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
|
|---|---|---|---|---|---|---|---|---|
2024-06-18T12:54:47Z | http://arxiv.org/pdf/2406.12950v1 | http://arxiv.org/abs/2406.12950v1 | Yuyan Liu, Sirui Ding, Sheng Zhou, Wenqi Fan, Qiaoyu Tan | 2024-06-18T12:54:47Z | q-bio.QM, cs.AI, cs.CE, cs.CL, cs.LG | Molecular property prediction (MPP) is a fundamental and crucial task in drug
discovery. However, prior methods are limited by the requirement for a large
number of labeled molecules and their restricted ability to generalize for
unseen and new tasks, both of which are essential for real-world applications.
To address these challenges, we present MolecularGPT for few-shot MPP. From a
perspective on instruction tuning, we fine-tune large language models (LLMs)
based on curated molecular instructions spanning over 1000 property prediction
tasks. This enables building a versatile and specialized LLM that can be
adapted to novel MPP tasks without any fine-tuning through zero- and few-shot
in-context learning (ICL). MolecularGPT exhibits competitive in-context
reasoning capabilities across 10 downstream evaluation datasets, setting new
benchmarks for few-shot molecular prediction tasks. More importantly, with just
two-shot examples, MolecularGPT can outperform standard supervised graph neural
network methods on 4 out of 7 datasets. It also excels state-of-the-art LLM
baselines by up to 16.6% increase on classification accuracy and decrease of
199.17 on regression metrics (e.g., RMSE) under zero-shot. This study
demonstrates the potential of LLMs as effective few-shot molecular property
predictors. The code is available at https://github.com/NYUSHCS/MolecularGPT. | q-bio.QM | MolecularGPT: Open Large Language Model (LLM) for Few-Shot Molecular Property Prediction |
2024-06-18T12:52:51Z | http://arxiv.org/pdf/2406.12566v2 | http://arxiv.org/abs/2406.12566v2 | Shuting Wang, Xin Yu, Mang Wang, Weipeng Chen, Yutao Zhu, Zhicheng Dou | 2024-06-21T18:12:15Z | cs.CL | Retrieval-augmented generation (RAG) effectively addresses issues of static
knowledge and hallucination in large language models. Existing studies mostly
focus on question scenarios with clear user intents and concise answers.
However, it is prevalent that users issue broad, open-ended queries with
diverse sub-intents, for which they desire rich and long-form answers covering
multiple relevant aspects. To tackle this important yet underexplored problem,
we propose a novel RAG framework, namely RichRAG. It includes a sub-aspect
explorer to identify potential sub-aspects of input questions, a multi-faceted
retriever to build a candidate pool of diverse external documents related to
these sub-aspects, and a generative list-wise ranker, which is a key module to
provide the top-k most valuable documents for the final generator. These ranked
documents sufficiently cover various query aspects and are aware of the
generator's preferences, hence incentivizing it to produce rich and
comprehensive responses for users. The training of our ranker involves a
supervised fine-tuning stage to ensure the basic coverage of documents, and a
reinforcement learning stage to align downstream LLM's preferences to the
ranking of documents. Experimental results on two publicly available datasets
prove that our framework effectively and efficiently provides comprehensive and
satisfying responses to users. | cs.CL | RichRAG: Crafting Rich Responses for Multi-faceted Queries in Retrieval-Augmented Generation |
2024-06-18T12:50:00Z | http://arxiv.org/pdf/2406.12564v1 | http://arxiv.org/abs/2406.12564v1 | Viktor Moskvoretskii, Nazarii Tupitsa, Chris Biemann, Samuel Horváth, Eduard Gorbunov, Irina Nikishina | 2024-06-18T12:50:00Z | cs.CL, cs.LG | We present a new approach based on the Personalized Federated Learning
algorithm MeritFed that can be applied to Natural Language Tasks with
heterogeneous data. We evaluate it on the Low-Resource Machine Translation
task, using the dataset from the Large-Scale Multilingual Machine Translation
Shared Task (Small Track #2) and the subset of Sami languages from the
multilingual benchmark for Finno-Ugric languages. In addition to its
effectiveness, MeritFed is also highly interpretable, as it can be applied to
track the impact of each language used for training. Our analysis reveals that
target dataset size affects weight distribution across auxiliary languages,
that unrelated languages do not interfere with the training, and auxiliary
optimizer parameters have minimal impact. Our approach is easy to apply with a
few lines of code, and we provide scripts for reproducing the experiments at
https://github.com/VityaVitalich/MeritFed | cs.CL | Low-Resource Machine Translation through the Lens of Personalized Federated Learning |
2024-06-18T12:26:09Z | http://arxiv.org/pdf/2406.12549v1 | http://arxiv.org/abs/2406.12549v1 | Dominik Macko, Jakub Kopal, Robert Moro, Ivan Srba | 2024-06-18T12:26:09Z | cs.CL, cs.AI | Recent LLMs are able to generate high-quality multilingual texts,
indistinguishable for humans from authentic human-written ones. Research in
machine-generated text detection is however mostly focused on the English
language and longer texts, such as news articles, scientific papers or student
essays. Social-media texts are usually much shorter and often feature informal
language, grammatical errors, or distinct linguistic items (e.g., emoticons,
hashtags). There is a gap in studying the ability of existing methods in
detection of such texts, reflected also in the lack of existing multilingual
benchmark datasets. To fill this gap we propose the first multilingual (22
languages) and multi-platform (5 social media platforms) dataset for
benchmarking machine-generated text detection in the social-media domain,
called MultiSocial. It contains 472,097 texts, of which about 58k are
human-written and approximately the same amount is generated by each of 7
multilingual LLMs. We use this benchmark to compare existing detection methods
in zero-shot as well as fine-tuned form. Our results indicate that the
fine-tuned detectors have no problem to be trained on social-media texts and
that the platform selection for training matters. | cs.CL | MultiSocial: Multilingual Benchmark of Machine-Generated Text Detection of Social-Media Texts |
2024-06-18T12:25:13Z | http://arxiv.org/pdf/2406.12548v1 | http://arxiv.org/abs/2406.12548v1 | Yuhao Dan, Jie Zhou, Qin Chen, Junfeng Tian, Liang He | 2024-06-18T12:25:13Z | cs.CL | Personalized large language models (LLMs) have attracted great attention in
many applications, such as intelligent education and emotional support. Most
work focuses on controlling the character settings based on the profile (e.g.,
age, skill, experience, and so on). Conversely, the psychological theory-based
personality traits with implicit expression and behavior are not well modeled,
limiting their potential application in more specialized fields such as the
psychological counseling agents. In this paper, we propose a mixture of experts
(MoE)-based personalized LLMs, named P-tailor, to model the Big Five
Personality Traits. Particularly, we learn specialized LoRA experts to
represent various traits, such as openness, conscientiousness, extraversion,
agreeableness and neuroticism. Then, we integrate P-Tailor with a personality
specialization loss, promoting experts to specialize in distinct personality
traits, thereby enhancing the efficiency of model parameter utilization. Due to
the lack of datasets, we also curate a high-quality personality crafting
dataset (PCD) to learn and develop the ability to exhibit different personality
traits across various topics. We conduct extensive experiments to verify the
great performance and effectiveness of P-Tailor in manipulation of the
fine-grained personality traits of LLMs. | cs.CL | P-Tailor: Customizing Personality Traits for Language Models via Mixture of Specialized LoRA Experts |
2024-06-18T12:24:22Z | http://arxiv.org/pdf/2406.12546v1 | http://arxiv.org/abs/2406.12546v1 | Philipp Mondorf, Barbara Plank | 2024-06-18T12:24:22Z | cs.CL | Knights and knaves problems represent a classic genre of logical puzzles
where characters either tell the truth or lie. The objective is to logically
deduce each character's identity based on their statements. The challenge
arises from the truth-telling or lying behavior, which influences the logical
implications of each statement. Solving these puzzles requires not only direct
deductions from individual statements, but the ability to assess the
truthfulness of statements by reasoning through various hypothetical scenarios.
As such, knights and knaves puzzles serve as compelling examples of
suppositional reasoning. In this paper, we introduce $\textit{TruthQuest}$, a
benchmark for suppositional reasoning based on the principles of knights and
knaves puzzles. Our benchmark presents problems of varying complexity,
considering both the number of characters and the types of logical statements
involved. Evaluations on $\textit{TruthQuest}$ show that large language models
like Llama 3 and Mixtral-8x7B exhibit significant difficulties solving these
tasks. A detailed error analysis of the models' output reveals that
lower-performing models exhibit a diverse range of reasoning errors, frequently
failing to grasp the concept of truth and lies. In comparison, more proficient
models primarily struggle with accurately inferring the logical implications of
potentially false statements. | cs.CL | Liar, Liar, Logical Mire: A Benchmark for Suppositional Reasoning in Large Language Models |
2024-06-18T12:09:02Z | http://arxiv.org/pdf/2406.12534v3 | http://arxiv.org/abs/2406.12534v3 | Qinyuan Cheng, Xiaonan Li, Shimin Li, Qin Zhu, Zhangyue Yin, Yunfan Shao, Linyang Li, Tianxiang Sun, Hang Yan, Xipeng Qiu | 2024-06-27T15:37:15Z | cs.CL | In Retrieval-Augmented Generation (RAG), retrieval is not always helpful and
applying it to every instruction is sub-optimal. Therefore, determining whether
to retrieve is crucial for RAG, which is usually referred to as Active
Retrieval. However, existing active retrieval methods face two challenges: 1.
They usually rely on a single criterion, which struggles with handling various
types of instructions. 2. They depend on specialized and highly differentiated
procedures, and thus combining them makes the RAG system more complicated and
leads to higher response latency. To address these challenges, we propose
Unified Active Retrieval (UAR). UAR contains four orthogonal criteria and casts
them into plug-and-play classification tasks, which achieves multifaceted
retrieval timing judgements with negligible extra inference cost. We further
introduce the Unified Active Retrieval Criteria (UAR-Criteria), designed to
process diverse active retrieval scenarios through a standardized procedure.
Experiments on four representative types of user instructions show that UAR
significantly outperforms existing work on the retrieval timing judgement and
the performance of downstream tasks, which shows the effectiveness of UAR and
its helpfulness to downstream tasks. | cs.CL | Unified Active Retrieval for Retrieval Augmented Generation |
2024-06-18T11:55:05Z | http://arxiv.org/pdf/2406.12527v1 | http://arxiv.org/abs/2406.12527v1 | Tianyuan Zou, Yang Liu, Peng Li, Jianqing Zhang, Jingjing Liu, Ya-Qin Zhang | 2024-06-18T11:55:05Z | cs.CL | Data generation-based zero-shot learning, although effective in training
Small Task-specific Models (STMs) via synthetic datasets generated by
Pre-trained Language Models (PLMs), is often limited by the low quality of such
synthetic datasets. Previous solutions have primarily focused on single PLM
settings, where synthetic datasets are typically restricted to specific
sub-spaces and often deviate from real-world distributions, leading to severe
distribution bias. To mitigate such bias, we propose FuseGen, a novel data
generation-based zero-shot learning framework that introduces a new criteria
for subset selection from synthetic datasets via utilizing multiple PLMs and
trained STMs. The chosen subset provides in-context feedback to each PLM,
enhancing dataset quality through iterative data generation. Trained STMs are
then used for sample re-weighting as well, further improving data quality.
Extensive experiments across diverse tasks demonstrate that FuseGen
substantially outperforms existing methods, highly effective in boosting STM
performance in a PLM-agnostic way. Code is provided in
https://github.com/LindaLydia/FuseGen. | cs.CL | FuseGen: PLM Fusion for Data-generation based Zero-shot Learning |
2024-06-18T11:05:37Z | http://arxiv.org/pdf/2406.12502v1 | http://arxiv.org/abs/2406.12502v1 | Leonidas Gee, Milan Gritta, Gerasimos Lampouras, Ignacio Iacobacci | 2024-06-18T11:05:37Z | cs.CL | Code Language Models have been trained to generate accurate solutions,
typically with no regard for runtime. On the other hand, previous works that
explored execution optimisation have observed corresponding drops in functional
correctness. To that end, we introduce Code-Optimise, a framework that
incorporates both correctness (passed, failed) and runtime (quick, slow) as
learning signals via self-generated preference data. Our framework is both
lightweight and robust as it dynamically selects solutions to reduce
overfitting while avoiding a reliance on larger models for learning signals.
