neural Machine Translation: A Systematic Comparison |
2024-04-03T15:55:39Z | http://arxiv.org/pdf/2404.02823v1 | http://arxiv.org/abs/2404.02823v1 | Haoran Sun, Lixin Liu, Junjie Li, Fengyu Wang, Baohua Dong, Ran Lin, Ruohui Huang | 2024-04-03T15:55:39Z | cs.CL, cs.AI, cs.LG | The ability of large language models (LLMs) to follow instructions is crucial
to real-world applications. Despite recent advances, several studies have
highlighted that LLMs struggle when faced with challenging instructions,
especially those that include complex constraints, hindering their
effectiveness in various tasks. To address this challenge, we introduce
Conifer, a novel instruction tuning dataset, designed to enhance LLMs to follow
multi-level instructions with complex constraints. Utilizing GPT-4, we curate
the dataset by a series of LLM-driven refinement processes to ensure high
quality. We also propose a progressive learning scheme that emphasizes an
easy-to-hard progression, and learning from process feedback. Models trained
with Conifer exhibit remarkable improvements in instruction-following
abilities, especially for instructions with complex constraints. On several
instruction-following benchmarks, our 7B model outperforms the state-of-the-art
open-source 7B models, even exceeds the performance of models 10 times larger
on certain metrics. All the code and Conifer dataset are available at
https://www.github.com/ConiferLM/Conifer. | cs.CL | Conifer: Improving Complex Constrained Instruction-Following Ability of Large Language Models |
2024-04-03T15:55:27Z | http://arxiv.org/pdf/2404.02822v2 | http://arxiv.org/abs/2404.02822v2 | Matyas Juhasz, Tina Marchand, Roshan Melwani, Kalyan Dutia, Sarah Goodenough, Harrison Pim, Henry Franks | 2024-04-04T11:23:59Z | cs.CY, cs.CL, cs.LG | Quantified policy targets are a fundamental element of climate policy,
typically characterised by domain-specific and technical language. Current
methods for curating comprehensive views of global climate policy targets
entail significant manual effort. At present there are few scalable methods for
extracting climate targets from national laws or policies, which limits
policymakers' and researchers' ability to (1) assess private and public sector
alignment with global goals and (2) inform policy decisions. In this paper we
present an approach for extracting mentions of climate targets from national
laws and policies. We create an expert-annotated dataset identifying three
categories of target ('Net Zero', 'Reduction' and 'Other' (e.g. renewable
energy targets)) and train a classifier to reliably identify them in text. We
investigate bias and equity impacts related to our model and identify specific
years and country names as problematic features. Finally, we investigate the
characteristics of the dataset produced by running this classifier on the
Climate Policy Radar (CPR) dataset of global national climate laws and policies
and UNFCCC submissions, highlighting the potential of automated and scalable
data collection for existing climate policy databases and supporting further
research. Our work represents a significant upgrade in the accessibility of
these key climate policy elements for policymakers and researchers. We publish
our model at https://huggingface.co/ClimatePolicyRadar/national-climate-targets
and related dataset at
https://huggingface.co/datasets/ClimatePolicyRadar/national-climate-targets. | cs.CY | Identifying Climate Targets in National Laws and Policies using Machine Learning |
2024-04-03T15:17:21Z | http://arxiv.org/pdf/2404.02800v1 | http://arxiv.org/abs/2404.02800v1 | Bernardo Leite, Henrique Lopes Cardoso | 2024-04-03T15:17:21Z | cs.CL, cs.AI | Question Generation aims to automatically generate questions based on a given
input provided as context. A controllable question generation scheme focuses on
generating questions with specific attributes, allowing better control. In this
study, we propose a few-shot prompting strategy for controlling the generation
of question-answer pairs from children's narrative texts. We aim to control two
attributes: the question's explicitness and underlying narrative elements. With
empirical evaluation, we show the effectiveness of controlling the generation
process by employing few-shot prompting side by side with a reference model.
Our experiments highlight instances where the few-shot strategy surpasses the
reference model, particularly in scenarios such as semantic closeness
evaluation and the diversity and coherency of question-answer pairs. However,
these improvements are not always statistically significant. The code is
publicly available at github.com/bernardoleite/few-shot-prompting-qg-control. | cs.CL | On Few-Shot Prompting for Controllable Question-Answer Generation in Narrative Comprehension |
2024-04-03T14:39:47Z | http://arxiv.org/pdf/2404.02772v2 | http://arxiv.org/abs/2404.02772v2 | Ziyang Wang, Sanwoo Lee, Hsiu-Yuan Huang, Yunfang Wu | 2024-04-10T04:25:09Z | cs.CL | Prompt-based methods have achieved promising results in most few-shot text
classification tasks. However, for readability assessment tasks, traditional
prompt methods lackcrucial linguistic knowledge, which has already been proven
to be essential. Moreover, previous studies on utilizing linguistic features
have shown non-robust performance in few-shot settings and may even impair
model performance.To address these issues, we propose a novel prompt-based
tuning framework that incorporates rich linguistic knowledge, called Feature
Prompt Tuning (FPT). Specifically, we extract linguistic features from the text
and embed them into trainable soft prompts. Further, we devise a new loss
function to calibrate the similarity ranking order between categories.