Code-Optimise achieves significant improvements in pass@k while decreasing the
competitive baseline runtimes by an additional 6% for in-domain data and up to
3% for out-of-domain data. As a byproduct, the average length of the generated
solutions is reduced by up to 48% on MBPP and 23% on HumanEval, resulting in
faster and cheaper inference. The generated data and codebase will be
open-sourced at www.open-source.link. | cs.CL | Code-Optimise: Self-Generated Preference Data for Correctness and Efficiency |
2024-06-18T10:57:27Z | http://arxiv.org/pdf/2406.12494v1 | http://arxiv.org/abs/2406.12494v1 | Masafumi Enomoto, Kunihiro Takeoka, Kosuke Akimoto, Kiril Gashteovski, Masafumi Oyamada | 2024-06-18T10:57:27Z | cs.CL | Open-Domain Multi-Document Summarization (ODMDS) is crucial for addressing
diverse information needs, which aims to generate a summary as answer to user's
query, synthesizing relevant content from multiple documents in a large
collection. Existing approaches that first find relevant passages and then
generate a summary using a language model are inadequate for ODMDS. This is
because open-ended queries often require additional context for the retrieved
passages to cover the topic comprehensively, making it challenging to retrieve
all relevant passages initially. While iterative retrieval methods have been
explored for multi-hop question answering (MQA), they are impractical for ODMDS
due to high latency from repeated large language model (LLM) inference for
reasoning. To address this issue, we propose LightPAL, a lightweight passage
retrieval method for ODMDS that constructs a graph representing passage
relationships using an LLM during indexing and employs random walk instead of
iterative reasoning and retrieval at inference time. Experiments on ODMDS
benchmarks show that LightPAL outperforms baseline retrievers in summary
quality while being significantly more efficient than an iterative MQA
approach. | cs.CL | LightPAL: Lightweight Passage Retrieval for Open Domain Multi-Document Summarization |
2024-06-18T10:36:21Z | http://arxiv.org/pdf/2406.12480v1 | http://arxiv.org/abs/2406.12480v1 | Stefan Sylvius Wagner, Maike Behrendt, Marc Ziegele, Stefan Harmeling | 2024-06-18T10:36:21Z | cs.CL, cs.LG | Stance detection holds great potential for enhancing the quality of online
political discussions, as it has shown to be useful for summarizing
discussions, detecting misinformation, and evaluating opinion distributions.
Usually, transformer-based models are used directly for stance detection, which
require large amounts of data. However, the broad range of debate questions in
online political discussion creates a variety of possible scenarios that the
model is faced with and thus makes data acquisition for model training
difficult. In this work, we show how to leverage LLM-generated synthetic data
to train and improve stance detection agents for online political
discussions:(i) We generate synthetic data for specific debate questions by
prompting a Mistral-7B model and show that fine-tuning with the generated
synthetic data can substantially improve the performance of stance detection.
(ii) We examine the impact of combining synthetic data with the most
informative samples from an unlabelled dataset. First, we use the synthetic
data to select the most informative samples, second, we combine both these
samples and the synthetic data for fine-tuning. This approach reduces labelling
effort and consistently surpasses the performance of the baseline model that is
trained with fully labeled data. Overall, we show in comprehensive experiments
that LLM-generated data greatly improves stance detection performance for
online political discussions. | cs.CL | The Power of LLM-Generated Synthetic Data for Stance Detection in Online Political Discussions |
2024-06-18T10:24:50Z | http://arxiv.org/pdf/2406.12474v1 | http://arxiv.org/abs/2406.12474v1 | Rongzhi Li, Takeru Matsuda, Hitomi Yanaka | 2024-06-18T10:24:50Z | cs.CL, stat.ME | Word embeddings represent words as multidimensional real vectors,
facilitating data analysis and processing, but are often challenging to
interpret. Independent Component Analysis (ICA) creates clearer semantic axes
by identifying independent key features. Previous research has shown ICA's
potential to reveal universal semantic axes across languages. However, it
lacked verification of the consistency of independent components within and
across languages. We investigated the consistency of semantic axes in two ways:
both within a single language and across multiple languages. We first probed
into intra-language consistency, focusing on the reproducibility of axes by
performing ICA multiple times and clustering the outcomes. Then, we
statistically examined inter-language consistency by verifying those axes'
correspondences using statistical tests. We newly applied statistical methods
to establish a robust framework that ensures the reliability and universality
of semantic axes. | cs.CL | Exploring Intra and Inter-language Consistency in Embeddings with ICA |
2024-06-18T10:20:36Z | http://arxiv.org/pdf/2406.12471v1 | http://arxiv.org/abs/2406.12471v1 | Branislav Pecher, Jan Cegin, Robert Belanec, Jakub Simko, Ivan Srba, Maria Bielikova | 2024-06-18T10:20:36Z | cs.CL | While fine-tuning of pre-trained language models generally helps to overcome
the lack of labelled training samples, it also displays model performance
instability. This instability mainly originates from randomness in
initialisation or data shuffling. To address this, researchers either modify
the training process or augment the available samples, which typically results
in increased computational costs. We propose a new mitigation strategy, called
Delayed Ensemble with Noisy Interpolation (DENI), that leverages the strengths
of ensembling, noise regularisation and model interpolation, while retaining
computational efficiency. We compare DENI with 9 representative mitigation
strategies across 3 models, 4 tuning strategies and 7 text classification
datasets. We show that: 1) DENI outperforms the best performing mitigation
strategy (Ensemble), while using only a fraction of its cost; 2) the mitigation
strategies are beneficial for parameter-efficient fine-tuning (PEFT) methods,
outperforming full fine-tuning in specific cases; and 3) combining DENI with
data augmentation often leads to even more effective instability mitigation. | cs.CL | Fighting Randomness with Randomness: Mitigating Optimisation Instability of Fine-Tuning using Delayed Ensemble and Noisy Interpolation |
2024-06-18T10:18:06Z | http://arxiv.org/pdf/2406.12468v1 | http://arxiv.org/abs/2406.12468v1 | Baolong Bi, Shenghua Liu, Yiwei Wang, Lingrui Mei, Hongcheng Gao, Yilong Xu, Xueqi Cheng | 2024-06-18T10:18:06Z | cs.CL, cs.AI | The parametric knowledge memorized by large language models (LLMs) becomes
outdated quickly. In-context editing (ICE) is currently the most effective
method for updating the knowledge of LLMs. Recent advancements involve
enhancing ICE by modifying the decoding strategy, obviating the need for
altering internal model structures or adjusting external prompts. However, this
enhancement operates across the entire sequence generation, encompassing a
plethora of non-critical tokens. In this work, we introduce $\textbf{A}$daptive
$\textbf{T}$oken $\textbf{Bias}$er ($\textbf{ATBias}$), a new decoding
technique designed to enhance ICE. It focuses on the tokens that are mostly
related to knowledge during decoding, biasing their logits by matching key
entities related to new and parametric knowledge. Experimental results show
that ATBias significantly enhances ICE performance, achieving up to a 32.3%
improvement over state-of-the-art ICE methods while incurring only half the
latency. ATBias not only improves the knowledge editing capabilities of ICE but
can also be widely applied to LLMs with negligible cost. | cs.CL | Adaptive Token Biaser: Knowledge Editing via Biasing Key Entities |
2024-06-18T09:46:44Z | http://arxiv.org/pdf/2406.12442v1 | http://arxiv.org/abs/2406.12442v1 | Ruixin Hong, Hongming Zhang, Xiaoman Pan, Dong Yu, Changshui Zhang | 2024-06-18T09:46:44Z | cs.CL, cs.AI | Abstract reasoning, the ability to reason from the abstract essence of a
problem, serves as a key to generalization in human reasoning. However,
eliciting language models to perform reasoning with abstraction remains
unexplored. This paper seeks to bridge this gap by introducing a novel
structured reasoning format called Abstraction-of-Thought (AoT). The uniqueness
of AoT lies in its explicit requirement for varying levels of abstraction
within the reasoning process. This approach could elicit language models to
first contemplate on the abstract level before incorporating concrete details,
which is overlooked by the prevailing step-by-step Chain-of-Thought (CoT)
method. To align models with the AoT format, we present AoT Collection, a
generic finetuning dataset consisting of 348k high-quality samples with AoT
reasoning processes, collected via an automated and scalable pipeline. We
finetune a wide range of language models with AoT Collection and conduct
extensive evaluations on 23 unseen tasks from the challenging benchmark
Big-Bench Hard. Experimental results indicate that models aligned to AoT
reasoning format substantially outperform those aligned to CoT in many
reasoning tasks. | cs.CL | Abstraction-of-Thought Makes Language Models Better Reasoners |
2024-06-18T09:25:35Z | http://arxiv.org/pdf/2406.12430v1 | http://arxiv.org/abs/2406.12430v1 | Myeonghwa Lee, Seonho An, Min-Soo Kim | 2024-06-18T09:25:35Z | cs.CL, cs.AI, cs.LG, I.2.7 | In this paper, we conduct a study to utilize LLMs as a solution for decision
making that requires complex data analysis. We define Decision QA as the task
of answering the best decision, $d_{best}$, for a decision-making question $Q$,
business rules $R$ and a database $D$. Since there is no benchmark that can
examine Decision QA, we propose Decision QA benchmark, DQA. It has two
scenarios, Locating and Building, constructed from two video games (Europa
Universalis IV and Victoria 3) that have almost the same goal as Decision QA.
To address Decision QA effectively, we also propose a new RAG technique called
the iterative plan-then-retrieval augmented generation (PlanRAG). Our
PlanRAG-based LM generates the plan for decision making as the first step, and
the retriever generates the queries for data analysis as the second step. The
proposed method outperforms the state-of-the-art iterative RAG method by 15.8%
in the Locating scenario and by 7.4% in the Building scenario, respectively. We
release our code and benchmark at https://github.com/myeon9h/PlanRAG. | cs.CL | PlanRAG: A Plan-then-Retrieval Augmented Generation for Generative Large Language Models as Decision Makers |
2024-06-18T09:23:54Z | http://arxiv.org/pdf/2406.12428v1 | http://arxiv.org/abs/2406.12428v1 | Kentaro Mitsui, Koh Mitsuda, Toshiaki Wakatsuki, Yukiya Hono, Kei Sawada | 2024-06-18T09:23:54Z | cs.CL, cs.AI, cs.LG, cs.SD, eess.AS | Multimodal language models that process both text and speech have a potential
for applications in spoken dialogue systems. However, current models face two
major challenges in response generation latency: (1) generating a spoken
response requires the prior generation of a written response, and (2) speech
sequences are significantly longer than text sequences. This study addresses
these issues by extending the input and output sequences of the language model
to support the parallel generation of text and speech. Our experiments on
spoken question answering tasks demonstrate that our approach improves latency
while maintaining the quality of response content. Additionally, we show that
latency can be further reduced by generating speech in multiple sequences. Demo
samples are available at https://rinnakk.github.io/research/publications/PSLM. | cs.CL | PSLM: Parallel Generation of Text and Speech with LLMs for Low-Latency Spoken Dialogue Systems |
2024-06-18T09:14:58Z | http://arxiv.org/pdf/2406.12422v1 | http://arxiv.org/abs/2406.12422v1 | Milan Straka, Jana Straková | 2024-06-18T09:14:58Z | cs.CL | We present an open-source web service for Czech morphosyntactic analysis. The
system combines a deep learning model with rescoring by a high-precision
morphological dictionary at inference time. We show that our hybrid method
surpasses two competitive baselines: While the deep learning model ensures
generalization for out-of-vocabulary words and better disambiguation, an
improvement over an existing morphological analyser MorphoDiTa, at the same
time, the deep learning model benefits from inference-time guidance of a
manually curated morphological dictionary. We achieve 50% error reduction in
lemmatization and 58% error reduction in POS tagging over MorphoDiTa, while
also offering dependency parsing. The model is trained on one of the currently
largest Czech morphosyntactic corpora, the PDT-C 1.0, with the trained models
available at https://hdl.handle.net/11234/1-5293. We provide the tool as a web
service deployed at https://lindat.mff.cuni.cz/services/udpipe/. The source
code is available at GitHub (https://github.com/ufal/udpipe/tree/udpipe-2),
along with a Python client for a simple use. The documentation for the models
can be found at https://ufal.mff.cuni.cz/udpipe/2/models#czech_pdtc1.0_model. | cs.CL | Open-Source Web Service with Morphological Dictionary-Supplemented Deep Learning for Morphosyntactic Analysis of Czech |
2024-06-18T09:14:17Z | http://arxiv.org/pdf/2406.12420v1 | http://arxiv.org/abs/2406.12420v1 | Philipp Seeberger, Dominik Wagner, Korbinian Riedhammer | 2024-06-18T09:14:17Z | cs.CL, cs.LG | With the advancement of multimedia technologies, news documents and
user-generated content are often represented as multiple modalities, making
Multimedia Event Extraction (MEE) an increasingly important challenge. However,
recent MEE methods employ weak alignment strategies and data augmentation with
simple classification models, which ignore the capabilities of natural
language-formulated event templates for the challenging Event Argument
Extraction (EAE) task. In this work, we focus on EAE and address this issue by
introducing a unified template filling model that connects the textual and
visual modalities via textual prompts. This approach enables the exploitation
of cross-ontology transfer and the incorporation of event-specific semantics.
Experiments on the M2E2 benchmark demonstrate the effectiveness of our
approach. Our system surpasses the current SOTA on textual EAE by +7% F1, and
performs generally better than the second-best systems for multimedia EAE. | cs.CL | MMUTF: Multimodal Multimedia Event Argument Extraction with Unified Template Filling |
2024-06-18T09:12:11Z | http://arxiv.org/pdf/2406.12419v1 | http://arxiv.org/abs/2406.12419v1 | Vilém Zouhar, Tom Kocmi, Mrinmaya Sachan | 2024-06-18T09:12:11Z | cs.CL | Annually, research teams spend large amounts of money to evaluate the quality
of machine translation systems (WMT, inter alia). This is expensive because it
requires detailed human labor. The recently proposed annotation protocol, Error
Span Annotation (ESA), has annotators marking erroneous parts of the
translation. In our work, we help the annotators by pre-filling the span
annotations with automatic quality estimation. With AI assistance, we obtain
more detailed annotations while cutting down the time per span annotation by
half (71s/error span $\rightarrow$ 31s/error span). The biggest advantage of
ESA$^\mathrm{AI}$ protocol is an accurate priming of annotators (pre-filled
error spans) before they assign the final score as opposed to starting from
scratch. In addition, the annotation budget can be reduced by up to 24% with
filtering of examples that the AI deems to be very likely to be correct. | cs.CL | AI-Assisted Human Evaluation of Machine Translation |
2024-06-18T09:07:30Z | http://arxiv.org/pdf/2406.12416v2 | http://arxiv.org/abs/2406.12416v2 | Hongbang Yuan, Yubo Chen, Pengfei Cao, Zhuoran Jin, Kang Liu, Jun Zhao | 2024-06-27T12:07:55Z | cs.CL, cs.AI | Large language models (LLMs) have achieved remarkable success but still tend
to generate factually erroneous responses, a phenomenon known as hallucination.
A recent trend is to use preference learning to fine-tune models to align with
factuality. However, existing work primarily evaluates fine-tuned models on
in-domain (ID) datasets and the factuality on out-of-domain (OOD) datasets
remains underexplored. In this paper, we conduct a comprehensive evaluation of
the factuality of different models tuned by various preference learning
algorithms and demonstrate that their performance on OOD datasets either
increases minimally or decreases. Subsequently, we reveal that the main cause
of model's failure to uphold factuality under a distribution shift is
\textbf{under-alignment}, rather than \textbf{over-alignment}, by analyzing the
token distribution shift of the models before and after tuning. Finally, we
propose \textbf{APEFT} (\textbf{A}tomic \textbf{P}reference \textbf{E}nhanced
\textbf{F}actuality \textbf{T}uning), a framework that enhances model's
awareness of factuality at the granularity of individual facts. Extensive
experiments demonstrate that APEFT improves model performance by an average of
$\boldsymbol{3.45\%}$ on both ID and OOD datasets, which is highly effective. | cs.CL | Beyond Under-Alignment: Atomic Preference Enhanced Factuality Tuning for Large Language Models |
2024-06-18T08:48:14Z | http://arxiv.org/pdf/2406.12403v1 | http://arxiv.org/abs/2406.12403v1 | Tao Fan, Yan Kang, Weijing Chen, Hanlin Gu, Yuanfeng Song, Lixin Fan, Kai Chen, Qiang Yang | 2024-06-18T08:48:14Z | cs.CL, cs.AI | In the context of real-world applications, leveraging large language models
(LLMs) for domain-specific tasks often faces two major challenges:
domain-specific knowledge privacy and constrained resources. To address these
issues, we propose PDSS, a privacy-preserving framework for step-by-step
distillation of LLMs. PDSS works on a server-client architecture, wherein
client transmits perturbed prompts to the server's LLM for rationale
generation. The generated rationales are then decoded by the client and used to
enrich the training of task-specific small language model(SLM) within a
multi-task learning paradigm. PDSS introduces two privacy protection
strategies: the Exponential Mechanism Strategy and the Encoder-Decoder
Strategy, balancing prompt privacy and rationale usability. Experiments
demonstrate the effectiveness of PDSS in various text generation tasks,
enabling the training of task-specific SLM with enhanced performance while
prioritizing data privacy protection. | cs.CL | PDSS: A Privacy-Preserving Framework for Step-by-Step Distillation of Large Language Models |
2024-06-18T08:44:45Z | http://arxiv.org/pdf/2406.12402v1 | http://arxiv.org/abs/2406.12402v1 | Irfan Robbani, Paul Reisert, Naoya Inoue, Surawat Pothong, Camélia Guerraoui, Wenzhi Wang, Shoichi Naito, Jungmin Choi, Kentaro Inui | 2024-06-18T08:44:45Z | cs.CL | Prior research in computational argumentation has mainly focused on scoring
the quality of arguments, with less attention on explicating logical errors. In
this work, we introduce four sets of explainable templates for common informal
logical fallacies designed to explicate a fallacy's implicit logic. Using our
templates, we conduct an annotation study on top of 400 fallacious arguments
taken from LOGIC dataset and achieve a high agreement score (Krippendorf's
alpha of 0.54) and reasonable coverage (0.83). Finally, we conduct an
experiment for detecting the structure of fallacies and discover that
state-of-the-art language models struggle with detecting fallacy templates
(0.47 accuracy). To facilitate research on fallacies, we make our dataset and
guidelines publicly available. | cs.CL | Flee the Flaw: Annotating the Underlying Logic of Fallacious Arguments Through Templates and Slot-filling |
2024-06-18T08:40:29Z | http://arxiv.org/pdf/2406.12399v1 | http://arxiv.org/abs/2406.12399v1 | Mae Sosto, Alberto Barrón-Cedeño | 2024-06-18T08:40:29Z | cs.CL, cs.AI, cs.CY | With the increasing role of Natural Language Processing (NLP) in various
applications, challenges concerning bias and stereotype perpetuation are
accentuated, which often leads to hate speech and harm. Despite existing
studies on sexism and misogyny, issues like homophobia and transphobia remain
underexplored and often adopt binary perspectives, putting the safety of
LGBTQIA+ individuals at high risk in online spaces. In this paper, we assess
the potential harm caused by sentence completions generated by English large
language models (LLMs) concerning LGBTQIA+ individuals. This is achieved using
QueerBench, our new assessment framework, which employs a template-based
approach and a Masked Language Modeling (MLM) task. The analysis indicates that
large language models tend to exhibit discriminatory behaviour more frequently
towards individuals within the LGBTQIA+ community, reaching a difference gap of
7.2% in the QueerBench score of harmfulness. | cs.CL | QueerBench: Quantifying Discrimination in Language Models Toward Queer Identities |
2024-06-18T08:38:59Z | http://arxiv.org/pdf/2406.12397v1 | http://arxiv.org/abs/2406.12397v1 | Jie Chen, Yupeng Zhang, Bingning Wang, Wayne Xin Zhao, Ji-Rong Wen, Weipeng Chen | 2024-06-18T08:38:59Z | cs.CL | Synthetic data has been proposed as a solution to address the issue of
high-quality data scarcity in the training of large language models (LLMs).
Studies have shown that synthetic data can effectively improve the performance
of LLMs on downstream benchmarks. However, despite its potential benefits, our
analysis suggests that there may be inherent flaws in synthetic data. The
uniform format of synthetic data can lead to pattern overfitting and cause
significant shifts in the output distribution, thereby reducing the model's
instruction-following capabilities. Our work delves into these specific flaws
associated with question-answer (Q-A) pairs, a prevalent type of synthetic
data, and presents a method based on unlearning techniques to mitigate these
flaws. The empirical results demonstrate the effectiveness of our approach,
which can reverse the instruction-following issues caused by pattern
overfitting without compromising performance on benchmarks at relatively low
cost. Our work has yielded key insights into the effective use of synthetic
data, aiming to promote more robust and efficient LLM training. | cs.CL | Unveiling the Flaws: Exploring Imperfections in Synthetic Data and Mitigation Strategies for Large Language Models |
2024-06-18T08:27:00Z | http://arxiv.org/pdf/2406.12946v1 | http://arxiv.org/abs/2406.12946v1 | Vahid Noroozi, Zhehuai Chen, Somshubra Majumdar, Steve Huang, Jagadeesh Balam, Boris Ginsburg | 2024-06-18T08:27:00Z | eess.AS, cs.AI, cs.CL, cs.LG | In this paper, we propose three methods for generating synthetic samples to
train and evaluate multimodal large language models capable of processing both
text and speech inputs. Addressing the scarcity of samples containing both
modalities, synthetic data generation emerges as a crucial strategy to enhance
the performance of such systems and facilitate the modeling of cross-modal
relationships between the speech and text domains. Our process employs large
language models to generate textual components and text-to-speech systems to
generate speech components. The proposed methods offer a practical and
effective means to expand the training dataset for these models. Experimental
results show progress in achieving an integrated understanding of text and
speech. We also highlight the potential of using unlabeled speech data to
generate synthetic samples comparable in quality to those with available
transcriptions, enabling the expansion of these models to more languages. | eess.AS | Instruction Data Generation and Unsupervised Adaptation for Speech Language Models |
2024-06-18T08:26:33Z | http://arxiv.org/pdf/2406.12389v1 | http://arxiv.org/abs/2406.12389v1 | Mia Huong Nguyen, Yasith Samaradivakara, Prasanth Sasikumar, Chitralekha Gupta, Suranga Nanayakkara | 2024-06-18T08:26:33Z | cs.CL | Emotion-Cause analysis has attracted the attention of researchers in recent
years. However, most existing datasets are limited in size and number of
emotion categories. They often focus on extracting parts of the document that
contain the emotion cause and fail to provide more abstractive, generalizable
root cause. To bridge this gap, we introduce a large-scale dataset of emotion
causes, derived from 9.8 million cleaned tweets over 15 years. We describe our
curation process, which includes a comprehensive pipeline for data gathering,
cleaning, labeling, and validation, ensuring the dataset's reliability and
richness. We extract emotion labels and provide abstractive summarization of
the events causing emotions. The final dataset comprises over 700,000 tweets
with corresponding emotion-cause pairs spanning 48 emotion classes, validated
by human evaluators. The novelty of our dataset stems from its broad spectrum
of emotion classes and the abstractive emotion cause that facilitates the
development of an emotion-cause knowledge graph for nuanced reasoning. Our
dataset will enable the design of emotion-aware systems that account for the
diverse emotional responses of different people for the same event. | cs.CL | EMO-KNOW: A Large Scale Dataset on Emotion and Emotion-cause |
2024-06-18T08:19:48Z | http://arxiv.org/pdf/2406.12387v1 | http://arxiv.org/abs/2406.12387v1 | Tejumade Afonja, Tobi Olatunji, Sewade Ogun, Naome A. Etori, Abraham Owodunni, Moshood Yekini | 2024-06-18T08:19:48Z | eess.AS, cs.CL, cs.SD | Recent strides in automatic speech recognition (ASR) have accelerated their
application in the medical domain where their performance on accented medical
named entities (NE) such as drug names, diagnoses, and lab results, is largely
unknown. We rigorously evaluate multiple ASR models on a clinical English
dataset of 93 African accents. Our analysis reveals that despite some models
achieving low overall word error rates (WER), errors in clinical entities are
higher, potentially posing substantial risks to patient safety. To empirically
demonstrate this, we extract clinical entities from transcripts, develop a
novel algorithm to align ASR predictions with these entities, and compute
medical NE Recall, medical WER, and character error rate. Our results show that
fine-tuning on accented clinical speech improves medical WER by a wide margin
(25-34 % relative), improving their practical applicability in healthcare
environments. | eess.AS | Performant ASR Models for Medical Entities in Accented Speech |
2024-06-18T08:18:18Z | http://arxiv.org/pdf/2406.12386v1 | http://arxiv.org/abs/2406.12386v1 | Qiyao Wang, Jianguo Huang, Shule Lu, Yuan Lin, Kan Xu, Liang Yang, Hongfei Lin | 2024-06-18T08:18:18Z | cs.CL | The rapid development of Large Language Models (LLMs) in vertical domains,
including intellectual property (IP), lacks a specific evaluation benchmark for
assessing their understanding, application, and reasoning abilities. To fill
this gap, we introduce IPEval, the first evaluation benchmark tailored for IP
agency and consulting tasks. IPEval comprises 2657 multiple-choice questions
across four major dimensions: creation, application, protection, and management
of IP. These questions span patent rights (inventions, utility models,
designs), trademarks, copyrights, trade secrets, and other related laws.