Experimental results demonstrate that our proposed method FTP not only exhibits
a significant performance improvement over the prior best prompt-based tuning
approaches, but also surpasses the previous leading methods that incorporate
linguistic features. Also, our proposed model significantly outperforms the
large language model gpt-3.5-turbo-16k in most cases. Our proposed method
establishes a new architecture for prompt tuning that sheds light on how
linguistic features can be easily adapted to linguistic-related tasks. | cs.CL | FPT: Feature Prompt Tuning for Few-shot Readability Assessment |
2024-04-03T14:07:02Z | http://arxiv.org/pdf/2404.02761v3 | http://arxiv.org/abs/2404.02761v3 | Maike Behrendt, Stefan Sylvius Wagner, Marc Ziegele, Lena Wilms, Anke Stoll, Dominique Heinbach, Stefan Harmeling | 2024-04-17T10:56:48Z | cs.CL, cs.AI, cs.LG | Measuring the quality of contributions in political online discussions is
crucial in deliberation research and computer science. Research has identified
various indicators to assess online discussion quality, and with deep learning
advancements, automating these measures has become feasible. While some studies
focus on analyzing specific quality indicators, a comprehensive quality score
incorporating various deliberative aspects is often preferred. In this work, we
introduce AQuA, an additive score that calculates a unified deliberative
quality score from multiple indices for each discussion post. Unlike other
singular scores, AQuA preserves information on the deliberative aspects present
in comments, enhancing model transparency. We develop adapter models for 20
deliberative indices, and calculate correlation coefficients between experts'
annotations and the perceived deliberativeness by non-experts to weigh the
individual indices into a single deliberative score. We demonstrate that the
AQuA score can be computed easily from pre-trained adapters and aligns well
with annotations on other datasets that have not be seen during training. The
analysis of experts' vs. non-experts' annotations confirms theoretical findings
in the social science literature. | cs.CL | AQuA -- Combining Experts' and Non-Experts' Views To Assess Deliberation Quality in Online Discussions Using LLMs |
2024-04-03T13:20:24Z | http://arxiv.org/pdf/2404.02717v1 | http://arxiv.org/abs/2404.02717v1 | Viet-Tung Do, Van-Khanh Hoang, Duy-Hung Nguyen, Shahab Sabahi, Jeff Yang, Hajime Hotta, Minh-Tien Nguyen, Hung Le | 2024-04-03T13:20:24Z | cs.CL, cs.LG | Large Language Models (LLMs) can perform various natural language processing
tasks with suitable instruction prompts. However, designing effective prompts
manually is challenging and time-consuming. Existing methods for automatic
prompt optimization either lack flexibility or efficiency. In this paper, we
propose an effective approach to automatically select the optimal prompt for a
given input from a finite set of synthetic candidate prompts. Our approach
consists of three steps: (1) clustering the training data and generating
candidate prompts for each cluster using an LLM-based prompt generator; (2)
synthesizing a dataset of input-prompt-output tuples for training a prompt
evaluator to rank the prompts based on their relevance to the input; (3) using
the prompt evaluator to select the best prompt for a new input at test time.
Our approach balances prompt generality-specificity and eliminates the need for
resource-intensive training and inference. It demonstrates competitive
performance on zero-shot question-answering datasets: GSM8K, MultiArith, and
AQuA. | cs.CL | Automatic Prompt Selection for Large Language Models |
2024-04-03T13:08:26Z | http://arxiv.org/pdf/2404.02710v1 | http://arxiv.org/abs/2404.02710v1 | Zheng Yuan, Dorina de Jong, Štefan Beňuš, Noël Nguyen, Ruitao Feng, Róbert Sabo, Luciano Fadiga, Alessandro D`Ausilio | 2024-04-03T13:08:26Z | cs.CL, eess.AS | We introduce the Alternating Reading Task (ART) Corpus, a collection of
dyadic sentence reading for studying the entrainment and imitation behaviour in
speech communication. The ART corpus features three experimental conditions -
solo reading, alternating reading, and deliberate imitation - as well as three
sub-corpora encompassing French-, Italian-, and Slovak-accented English. This
design allows systematic investigation of speech entrainment in a controlled
and less-spontaneous setting. Alongside detailed transcriptions, it includes
English proficiency scores, demographics, and in-experiment questionnaires for
probing linguistic, personal and interpersonal influences on entrainment. Our
presentation covers its design, collection, annotation processes, initial
analysis, and future research prospects. | cs.CL | ART: The Alternating Reading Task Corpus for Speech Entrainment and Imitation |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.