Evaluation methods include zero-shot, 5-few-shot, and Chain of Thought (CoT)
for seven LLM types, predominantly in English or Chinese. Results show superior
English performance by models like GPT series and Qwen series, while
Chinese-centric LLMs excel in Chinese tests, albeit specialized IP LLMs lag
behind general-purpose ones. Regional and temporal aspects of IP underscore the
need for LLMs to grasp legal nuances and evolving laws. IPEval aims to
accurately gauge LLM capabilities in IP and spur development of specialized
models. Website: \url{https://ipeval.github.io/} | cs.CL | IPEval: A Bilingual Intellectual Property Agency Consultation Evaluation Benchmark for Large Language Models |
2024-06-18T08:14:28Z | http://arxiv.org/pdf/2406.12382v1 | http://arxiv.org/abs/2406.12382v1 | Huanxuan Liao, Yao Xu, Shizhu He, Yuanzhe Zhang, Yanchao Hao, Shengping Liu, Kang Liu, Jun Zhao | 2024-06-18T08:14:28Z | cs.CL | Large language models (LLMs) have acquired the ability to solve general tasks
by utilizing instruction finetuning (IFT). However, IFT still relies heavily on
instance training of extensive task data, which greatly limits the adaptability
of LLMs to real-world scenarios where labeled task instances are scarce and
broader task generalization becomes paramount. Contrary to LLMs, humans acquire
skills and complete tasks not merely through repeated practice but also by
understanding and following instructional guidelines. This paper is dedicated
to simulating human learning to address the shortcomings of instance training,
focusing on instruction learning to enhance cross-task generalization. Within
this context, we introduce Task Adapters Generation from Instructions (TAGI),
which automatically constructs the task-specific model in a parameter
generation manner based on the given task instructions without retraining for
unseen tasks. Specifically, we utilize knowledge distillation to enhance the
consistency between TAGI developed through Learning with Instruction and
task-specific models developed through Training with Instance, by aligning the
labels, output logits, and adapter parameters between them. TAGI is endowed
with cross-task generalization capabilities through a two-stage training
process that includes hypernetwork pretraining and finetuning. We evaluate TAGI
on the Super-Natural Instructions and P3 datasets. The experimental results
demonstrate that TAGI can match or even outperform traditional meta-trained
models and other hypernetwork models, while significantly reducing
computational requirements. | cs.CL | From Instance Training to Instruction Learning: Task Adapters Generation from Instructions |
2024-06-18T08:09:58Z | http://arxiv.org/pdf/2406.12381v2 | http://arxiv.org/abs/2406.12381v2 | Jincheng Zhou | 2024-07-02T09:21:03Z | cs.CL, cs.LG | Question-Options Generation (QOG) is a task that involves generating a set of
question-options pairs given context. This task has various applications,
including fine-tuning large models, information retrieval, and automated
multiple-choice question generation for education. In this paper, we develop
QOG models using three different methods based on fine-tuning
sequence-to-sequence language models (LMs). Experiments demonstrate that the
end-to-end QOG model is computationally efficient and stable during both
training and inference, outperforming other methods. Furthermore, our analysis
indicates that our QOG models are competitive on the QOG task compared to the
large language model Llama 3-8B. | cs.CL | QOG:Question and Options Generation based on Language Model |
2024-06-18T07:58:33Z | http://arxiv.org/pdf/2406.12373v2 | http://arxiv.org/abs/2406.12373v2 | Yichen Pan, Dehan Kong, Sida Zhou, Cheng Cui, Yifei Leng, Bing Jiang, Hangyu Liu, Yanyi Shang, Shuyan Zhou, Tongshuang Wu, Zhengyang Wu | 2024-06-27T16:56:13Z | cs.CL, cs.AI, cs.LG, 68T50, I.2.7 | For web agents to be practically useful, they must adapt to the continuously
evolving web environment characterized by frequent updates to user interfaces
and content. However, most existing benchmarks only capture the static aspects
of the web. To bridge this gap, we introduce WebCanvas, an innovative online
evaluation framework for web agents that effectively addresses the dynamic
nature of web interactions. WebCanvas contains three main components to
facilitate realistic assessments: (1) A novel evaluation metric which reliably
capture critical intermediate actions or states necessary for task completions
while disregarding noise caused by insignificant events or changed
web-elements. (2) A benchmark dataset called Mind2Web-Live, a refined version
of original Mind2Web static dataset containing 542 tasks with 2439 intermediate
evaluation states; (3) Lightweight and generalizable annotation tools and
testing pipelines that enables the community to collect and maintain the
high-quality, up-to-date dataset. Building on WebCanvas, we open-source an
agent framework with extensible modules for reasoning, providing a foundation
for the community to conduct online inference and evaluations. Our
best-performing agent achieves a task success rate of 23.1% and a task
completion rate of 48.8% on the Mind2Web-Live test set. Additionally, we
analyze the performance discrepancies across various websites, domains, and
experimental environments. We encourage the community to contribute further
insights on online agent evaluation, thereby advancing this field of research. | cs.CL | WebCanvas: Benchmarking Web Agents in Online Environments |
2024-06-18T07:51:44Z | http://arxiv.org/pdf/2406.12369v1 | http://arxiv.org/abs/2406.12369v1 | Sarah Alyami, Hamzah Luqman | 2024-06-18T07:51:44Z | cs.CL, cs.CV | Continuous Sign Language Recognition (CSLR) focuses on the interpretation of
a sequence of sign language gestures performed continually without pauses. In
this study, we conduct an empirical evaluation of recent deep learning CSLR
techniques and assess their performance across various datasets and sign
languages. The models selected for analysis implement a range of approaches for
extracting meaningful features and employ distinct training strategies. To
determine their efficacy in modeling different sign languages, these models
were evaluated using multiple datasets, specifically RWTH-PHOENIX-Weather-2014,
ArabSign, and GrSL, each representing a unique sign language. The performance
of the models was further tested with unseen signers and sentences. The
conducted experiments establish new benchmarks on the selected datasets and
provide valuable insights into the robustness and generalization of the
evaluated techniques under challenging scenarios. | cs.CL | A Comparative Study of Continuous Sign Language Recognition Techniques |
2024-06-18T07:44:13Z | http://arxiv.org/pdf/2406.12364v1 | http://arxiv.org/abs/2406.12364v1 | Harritxu Gete, Thierry Etchegoyhen | 2024-06-18T07:44:13Z | cs.CL | Neural Machine Translation models tend to perpetuate gender bias present in
their training data distribution. Context-aware models have been previously
suggested as a means to mitigate this type of bias. In this work, we examine
this claim by analysing in detail the translation of stereotypical professions
in English to German, and translation with non-informative context in Basque to
Spanish. Our results show that, although context-aware models can significantly
enhance translation accuracy for feminine terms, they can still maintain or
even amplify gender bias. These results highlight the need for more
fine-grained approaches to bias mitigation in Neural Machine Translation. | cs.CL | Does Context Help Mitigate Gender Bias in Neural Machine Translation? |
2024-06-18T07:40:18Z | http://arxiv.org/pdf/2406.12354v1 | http://arxiv.org/abs/2406.12354v1 | Minseok Choi, Kyunghyun Min, Jaegul Choo | 2024-06-18T07:40:18Z | cs.CL | Pretrained language models memorize vast amounts of information, including
private and copyrighted data, raising significant safety concerns. Retraining
these models after excluding sensitive data is prohibitively expensive, making
machine unlearning a viable, cost-effective alternative. Previous research has
focused on machine unlearning for monolingual models, but we find that
unlearning in one language does not necessarily transfer to others. This
vulnerability makes models susceptible to low-resource language attacks, where
sensitive information remains accessible in less dominant languages. This paper
presents a pioneering approach to machine unlearning for multilingual language
models, selectively erasing information across different languages while
maintaining overall performance. Specifically, our method employs an adaptive
unlearning scheme that assigns language-dependent weights to address different
language performances of multilingual language models. Empirical results
demonstrate the effectiveness of our framework compared to existing unlearning
baselines, setting a new standard for secure and adaptable multilingual
language models. | cs.CL | Cross-Lingual Unlearning of Selective Knowledge in Multilingual Language Models |
2024-06-18T07:28:15Z | http://arxiv.org/pdf/2406.12347v1 | http://arxiv.org/abs/2406.12347v1 | Nirmalendu Prakash, Lee Ka Wei Roy | 2024-06-18T07:28:15Z | cs.CL | Large Language Models (LLMs) such as Mistral and LLaMA have showcased
remarkable performance across various natural language processing (NLP) tasks.
Despite their success, these models inherit social biases from the diverse
datasets on which they are trained. This paper investigates the propagation of
biases within LLMs through a novel feature-based analytical approach. Drawing
inspiration from causal mediation analysis, we hypothesize the evolution of
bias-related features and validate them using interpretability techniques like
activation and attribution patching. Our contributions are threefold: (1) We
introduce and empirically validate a feature-based method for bias analysis in
LLMs, applied to LLaMA-2-7B, LLaMA-3-8B, and Mistral-7B-v0.3 with templates
from a professions dataset. (2) We extend our method to another form of gender
bias, demonstrating its generalizability. (3) We differentiate the roles of
MLPs and attention heads in bias propagation and implement targeted debiasing
using a counterfactual dataset. Our findings reveal the complex nature of bias
in LLMs and emphasize the necessity for tailored debiasing strategies, offering
a deeper understanding of bias mechanisms and pathways for effective
mitigation. | cs.CL | Interpreting Bias in Large Language Models: A Feature-Based Approach |
2024-06-18T07:03:34Z | http://arxiv.org/pdf/2406.12336v1 | http://arxiv.org/abs/2406.12336v1 | Sujoy Roychowdhury, Sumit Soman, H. G. Ranjani, Vansh Chhabra, Neeraj Gunda, Subhadip Bandyopadhyay, Sai Krishna Bala | 2024-06-18T07:03:34Z | cs.CL, cs.LG, 68T50, I.2.7 | A plethora of sentence embedding models makes it challenging to choose one,
especially for domains such as telecom, rich with specialized vocabulary. We
evaluate multiple embeddings obtained from publicly available models and their
domain-adapted variants, on both point retrieval accuracies as well as their
(95\%) confidence intervals. We establish a systematic method to obtain
thresholds for similarity scores for different embeddings. We observe that
fine-tuning improves mean bootstrapped accuracies as well as tightens
confidence intervals. The pre-training combined with fine-tuning makes
confidence intervals even tighter. To understand these variations, we analyse
and report significant correlations between the distributional overlap between
top-$K$, correct and random sentence similarities with retrieval accuracies and
similarity thresholds. Following current literature, we analyze if retrieval
accuracy variations can be attributed to isotropy of embeddings. Our
conclusions are that isotropy of embeddings (as measured by two independent
state-of-the-art isotropy metric definitions) cannot be attributed to better
retrieval performance. However, domain adaptation which improves retrieval
accuracies also improves isotropy. We establish that domain adaptation moves
domain specific embeddings further away from general domain embeddings. | cs.CL | A Compass for Navigating the World of Sentence Embeddings for the Telecom Domain |
2024-06-18T07:01:11Z | http://arxiv.org/pdf/2406.12335v1 | http://arxiv.org/abs/2406.12335v1 | Zhiyu Guo, Hidetaka Kamigaito, Taro Watanabe | 2024-06-18T07:01:11Z | cs.CL, cs.LG | Scaling the context size of large language models (LLMs) enables them to
perform various new tasks, e.g., book summarization. However, the memory cost
of the Key and Value (KV) cache in attention significantly limits the practical
applications of LLMs. Recent works have explored token pruning for KV cache
reduction in LLMs, relying solely on attention scores as a token importance
indicator. However, our investigation into value vector norms revealed a
notably non-uniform pattern questioning their reliance only on attention
scores. Inspired by this, we propose a new method: Value-Aware Token Pruning
(VATP) which uses both attention scores and the $ \ell_{1} $ norm of value
vectors to evaluate token importance. Extensive experiments on LLaMA2-7B-chat
and Vicuna-v1.5-7B across 16 LongBench tasks demonstrate VATP's superior
performance. | cs.CL | Attention Score is not All You Need for Token Importance Indicator in KV Cache Reduction: Value Also Matters |
2024-06-18T06:54:28Z | http://arxiv.org/pdf/2406.12331v1 | http://arxiv.org/abs/2406.12331v1 | Weizhi Fei, Xueyan Niu, Guoqing Xie, Yanhua Zhang, Bo Bai, Lei Deng, Wei Han | 2024-06-18T06:54:28Z | cs.CL, cs.AI | Current Large Language Models (LLMs) face inherent limitations due to their
pre-defined context lengths, which impede their capacity for multi-hop
reasoning within extensive textual contexts. While existing techniques like
Retrieval-Augmented Generation (RAG) have attempted to bridge this gap by
sourcing external information, they fall short when direct answers are not
readily available. We introduce a novel approach that re-imagines information
retrieval through dynamic in-context editing, inspired by recent breakthroughs
in knowledge editing. By treating lengthy contexts as malleable external
knowledge, our method interactively gathers and integrates relevant
information, thereby enabling LLMs to perform sophisticated reasoning steps.
Experimental results demonstrate that our method effectively empowers
context-limited LLMs, such as Llama2, to engage in multi-hop reasoning with
improved performance, which outperforms state-of-the-art context window
extrapolation methods and even compares favorably to more advanced commercial
long-context models. Our interactive method not only enhances reasoning
capabilities but also mitigates the associated training and computational
costs, making it a pragmatic solution for enhancing LLMs' reasoning within
expansive contexts. | cs.CL | Retrieval Meets Reasoning: Dynamic In-Context Editing for Long-Text Understanding |
2024-06-18T06:54:05Z | http://arxiv.org/pdf/2406.12329v1 | http://arxiv.org/abs/2406.12329v1 | Minseok Choi, Daniel Rim, Dohyun Lee, Jaegul Choo | 2024-06-18T06:54:05Z | cs.CL | Instruction-following large language models (LLMs), such as ChatGPT, have
become increasingly popular with the general audience, many of whom are
incorporating them into their daily routines. However, these LLMs inadvertently
disclose personal or copyrighted information, which calls for a machine
unlearning method to remove selective knowledge. Previous attempts sought to
forget the link between the target information and its associated entities, but
it rather led to generating undesirable responses about the target,
compromising the end-user experience. In this work, we propose SNAP, an
innovative framework designed to selectively unlearn information by 1) training
an LLM with negative instructions to generate obliterated responses, 2)
augmenting hard positives to retain the original LLM performance, and 3)
applying the novel Wasserstein regularization to ensure adequate deviation from
the initial weights of the LLM. We evaluate our framework on various NLP
benchmarks and demonstrate that our approach retains the original LLM
capabilities, while successfully unlearning the specified information. | cs.CL | SNAP: Unlearning Selective Knowledge in Large Language Models with Negative Instructions |
2024-06-18T06:43:46Z | http://arxiv.org/pdf/2406.12321v1 | http://arxiv.org/abs/2406.12321v1 | Alessandro Conti, Enrico Fini, Paolo Rota, Yiming Wang, Massimiliano Mancini, Elisa Ricci | 2024-06-18T06:43:46Z | cs.AI, cs.CL, cs.CV | Assessing the capabilities of large multimodal models (LMMs) often requires
the creation of ad-hoc evaluations. Currently, building new benchmarks requires
tremendous amounts of manual work for each specific analysis. This makes the
evaluation process tedious and costly. In this paper, we present APEx,
Automatic Programming of Experiments, the first framework for automatic
benchmarking of LMMs. Given a research question expressed in natural language,
APEx leverages a large language model (LLM) and a library of pre-specified
tools to generate a set of experiments for the model at hand, and progressively
compile a scientific report. The report drives the testing procedure: based on
the current status of the investigation, APEx chooses which experiments to
perform and whether the results are sufficient to draw conclusions. Finally,
the LLM refines the report, presenting the results to the user in natural
language. Thanks to its modularity, our framework is flexible and extensible as
new tools become available. Empirically, APEx reproduces the findings of
existing studies while allowing for arbitrary analyses and hypothesis testing. | cs.AI | Automatic benchmarking of large multimodal models via iterative experiment programming |
2024-06-18T06:43:04Z | http://arxiv.org/pdf/2406.12319v1 | http://arxiv.org/abs/2406.12319v1 | Hawon Jeong, ChaeHun Park, Jimin Hong, Jaegul Choo | 2024-06-18T06:43:04Z | cs.CL | Pairwise evaluation using large language models (LLMs) is widely used for
evaluating natural language generation (NLG) tasks. However, the reliability of
LLMs is often compromised by biases, such as favoring verbosity and
authoritative tone. In the study, we focus on the comparison of two LLM-based
evaluation approaches, pointwise and pairwise. Our findings demonstrate that
pointwise evaluators exhibit more robustness against undesirable preferences.
Further analysis reveals that pairwise evaluators can accurately identify the
shortcomings of low-quality outputs even when their judgment is incorrect.
These results indicate that LLMs are more severely influenced by their bias in
a pairwise evaluation setup. To mitigate this, we propose a hybrid method that
integrates pointwise reasoning into pairwise evaluation. Experimental results
show that our method enhances the robustness of pairwise evaluators against
adversarial samples while preserving accuracy on normal samples. | cs.CL | PRePair: Pointwise Reasoning Enhance Pairwise Evaluating for Robust Instruction-Following Assessments |
2024-06-18T06:39:41Z | http://arxiv.org/pdf/2406.12317v1 | http://arxiv.org/abs/2406.12317v1 | Hayato Futami, Siddhant Arora, Yosuke Kashiwagi, Emiru Tsunoo, Shinji Watanabe | 2024-06-18T06:39:41Z | cs.CL, eess.AS | Recently, multi-task spoken language understanding (SLU) models have emerged,
designed to address various speech processing tasks. However, these models
often rely on a large number of parameters. Also, they often encounter
difficulties in adapting to new data for a specific task without experiencing
catastrophic forgetting of previously trained tasks. In this study, we propose
finding task-specific subnetworks within a multi-task SLU model via neural
network pruning. In addition to model compression, we expect that the
forgetting of previously trained tasks can be mitigated by updating only a
task-specific subnetwork. We conduct experiments on top of the state-of-the-art
multi-task SLU model ``UniverSLU'', trained for several tasks such as emotion
recognition (ER), intent classification (IC), and automatic speech recognition
(ASR). We show that pruned models were successful in adapting to additional ASR
or IC data with minimal performance degradation on previously trained tasks. | cs.CL | Finding Task-specific Subnetworks in Multi-task Spoken Language Understanding Model |
2024-06-18T06:28:06Z | http://arxiv.org/pdf/2406.12307v1 | http://arxiv.org/abs/2406.12307v1 | Seungbin Yang, ChaeHun Park, Taehee Kim, Jaegul Choo | 2024-06-18T06:28:06Z | cs.CL | Recent advancements in integrating large language models (LLMs) with tools
have allowed the models to interact with real-world environments. However,
these tool-augmented LLMs often encounter incomplete scenarios when users
provide partial information or the necessary tools are unavailable. Recognizing
and managing such scenarios is crucial for LLMs to ensure their reliability,
but this exploration remains understudied. This study examines whether LLMs can
identify incomplete conditions and appropriately determine when to refrain from
using tools. To this end, we address a dataset by manipulating instances from
two datasets by removing necessary tools or essential information for tool
invocation. We confirm that most LLMs are challenged to identify the additional
information required to utilize specific tools and the absence of appropriate
tools. Our research can contribute to advancing reliable LLMs by addressing
scenarios that commonly arise during interactions between humans and LLMs. | cs.CL | Can Tool-augmented Large Language Models be Aware of Incomplete Conditions? |
2024-06-18T06:24:26Z | http://arxiv.org/pdf/2406.12304v1 | http://arxiv.org/abs/2406.12304v1 | Linhao Zhang, Li Jin, Guangluan Xu, Xiaoyu Li, Xian Sun | 2024-06-18T06:24:26Z | cs.CL, 68U15, I.2.7 | Counter-narratives, which are direct responses consisting of non-aggressive
fact-based arguments, have emerged as a highly effective approach to combat the
proliferation of hate speech. Previous methodologies have primarily focused on
fine-tuning and post-editing techniques to ensure the fluency of generated
contents, while overlooking the critical aspects of individualization and
relevance concerning the specific hatred targets, such as LGBT groups,
immigrants, etc. This research paper introduces a novel framework based on
contrastive optimal transport, which effectively addresses the challenges of
maintaining target interaction and promoting diversification in generating
counter-narratives. Firstly, an Optimal Transport Kernel (OTK) module is
leveraged to incorporate hatred target information in the token
representations, in which the comparison pairs are extracted between original
and transported features. Secondly, a self-contrastive learning module is
employed to address the issue of model degeneration. This module achieves this
by generating an anisotropic distribution of token representations. Finally, a
target-oriented search method is integrated as an improved decoding strategy to
explicitly promote domain relevance and diversification in the inference
process. This strategy modifies the model's confidence score by considering
both token similarity and target relevance. Quantitative and qualitative
experiments have been evaluated on two benchmark datasets, which demonstrate
that our proposed model significantly outperforms current methods evaluated by
metrics from multiple aspects. | cs.CL | COT: A Generative Approach for Hate Speech Counter-Narratives via Contrastive Optimal Transport |
2024-06-18T05:59:28Z | http://arxiv.org/pdf/2406.12295v1 | http://arxiv.org/abs/2406.12295v1 | Kaiyan Zhang, Jianyu Wang, Ning Ding, Biqing Qi, Ermo Hua, Xingtai Lv, Bowen Zhou | 2024-06-18T05:59:28Z | cs.CL | Large Language Models (LLMs) demonstrate impressive performance in diverse
applications, yet they face significant drawbacks, including high inference
latency, expensive training cost, and generation of hallucination.
Collaborative decoding between large and small language models (SLMs) offers a
novel approach to address these challenges. Inspired by dual-process cognitive
theory, we integrate these methods into a unified framework termed Fast and
Slow Generating (FS-GEN). This paper explores several techniques within the
FS-GEN framework, including speculative decoding, contrastive decoding, and
emulator or proxy fine-tuning. We provide a comprehensive analysis of these
methodologies, offering insights into their similarities and differences under
this framework. Our study delves into the differential knowledge capabilities
of LLMs versus SLMs through the FS-GEN lens, revealing that fewer than 20% of
collaborative interactions are required across various methods. These
interactions adhere to a scaling law relative to the parameter ratios, thereby
facilitating predictable collaboration. Furthermore, we investigate the
specific positions where collaboration is most effective from an uncertainty
perspective, yielding novel insights that could refine FS-GEN methods. Our
findings reveal that the essential difference between models of different sizes
lies in the uncertainty of the next token prediction, where interventions by
larger models are most needed to assist the smaller ones. Code for
Reproduction: https://github.com/TsinghuaC3I/FS-GEN | cs.CL | Fast and Slow Generating: An Empirical Study on Large and Small Language Models Collaborative Decoding |
2024-06-18T05:11:35Z | http://arxiv.org/pdf/2406.12277v1 | http://arxiv.org/abs/2406.12277v1 | Xin Zhao, Naoki Yoshinaga, Daisuke Oba | 2024-06-18T05:11:35Z | cs.CL | Large language models (LLMs) face issues in handling factual knowledge,
making it vital to evaluate their true ability to understand facts. In this
study, we introduce knowledge probing frameworks, BELIEF(-ICL), to evaluate the
knowledge understanding ability of not only encoder-based PLMs but also
decoder-based PLMs from diverse perspectives. BELIEFs utilize a multi-prompt
dataset to evaluate PLM's accuracy, consistency, and reliability in factual
knowledge understanding. To provide a more reliable evaluation with BELIEFs, we
semi-automatically create MyriadLAMA, which has more diverse prompts than
existing datasets. We validate the effectiveness of BELIEFs in correctly and
comprehensively evaluating PLM's factual understanding ability through
extensive evaluations. We further investigate key factors in learning facts in
LLMs, and reveal the limitation of the prompt-based knowledge probing. The
dataset is anonymously publicized. | cs.CL | What Matters in Learning Facts in Language Models? Multifaceted Knowledge Probing with Diverse Multi-Prompt Datasets |
2024-06-18T05:10:38Z | http://arxiv.org/pdf/2406.12276v1 | http://arxiv.org/abs/2406.12276v1 | Tanmay Gupta, Luca Weihs, Aniruddha Kembhavi | 2024-06-18T05:10:38Z | cs.AI, cs.CL, cs.SE | We present CodeNav, an LLM agent that navigates and leverages previously
unseen code repositories to solve user queries. In contrast to tool-use LLM
agents that require ``registration'' of all relevant tools via manual
descriptions within the LLM context, CodeNav automatically indexes and searches
over code blocks in the target codebase, finds relevant code snippets, imports
them, and uses them to iteratively generate a solution with execution feedback.
To highlight the core-capabilities of CodeNav, we first showcase three case
studies where we use CodeNav for solving complex user queries using three
diverse codebases. Next, on three benchmarks, we quantitatively compare the
effectiveness of code-use (which only has access to the target codebase) to
tool-use (which has privileged access to all tool names and descriptions).
Finally, we study the effect of varying kinds of tool and library descriptions
on code-use performance, as well as investigate the advantage of the agent
seeing source code as opposed to natural descriptions of code. All code will be
made open source under a permissive license. | cs.AI | CodeNav: Beyond tool-use to using real-world codebases with LLM agents |
2024-06-18T05:03:23Z | http://arxiv.org/pdf/2406.12274v1 | http://arxiv.org/abs/2406.12274v1 | Somnath Banerjee, Soham Tripathy, Sayan Layek, Shanu Kumar, Animesh Mukherjee, Rima Hazra | 2024-06-18T05:03:23Z | cs.CL | Safety-aligned language models often exhibit fragile and imbalanced safety
mechanisms, increasing the likelihood of generating unsafe content. In
addition, incorporating new knowledge through editing techniques to language
models can further compromise safety. To address these issues, we propose
SafeInfer, a context-adaptive, decoding-time safety alignment strategy for
generating safe responses to user queries. SafeInfer comprises two phases: the
safety amplification phase, which employs safe demonstration examples to adjust
the model's hidden states and increase the likelihood of safer outputs, and the
safety-guided decoding phase, which influences token selection based on
safety-optimized distributions, ensuring the generated content complies with
ethical guidelines. Further, we present HarmEval, a novel benchmark for
extensive safety evaluations, designed to address potential misuse scenarios in
accordance with the policies of leading AI tech giants. | cs.CL | SafeInfer: Context Adaptive Decoding Time Safety Alignment for Large Language Models |
2024-06-18T04:55:09Z | http://arxiv.org/pdf/2406.12269v1 | http://arxiv.org/abs/2406.12269v1 | Kwangwook Seo, Jinyoung Yeo, Dongha Lee | 2024-06-18T04:55:09Z | cs.CL | Implicit knowledge hidden within the explicit table cells, such as data
insights, is the key to generating a high-quality table summary. However,
unveiling such implicit knowledge is a non-trivial task. Due to the complex
nature of structured tables, it is challenging even for large language models
(LLMs) to mine the implicit knowledge in an insightful and faithful manner. To
address this challenge, we propose a novel table reasoning framework
Question-then-Pinpoint. Our work focuses on building a plug-and-play table
reasoner that can self-question the insightful knowledge and answer it by
faithfully pinpointing evidence on the table to provide explainable guidance
for the summarizer. To train a reliable reasoner, we collect table knowledge by
guiding a teacher LLM to follow the coarse-to-fine reasoning paths and refine
it through two quality enhancement strategies to selectively distill the
high-quality knowledge to the reasoner. Extensive experiments on two table
summarization datasets, including our newly proposed InsTaSumm, validate the
general effectiveness of our framework. | cs.CL | Unveiling Implicit Table Knowledge with Question-Then-Pinpoint Reasoner for Insightful Table Summarization |
2024-06-18T04:46:55Z | http://arxiv.org/pdf/2406.12266v2 | http://arxiv.org/abs/2406.12266v2 | Jiashuo Wang, Yang Xiao, Yanran Li, Changhe Song, Chunpu Xu, Chenhao Tan, Wenjie Li | 2024-06-20T22:06:47Z | cs.CL | Although there is a growing belief that LLMs can be used as therapists,
exploring LLMs' capabilities and inefficacy, particularly from the client's
perspective, is limited. This work focuses on a client-centered assessment of
LLM therapists with the involvement of simulated clients, a standard approach
in clinical medical education. However, there are two challenges when applying
the approach to assess LLM therapists at scale. Ethically, asking humans to
frequently mimic clients and exposing them to potentially harmful LLM outputs
can be risky and unsafe. Technically, it can be difficult to consistently
compare the performances of different LLM therapists interacting with the same
client. To this end, we adopt LLMs to simulate clients and propose ClientCAST,
a client-centered approach to assessing LLM therapists by client simulation.
Specifically, the simulated client is utilized to interact with LLM therapists
and complete questionnaires related to the interaction. Based on the
questionnaire results, we assess LLM therapists from three client-centered
aspects: session outcome, therapeutic alliance, and self-reported feelings. We
conduct experiments to examine the reliability of ClientCAST and use it to
evaluate LLMs therapists implemented by Claude-3, GPT-3.5, LLaMA3-70B, and
Mixtral 8*7B. Codes are released at https://github.com/wangjs9/ClientCAST. | cs.CL | Towards a Client-Centered Assessment of LLM Therapists by Client Simulation |
2024-06-18T04:39:40Z | http://arxiv.org/pdf/2406.12263v1 | http://arxiv.org/abs/2406.12263v1 | Lin Ai, Tharindu Kumarage, Amrita Bhattacharjee, Zizhou Liu, Zheng Hui, Michael Davinroy, James Cook, Laura Cassani, Kirill Trapeznikov, Matthias Kirchner, Arslan Basharat, Anthony Hoogs, Joshua Garland, Huan Liu, Julia Hirschberg | 2024-06-18T04:39:40Z | cs.CL | The proliferation of Large Language Models (LLMs) poses challenges in
detecting and mitigating digital deception, as these models can emulate human
conversational patterns and facilitate chat-based social engineering (CSE)
attacks. This study investigates the dual capabilities of LLMs as both
facilitators and defenders against CSE threats. We develop a novel dataset,
SEConvo, simulating CSE scenarios in academic and recruitment contexts, and
designed to examine how LLMs can be exploited in these situations. Our findings
reveal that, while off-the-shelf LLMs generate high-quality CSE content, their
detection capabilities are suboptimal, leading to increased operational costs
for defense. In response, we propose ConvoSentinel, a modular defense pipeline
that improves detection at both the message and the conversation levels,
offering enhanced adaptability and cost-effectiveness. The retrieval-augmented
module in ConvoSentinel identifies malicious intent by comparing messages to a
database of similar conversations, enhancing CSE detection at all stages. Our
study highlights the need for advanced strategies to leverage LLMs in
cybersecurity. | cs.CL | Defending Against Social Engineering Attacks in the Age of LLMs |
2024-06-18T04:25:56Z | http://arxiv.org/pdf/2406.12260v1 | http://arxiv.org/abs/2406.12260v1 | Kukjin Choi, Jihun Yi, Jisoo Mok, Sungroh Yoon | 2024-06-18T04:25:56Z | cs.LG, cs.AI, cs.CL | Continuous efforts are being made to advance anomaly detection in various
manufacturing processes to increase the productivity and safety of industrial
sites. Deep learning replaced rule-based methods and recently emerged as a
promising method for anomaly detection in diverse industries. However, in the
real world, the scarcity of abnormal data and difficulties in obtaining labeled
data create limitations in the training of detection models. In this study, we
addressed these shortcomings by proposing a learnable data augmentation-based
time-series anomaly detection (LATAD) technique that is trained in a
self-supervised manner. LATAD extracts discriminative features from time-series
data through contrastive learning. At the same time, learnable data
augmentation produces challenging negative samples to enhance learning
efficiency. We measured anomaly scores of the proposed technique based on
latent feature similarities. As per the results, LATAD exhibited comparable or
improved performance to the state-of-the-art anomaly detection assessments on
several benchmark datasets and provided a gradient-based diagnosis technique to
help identify root causes. | cs.LG | Self-Supervised Time-Series Anomaly Detection Using Learnable Data Augmentation |
2024-06-18T04:07:13Z | http://arxiv.org/pdf/2406.12255v1 | http://arxiv.org/abs/2406.12255v1 | Lijie Hu, Liang Liu, Shu Yang, Xin Chen, Hongru Xiao, Mengdi Li, Pan Zhou, Muhammad Asif Ali, Di Wang | 2024-06-18T04:07:13Z | cs.CL, cs.AI, cs.HC, cs.LG | Chain-of-Thought (CoT) holds a significant place in augmenting the reasoning
performance for large language models (LLMs). While some studies focus on
improving CoT accuracy through methods like retrieval enhancement, yet a
rigorous explanation for why CoT achieves such success remains unclear. In this
paper, we analyze CoT methods under two different settings by asking the
following questions: (1) For zero-shot CoT, why does prompting the model with
"let's think step by step" significantly impact its outputs? (2) For few-shot
CoT, why does providing examples before questioning the model could
substantially improve its reasoning ability? To answer these questions, we
conduct a top-down explainable analysis from the Hopfieldian view and propose a
Read-and-Control approach for controlling the accuracy of CoT. Through
extensive experiments on seven datasets for three different tasks, we
demonstrate that our framework can decipher the inner workings of CoT, provide
reasoning error localization, and control to come up with the correct reasoning
path. | cs.CL | A Hopfieldian View-based Interpretation for Chain-of-Thought Reasoning |
2024-06-18T03:59:26Z | http://arxiv.org/pdf/2406.12252v1 | http://arxiv.org/abs/2406.12252v1 | Haotian Xia, Zhengbang Yang, Yun Zhao, Yuqing Wang, Jingxi Li, Rhys Tracy, Zhuangdi Zhu, Yuan-fang Wang, Hanjie Chen, Weining Shen | 2024-06-18T03:59:26Z | cs.CL | Recent integration of Natural Language Processing (NLP) and multimodal models
has advanced the field of sports analytics. This survey presents a
comprehensive review of the datasets and applications driving these innovations
post-2020. We overviewed and categorized datasets into three primary types:
language-based, multimodal, and convertible datasets. Language-based and
multimodal datasets are for tasks involving text or multimodality (e.g., text,
video, audio), respectively. Convertible datasets, initially single-modal
(video), can be enriched with additional annotations, such as explanations of
actions and video descriptions, to become multimodal, offering future potential
for richer and more diverse applications. Our study highlights the
contributions of these datasets to various applications, from improving fan
experiences to supporting tactical analysis and medical diagnostics. We also
discuss the challenges and future directions in dataset development,
emphasizing the need for diverse, high-quality data to support real-time
processing and personalized user experiences. This survey provides a
foundational resource for researchers and practitioners aiming to leverage NLP
and multimodal models in sports, offering insights into current trends and
future opportunities in the field. | cs.CL | Language and Multimodal Models in Sports: A Survey of Datasets and Applications |
2024-06-18T03:57:49Z | http://arxiv.org/pdf/2406.12251v1 | http://arxiv.org/abs/2406.12251v1 | Chenyuan Wu, Gangwei Jiang, Defu Lian | 2024-06-18T03:57:49Z | cs.CL, cs.AI, cs.LG | Lifelong prompt tuning has significantly advanced parameter-efficient
lifelong learning with its efficiency and minimal storage demands on various
tasks. Our empirical studies, however, highlights certain transferability
constraints in the current methodologies: a universal algorithm that guarantees
consistent positive transfer across all tasks is currently unattainable,
especially when dealing dissimilar tasks that may engender negative transfer.
Identifying the misalignment between algorithm selection and task specificity
as the primary cause of negative transfer, we present the Similarity Heuristic
Lifelong Prompt Tuning (SHLPT) framework. This innovative strategy partitions
tasks into two distinct subsets by harnessing a learnable similarity metric,
thereby facilitating fruitful transfer from tasks regardless of their
similarity or dissimilarity. Additionally, SHLPT incorporates a parameter pool
to combat catastrophic forgetting effectively. Our experiments shows that SHLPT
outperforms state-of-the-art techniques in lifelong learning benchmarks and
demonstrates robustness against negative transfer in diverse task sequences. | cs.CL | Mitigate Negative Transfer with Similarity Heuristic Lifelong Prompt Tuning |
2024-06-18T03:42:00Z | http://arxiv.org/pdf/2406.12246v2 | http://arxiv.org/abs/2406.12246v2 | Byung-Kwan Lee, Sangyun Chung, Chae Won Kim, Beomchan Park, Yong Man Ro | 2024-06-19T21:40:03Z | cs.LG, cs.CL, cs.CV | Large language and vision models (LLVMs) have been driven by the
generalization power of large language models (LLMs) and the advent of visual
instruction tuning. Along with scaling them up directly, these models enable
LLVMs to showcase powerful vision language (VL) performances by covering
diverse tasks via natural language instructions. However, existing open-source
LLVMs that perform comparably to closed-source LLVMs such as GPT-4V are often
considered too large (e.g., 26B, 34B, and 110B parameters), having a larger
number of layers. These large models demand costly, high-end resources for both
training and inference. To address this issue, we present a new efficient LLVM
family with 1.8B, 3.8B, and 7B LLM model sizes, Traversal of Layers (TroL),
which enables the reuse of layers in a token-wise manner. This layer traversing
technique simulates the effect of looking back and retracing the answering
stream while increasing the number of forward propagation layers without
physically adding more layers. We demonstrate that TroL employs a simple layer
traversing approach yet efficiently outperforms the open-source LLVMs with
larger model sizes and rivals the performances of the closed-source LLVMs with
substantial sizes. | cs.LG | TroL: Traversal of Layers for Large Language and Vision Models |
2024-06-18T03:27:09Z | http://arxiv.org/pdf/2406.12238v1 | http://arxiv.org/abs/2406.12238v1 | Haoyan Yang, Zhitao Li, Yong Zhang, Jianzong Wang, Ning Cheng, Ming Li, Jing Xiao | 2024-06-18T03:27:09Z | cs.CL | This paper introduces a novel privacy-preservation framework named PFID for
LLMs that addresses critical privacy concerns by localizing user data through
model sharding and singular value decomposition. When users are interacting
with LLM systems, their prompts could be subject to being exposed to
eavesdroppers within or outside LLM system providers who are interested in
collecting users' input. In this work, we proposed a framework to camouflage
user input, so as to alleviate privacy issues. Our framework proposes to place
model shards on the client and the public server, we sent compressed hidden
states instead of prompts to and from servers. Clients have held back
information that can re-privatized the hidden states so that overall system
performance is comparable to traditional LLMs services. Our framework was
designed to be communication efficient, computation can be delegated to the
local client so that the server's computation burden can be lightened. We
conduct extensive experiments on machine translation tasks to verify our
framework's performance. | cs.CL | PFID: Privacy First Inference Delegation Framework for LLMs |
2024-06-18T03:14:22Z | http://arxiv.org/pdf/2406.12233v1 | http://arxiv.org/abs/2406.12233v1 | Young Jin Ahn, Jungwoo Park, Sangha Park, Jonghyun Choi, Kee-Eung Kim | 2024-06-18T03:14:22Z | cs.AI, cs.CL, cs.CV | Visual Speech Recognition (VSR) stands at the intersection of computer vision
and speech recognition, aiming to interpret spoken content from visual cues. A
prominent challenge in VSR is the presence of homophenes-visually similar lip
gestures that represent different phonemes. Prior approaches have sought to
distinguish fine-grained visemes by aligning visual and auditory semantics, but
often fell short of full synchronization. To address this, we present SyncVSR,
an end-to-end learning framework that leverages quantized audio for frame-level
crossmodal supervision. By integrating a projection layer that synchronizes
visual representation with acoustic data, our encoder learns to generate
discrete audio tokens from a video sequence in a non-autoregressive manner.
SyncVSR shows versatility across tasks, languages, and modalities at the cost
of a forward pass. Our empirical evaluations show that it not only achieves
state-of-the-art results but also reduces data usage by up to ninefold. | cs.AI | SyncVSR: Data-Efficient Visual Speech Recognition with End-to-End Crossmodal Audio Token Synchronization |
2024-06-18T03:11:43Z | http://arxiv.org/pdf/2406.12232v1 | http://arxiv.org/abs/2406.12232v1 | Huy Nghiem, John Prindle, Jieyu Zhao, Hal Daumé III | 2024-06-18T03:11:43Z | cs.AI, cs.CL | Social science research has shown that candidates with names indicative of
certain races or genders often face discrimination in employment practices.
Similarly, Large Language Models (LLMs) have demonstrated racial and gender
biases in various applications. In this study, we utilize GPT-3.5-Turbo and
Llama 3-70B-Instruct to simulate hiring decisions and salary recommendations
for candidates with 320 first names that strongly signal their race and gender,
across over 750,000 prompts. Our empirical results indicate a preference among
these models for hiring candidates with White female-sounding names over other
demographic groups across 40 occupations. Additionally, even among candidates
with identical qualifications, salary recommendations vary by as much as 5%
between different subgroups. A comparison with real-world labor data reveals
inconsistent alignment with U.S. labor market characteristics, underscoring the
necessity of risk investigation of LLM-powered systems. | cs.AI | "You Gotta be a Doctor, Lin": An Investigation of Name-Based Bias of Large Language Models in Employment Recommendations |
2024-06-18T03:08:01Z | http://arxiv.org/pdf/2406.12230v1 | http://arxiv.org/abs/2406.12230v1 | Hua Yang, Duohai Li, Shiman Li | 2024-06-18T03:08:01Z | cs.CL, cs.AI | Transformers excel in Natural Language Processing (NLP) due to their prowess
in capturing long-term dependencies but suffer from exponential resource
consumption with increasing sequence lengths. To address these challenges, we
propose MCSD model, an efficient language model with linear scaling and fast
inference speed. MCSD model leverages diverse feature fusion, primarily through
the multi-channel slope and decay (MCSD) block, to robustly represent features.
This block comprises slope and decay sections that extract features across
diverse temporal receptive fields, facilitating capture of both local and
global information. In addition, MCSD block conducts element-wise fusion of
diverse features to further enhance the delicate feature extraction capability.
For inference, we formulate the inference process into a recurrent
representation, slashing space complexity to $O(1)$ and time complexity to
$O(N)$ respectively. Our experiments show that MCSD attains higher throughput
and lower GPU memory consumption compared to Transformers, while maintaining
comparable performance to larger-scale language learning models on benchmark
tests. These attributes position MCSD as a promising base for edge deployment
and embodied intelligence. | cs.CL | MCSD: An Efficient Language Model with Diverse Fusion |
2024-06-18T02:44:56Z | http://arxiv.org/pdf/2406.12223v1 | http://arxiv.org/abs/2406.12223v1 | Yunze Xiao, Yujia Hu, Kenny Tsu Wei Choo, Roy Ka-wei Lee | 2024-06-18T02:44:56Z | cs.CL, cs.CY | Detecting hate speech and offensive language is essential for maintaining a
safe and respectful digital environment. This study examines the limitations of
state-of-the-art large language models (LLMs) in identifying offensive content
within systematically perturbed data, with a focus on Chinese, a language
particularly susceptible to such perturbations. We introduce
\textsf{ToxiCloakCN}, an enhanced dataset derived from ToxiCN, augmented with
homophonic substitutions and emoji transformations, to test the robustness of
LLMs against these cloaking perturbations. Our findings reveal that existing
models significantly underperform in detecting offensive content when these
perturbations are applied. We provide an in-depth analysis of how different
types of offensive content are affected by these perturbations and explore the
alignment between human and model explanations of offensiveness. Our work
highlights the urgent need for more advanced techniques in offensive language
detection to combat the evolving tactics used to evade detection mechanisms. | cs.CL | ToxiCloakCN: Evaluating Robustness of Offensive Language Detection in Chinese with Cloaking Perturbations |
2024-06-18T02:43:49Z | http://arxiv.org/pdf/2406.12221v1 | http://arxiv.org/abs/2406.12221v1 | Xueru Wen, Xinyu Lu, Xinyan Guan, Yaojie Lu, Hongyu Lin, Ben He, Xianpei Han, Le Sun | 2024-06-18T02:43:49Z | cs.CL | Hallucination occurs when large language models (LLMs) exhibit behavior that
deviates from the boundaries of their knowledge during the response generation
process. Previous learning-based methods focus on detecting knowledge
boundaries and finetuning models with instance-level feedback, but they suffer
from inaccurate signals due to off-policy data sampling and coarse-grained
feedback. In this paper, we introduce \textit{\b{R}einforcement \b{L}earning
\b{f}or \b{H}allucination} (RLFH), a fine-grained feedback-based online
reinforcement learning method for hallucination mitigation. Unlike previous
learning-based methods, RLFH enables LLMs to explore the boundaries of their
internal knowledge and provide on-policy, fine-grained feedback on these
explorations. To construct fine-grained feedback for learning reliable
generation behavior, RLFH decomposes the outcomes of large models into atomic
facts, provides statement-level evaluation signals, and traces back the signals
to the tokens of the original responses. Finally, RLFH adopts the online
reinforcement algorithm with these token-level rewards to adjust model behavior
for hallucination mitigation. For effective on-policy optimization, RLFH also
introduces an LLM-based fact assessment framework to verify the truthfulness
and helpfulness of atomic facts without human intervention. Experiments on
HotpotQA, SQuADv2, and Biography benchmarks demonstrate that RLFH can balance
their usage of internal knowledge during the generation process to eliminate
the hallucination behavior of LLMs. | cs.CL | On-Policy Fine-grained Knowledge Feedback for Hallucination Mitigation |
2024-06-18T02:32:57Z | http://arxiv.org/pdf/2406.12216v1 | http://arxiv.org/abs/2406.12216v1 | Yongyi Ji, Zhisheng Tang, Mayank Kejriwal | 2024-06-18T02:32:57Z | cs.CL, cs.AI | Personality, a fundamental aspect of human cognition, contains a range of
traits that influence behaviors, thoughts, and emotions. This paper explores
the capabilities of large language models (LLMs) in reconstructing these
complex cognitive attributes based only on simple descriptions containing
socio-demographic and personality type information. Utilizing the HEXACO
personality framework, our study examines the consistency of LLMs in recovering
and predicting underlying (latent) personality dimensions from simple
descriptions. Our experiments reveal a significant degree of consistency in
personality reconstruction, although some inconsistencies and biases, such as a
tendency to default to positive traits in the absence of explicit information,
are also observed. Additionally, socio-demographic factors like age and number
of children were found to influence the reconstructed personality dimensions.
These findings have implications for building sophisticated agent-based
simulacra using LLMs and highlight the need for further research on robust
personality generation in LLMs. | cs.CL | Is persona enough for personality? Using ChatGPT to reconstruct an agent's latent personality from simple descriptions |
2024-06-18T02:25:33Z | http://arxiv.org/pdf/2406.12213v3 | http://arxiv.org/abs/2406.12213v3 | Jie Wang | 2024-07-03T12:59:21Z | cs.CL, cs.AI, cs.FL, F.1.1; F.4.1; I.2.0 | Contemporary AI applications leverage large language models (LLMs) to harness
their knowledge and reasoning abilities for natural language processing tasks.
This approach shares similarities with the concept of oracle Turing machines
(OTMs). To capture the broader potential of these computations, including those
not yet realized, we propose an extension to OTMs: the LLM-oracle machine
(LLM-OM), by employing a cluster of LLMs as the oracle. Each LLM acts as a
black box, capable of answering queries within its expertise, albeit with a
delay. We introduce four variants of the LLM-OM: basic, augmented,
fault-avoidance, and $\epsilon$-fault. The first two are commonly observed in
existing AI applications. The latter two are specifically designed to address
the challenges of LLM hallucinations, biases, and inconsistencies, aiming to
ensure reliable outcomes. | cs.CL | LLM-Oracle Machines |
2024-06-18T02:13:48Z | http://arxiv.org/pdf/2406.12209v1 | http://arxiv.org/abs/2406.12209v1 | Yi-Jen Shih, David Harwath | 2024-06-18T02:13:48Z | cs.SD, cs.CL, eess.AS | Self-supervised speech (SSL) models have recently become widely adopted for
many downstream speech processing tasks. The general usage pattern is to employ
SSL models as feature extractors, and then train a downstream prediction head
to solve a specific task. However, different layers of SSL models have been
shown to capture different types of information, and the methods of combining
them are not well studied. To this end, we extend the general framework for SSL
model utilization by proposing the interface that connects the upstream and
downstream. Under this view, the dominant technique of combining features via a
layerwise weighted sum can be regarded as a specific interface. We propose
several alternative interface designs and demonstrate that the weighted sum
interface is suboptimal for many tasks. In particular, we show that a
convolutional interface whose depth scales logarithmically with the depth of
the upstream model consistently outperforms many other interface designs. | cs.SD | Interface Design for Self-Supervised Speech Models |
2024-06-18T02:12:34Z | http://arxiv.org/pdf/2406.12208v1 | http://arxiv.org/abs/2406.12208v1 | Guodong Du, Jing Li, Hanting Liu, Runhua Jiang, Shuyang Yu, Yifei Guo, Sim Kuan Goh, Ho-Kin Tang | 2024-06-18T02:12:34Z | cs.CL, cs.AI, cs.CV, cs.NE | Fine-tuning pre-trained language models, particularly large language models,
demands extensive computing resources and can result in varying performance
outcomes across different domains and datasets. This paper examines the
approach of integrating multiple models from diverse training scenarios into a
unified model. This unified model excels across various data domains and
exhibits the ability to generalize well on out-of-domain data. We propose a
knowledge fusion method named Evolver, inspired by evolutionary algorithms,
which does not need further training or additional training data. Specifically,
our method involves aggregating the weights of different language models into a
population and subsequently generating offspring models through mutation and
crossover operations. These offspring models are then evaluated against their
parents, allowing for the preservation of those models that show enhanced
performance on development datasets. Importantly, our model evolving strategy
can be seamlessly integrated with existing model merging frameworks, offering a
versatile tool for model enhancement. Experimental results on mainstream
language models (i.e., encoder-only, decoder-only, encoder-decoder) reveal that
Evolver outperforms previous state-of-the-art models by large margins. The code
is publicly available at {https://github.com/duguodong7/model-evolution}. | cs.CL | Knowledge Fusion By Evolving Weights of Language Models |
2024-06-18T01:53:49Z | http://arxiv.org/pdf/2406.12197v1 | http://arxiv.org/abs/2406.12197v1 | Sijia Wang, Lifu Huang | 2024-06-18T01:53:49Z | cs.CL, cs.AI | We propose a multi-agent debate as optimization (DAO) system for event
extraction, where the primary objective is to iteratively refine the large
language models (LLMs) outputs through debating without parameter tuning. In
DAO, we introduce two novel modules: the Diverse-RAG (DRAG) module and the
Adaptive Conformal Prediction (AdaCP) module. DRAG systematically retrieves
supporting information that best fits the debate discussion, while AdaCP
enhances the accuracy and reliability of event extraction by effectively
rejecting less promising answers. Experimental results demonstrate a
significant reduction in the performance gap between supervised approaches and
tuning-free LLM-based methods by 18.1% and 17.8% on ACE05 and 17.9% and 15.2%
on CASIE for event detection and argument extraction respectively. | cs.CL | Debate as Optimization: Adaptive Conformal Prediction and Diverse Retrieval for Event Extraction |
2024-06-18T01:30:07Z | http://arxiv.org/pdf/2406.12182v1 | http://arxiv.org/abs/2406.12182v1 | Lulu Zhao, Weihao Zeng, Xiaofeng Shi, Hua Zhou, Donglin Hao, Yonghua Lin | 2024-06-18T01:30:07Z | cs.CL, cs.AI | Recently, both closed-source LLMs and open-source communities have made
significant strides, outperforming humans in various general domains. However,
their performance in specific professional fields such as medicine, especially
within the open-source community, remains suboptimal due to the complexity of
medical knowledge. We propose Aquila-Med, a bilingual medical LLM based on
Aquila, addressing these challenges through continue pre-training, supervised
fine-tuning (SFT), and reinforcement learning from human feedback (RLHF). We
construct a large-scale Chinese and English medical dataset for continue
pre-training and a high-quality SFT dataset, covering extensive medical
specialties. Additionally, we develop a high-quality Direct Preference
Optimization (DPO) dataset for further alignment. Aquila-Med achieves notable
results across single-turn, multi-turn dialogues, and medical multiple-choice
questions, demonstrating the effectiveness of our approach. We open-source the
datasets and the entire training process, contributing valuable resources to
the research community. Our models and datasets will released at
https://huggingface.co/BAAI/AquilaMed-RL. | cs.CL | Aqulia-Med LLM: Pioneering Full-Process Open-Source Medical Language Models |
2024-06-18T00:41:40Z | http://arxiv.org/pdf/2406.12168v2 | http://arxiv.org/abs/2406.12168v2 | Wenda Xu, Jiachen Li, William Yang Wang, Lei Li | 2024-06-19T05:25:27Z | cs.LG, cs.AI, cs.CL | Direct alignment from preferences (DAP) has emerged as a promising paradigm
for aligning large language models (LLMs) to human desiderata from
pre-collected, offline preference datasets. While recent studies indicate that
existing offline DAP methods can directly benefit from online training samples,
we highlight the need to develop specific online DAP algorithms to fully
harness the power of online training. Specifically, we identify that the
learned LLM should adhere to the proximity of the behavior LLM, which collects
the training samples. To this end, we propose online Preference Optimization in
proximity to the Behavior LLM (BPO), emphasizing the importance of constructing
a proper trust region for LLM alignment.
We conduct extensive experiments to validate the effectiveness and
applicability of our approach by integrating it with various DAP methods,
resulting in significant performance improvements across a wide range of tasks
when training with the same amount of preference data. Even when only
introducing one additional data collection phase, our online BPO improves its
offline DAP baseline from 72.0% to 80.2% on TL;DR and from 82.2% to 89.1% on
Anthropic Helpfulness in terms of win rate against human reference text. | cs.LG | BPO: Supercharging Online Preference Learning by Adhering to the Proximity of Behavior LLM |
2024-06-18T00:35:02Z | http://arxiv.org/pdf/2406.12165v1 | http://arxiv.org/abs/2406.12165v1 | Andrea Vallebueno, Cassandra Handan-Nader, Christopher D. Manning, Daniel E. Ho | 2024-06-18T00:35:02Z | cs.CL | Static word embeddings are ubiquitous in computational social science
applications and contribute to practical decision-making in a variety of fields
including law and healthcare. However, assessing the statistical uncertainty in
downstream conclusions drawn from word embedding statistics has remained
challenging. When using only point estimates for embeddings, researchers have
no streamlined way of assessing the degree to which their model selection
criteria or scientific conclusions are subject to noise due to sparsity in the
underlying data used to generate the embeddings. We introduce a method to
obtain approximate, easy-to-use, and scalable reconstruction error variance
estimates for GloVe (Pennington et al., 2014), one of the most widely used word
embedding models, using an analytical approximation to a multivariate normal
model. To demonstrate the value of embeddings with variance (GloVe-V), we
illustrate how our approach enables principled hypothesis testing in core word
embedding tasks, such as comparing the similarity between different word pairs
in vector space, assessing the performance of different models, and analyzing
the relative degree of ethnic or gender bias in a corpus using different word
lists. | cs.CL | Statistical Uncertainty in Word Embeddings: GloVe-V |
2024-06-18T00:17:30Z | http://arxiv.org/pdf/2406.12159v1 | http://arxiv.org/abs/2406.12159v1 | Anna C. Marbut, John W. Chandler, Travis J. Wheeler | 2024-06-18T00:17:30Z | cs.CL, cs.LG | It is generally thought that transformer-based large language models benefit
from pre-training by learning generic linguistic knowledge that can be focused
on a specific task during fine-tuning. However, we propose that much of the
benefit from pre-training may be captured by geometric characteristics of the
latent space representations, divorced from any specific linguistic knowledge.
In this work we explore the relationship between GLUE benchmarking task
performance and a variety of measures applied to the latent space resulting
from BERT-type contextual language models. We find that there is a strong
linear relationship between a measure of quantized cell density and average
GLUE performance and that these measures may be predictive of otherwise
surprising GLUE performance for several non-standard BERT-type models from the
literature. These results may be suggestive of a strategy for decreasing
pre-training requirements, wherein model initialization can be informed by the
geometric characteristics of the model's latent space. | cs.CL | Exploring the Impact of a Transformer's Latent Space Geometry on Downstream Task Performance |
2024-06-18T00:14:07Z | http://arxiv.org/pdf/2406.12158v1 | http://arxiv.org/abs/2406.12158v1 | Nitish Joshi, Abulhair Saparov, Yixin Wang, He He | 2024-06-18T00:14:07Z | cs.CL, cs.AI | Recent work shows that causal facts can be effectively extracted from LLMs
through prompting, facilitating the creation of causal graphs for causal
inference tasks. However, it is unclear if this success is limited to
explicitly-mentioned causal facts in the pretraining data which the model can
memorize. Thus, this work investigates: Can LLMs infer causal relations from
other relational data in text? To disentangle the role of memorized causal
facts vs inferred causal relations, we finetune LLMs on synthetic data
containing temporal, spatial and counterfactual relations, and measure whether
the LLM can then infer causal relations. We find that: (a) LLMs are susceptible
to inferring causal relations from the order of two entity mentions in text
(e.g. X mentioned before Y implies X causes Y); (b) if the order is randomized,
LLMs still suffer from the post hoc fallacy, i.e. X occurs before Y (temporal
relation) implies X causes Y. We also find that while LLMs can correctly deduce
the absence of causal relations from temporal and spatial relations, they have
difficulty inferring causal relations from counterfactuals, questioning their
understanding of causality. | cs.CL | LLMs Are Prone to Fallacies in Causal Inference |
2024-06-18T00:02:15Z | http://arxiv.org/pdf/2406.15487v1 | http://arxiv.org/abs/2406.15487v1 | Zhifeng Kong, Sang-gil Lee, Deepanway Ghosal, Navonil Majumder, Ambuj Mehrish, Rafael Valle, Soujanya Poria, Bryan Catanzaro | 2024-06-18T00:02:15Z | cs.CL, cs.LG, cs.SD, eess.AS | It is an open challenge to obtain high quality training data, especially
captions, for text-to-audio models. Although prior methods have leveraged
\textit{text-only language models} to augment and improve captions, such
methods have limitations related to scale and coherence between audio and
captions. In this work, we propose an audio captioning pipeline that uses an
\textit{audio language model} to synthesize accurate and diverse captions for
audio at scale. We leverage this pipeline to produce a dataset of synthetic
captions for AudioSet, named \texttt{AF-AudioSet}, and then evaluate the
benefit of pre-training text-to-audio models on these synthetic captions.
Through systematic evaluations on AudioCaps and MusicCaps, we find leveraging
our pipeline and synthetic captions leads to significant improvements on audio
generation quality, achieving a new \textit{state-of-the-art}. | cs.CL | Improving Text-To-Audio Models with Synthetic Captions |
2024-06-17T23:07:53Z | http://arxiv.org/pdf/2406.12141v1 | http://arxiv.org/abs/2406.12141v1 | Gaëlle Laperrière, Sahar Ghannay, Bassam Jabaian, Yannick Estève | 2024-06-17T23:07:53Z | cs.CL, cs.SD, eess.AS | Self-Supervised Learning is vastly used to efficiently represent speech for
Spoken Language Understanding, gradually replacing conventional approaches.
Meanwhile, textual SSL models are proposed to encode language-agnostic
semantics. SAMU-XLSR framework employed this semantic information to enrich
multilingual speech representations. A recent study investigated SAMU-XLSR
in-domain semantic enrichment by specializing it on downstream transcriptions,
leading to state-of-the-art results on a challenging SLU task. This study's
interest lies in the loss of multilingual performances and lack of
specific-semantics training induced by such specialization in close languages
without any SLU implication. We also consider SAMU-XLSR's loss of initial
cross-lingual abilities due to a separate SLU fine-tuning. Therefore, this
paper proposes a dual task learning approach to improve SAMU-XLSR semantic
enrichment while considering distant languages for multilingual and language
portability experiments. | cs.CL | A dual task learning approach to fine-tune a multilingual semantic speech encoder for Spoken Language Understanding |
2024-06-17T22:58:36Z | http://arxiv.org/pdf/2406.12138v1 | http://arxiv.org/abs/2406.12138v1 | Vasyl Rakivnenko, Nestor Maslej, Jessica Cervi, Volodymyr Zhukov | 2024-06-17T22:58:36Z | cs.AI, cs.CL | Text embedding is becoming an increasingly popular AI methodology, especially
among businesses, yet the potential of text embedding models to be biased is
not well understood. This paper examines the degree to which a selection of
popular text embedding models are biased, particularly along gendered
dimensions. More specifically, this paper studies the degree to which these
models associate a list of given professions with gendered terms. The analysis
reveals that text embedding models are prone to gendered biases but in varying
ways. Although there are certain inter-model commonalities, for instance,
greater association of professions like nurse, homemaker, and socialite with
female identifiers, and greater association of professions like CEO, manager,
and boss with male identifiers, not all models make the same gendered
associations for each occupation. Furthermore, the magnitude and directionality
of bias can also vary on a model-by-model basis and depend on the particular
words models are prompted with. This paper demonstrates that gender bias
afflicts text embedding models and suggests that businesses using this
technology need to be mindful of the specific dimensions of this problem. | cs.AI | Bias in Text Embedding Models |
2024-06-17T22:42:14Z | http://arxiv.org/pdf/2406.12131v1 | http://arxiv.org/abs/2406.12131v1 | Peter Zeng, Eric Sclafani, Owen Rambow | 2024-06-17T22:42:14Z | cs.CL | We present Gram2Vec, a grammatical style embedding algorithm that embeds
documents into a higher dimensional space by extracting the normalized relative
frequencies of grammatical features present in the text. Compared to neural
approaches, Gram2Vec offers inherent interpretability based on how the feature
vectors are generated. In our demo, we present a way to visualize a mapping of
authors to documents based on their Gram2Vec vectors and highlight the ability
to drop or add features to view which authors make certain linguistic choices.
Next, we use authorship attribution as an application to show how Gram2Vec can
explain why a document is attributed to a certain author, using cosine
similarities between the Gram2Vec feature vectors to calculate the distances
between candidate documents and a query document. | cs.CL | Gram2Vec: An Interpretable Document Vectorizer |
2024-06-17T22:19:00Z | http://arxiv.org/pdf/2406.12128v1 | http://arxiv.org/abs/2406.12128v1 | Giovanni Puccetti, Anna Rogers, Chiara Alzetta, Felice Dell'Orletta, Andrea Esuli | 2024-06-17T22:19:00Z | cs.CL | Large Language Models (LLMs) are increasingly used as "content farm" models
(CFMs), to generate synthetic text that could pass for real news articles. This
is already happening even for languages that do not have high-quality
monolingual LLMs. We show that fine-tuning Llama (v1), mostly trained on
English, on as little as 40K Italian news articles, is sufficient for producing
news-like texts that native speakers of Italian struggle to identify as
synthetic.
We investigate three LLMs and three methods of detecting synthetic texts
(log-likelihood, DetectGPT, and supervised classification), finding that they
all perform better than human raters, but they are all impractical in the real
world (requiring either access to token likelihood information or a large
dataset of CFM texts). We also explore the possibility of creating a proxy CFM:
an LLM fine-tuned on a similar dataset to one used by the real "content farm".
We find that even a small amount of fine-tuning data suffices for creating a
successful detector, but we need to know which base LLM is used, which is a
major challenge.
Our results suggest that there are currently no practical methods for
detecting synthetic news-like texts 'in the wild', while generating them is too
easy. We highlight the urgency of more NLP research on this problem. | cs.CL | AI "News" Content Farms Are Easy to Make and Hard to Detect: A Case Study in Italian |
2024-06-17T22:13:22Z | http://arxiv.org/pdf/2406.12125v1 | http://arxiv.org/abs/2406.12125v1 | Dingyang Chen, Qi Zhang, Yinglun Zhu | 2024-06-17T22:13:22Z | cs.LG, cs.CL | This paper focuses on extending the success of large language models (LLMs)
to sequential decision making. Existing efforts either (i) re-train or finetune
LLMs for decision making, or (ii) design prompts for pretrained LLMs. The
former approach suffers from the computational burden of gradient updates, and
the latter approach does not show promising results. In this paper, we propose
a new approach that leverages online model selection algorithms to efficiently
incorporate LLMs agents into sequential decision making. Statistically, our
approach significantly outperforms both traditional decision making algorithms
and vanilla LLM agents. Computationally, our approach avoids the need for
expensive gradient updates of LLMs, and throughout the decision making process,
it requires only a small number of LLM calls. We conduct extensive experiments
to verify the effectiveness of our proposed approach. As an example, on a
large-scale Amazon dataset, our approach achieves more than a $6$x performance
gain over baselines while calling LLMs in only $1.5$\% of the time steps. | cs.LG | Efficient Sequential Decision Making with Large Language Models |
2024-06-17T21:56:57Z | http://arxiv.org/pdf/2406.12117v1 | http://arxiv.org/abs/2406.12117v1 | Layla Bouzoubaa, Elham Aghakhani, Max Song, Minh Trinh, Rezvaneh Rezapour | 2024-06-17T21:56:57Z | cs.CL | Online communities such as drug-related subreddits serve as safe spaces for
people who use drugs (PWUD), fostering discussions on substance use
experiences, harm reduction, and addiction recovery. Users' shared narratives
on these forums provide insights into the likelihood of developing a substance
use disorder (SUD) and recovery potential. Our study aims to develop a
multi-level, multi-label classification model to analyze online user-generated
texts about substance use experiences. For this purpose, we first introduce a
novel taxonomy to assess the nature of posts, including their intended
connections (Inquisition or Disclosure), subjects (e.g., Recovery, Dependency),
and specific objectives (e.g., Relapse, Quality, Safety). Using various
multi-label classification algorithms on a set of annotated data, we show that
GPT-4, when prompted with instructions, definitions, and examples, outperformed
all other models. We apply this model to label an additional 1,000 posts and
analyze the categories of linguistic expression used within posts in each
class. Our analysis shows that topics such as Safety, Combination of
Substances, and Mental Health see more disclosure, while discussions about
physiological Effects focus on harm reduction. Our work enriches the
understanding of PWUD's experiences and informs the broader knowledge base on
SUD and drug use. | cs.CL | Decoding the Narratives: Analyzing Personal Drug Experiences Shared on Reddit |
2024-06-17T21:45:48Z | http://arxiv.org/pdf/2406.12114v1 | http://arxiv.org/abs/2406.12114v1 | Hamidreza Rouzegar, Masoud Makrehchi | 2024-06-17T21:45:48Z | cs.CL, cs.AI, cs.LG, 68T50, I.2.7 | In the context of text classification, the financial burden of annotation
exercises for creating training data is a critical issue. Active learning
techniques, particularly those rooted in uncertainty sampling, offer a
cost-effective solution by pinpointing the most instructive samples for manual
annotation. Similarly, Large Language Models (LLMs) such as GPT-3.5 provide an
alternative for automated annotation but come with concerns regarding their
reliability. This study introduces a novel methodology that integrates human
annotators and LLMs within an Active Learning framework. We conducted
evaluations on three public datasets. IMDB for sentiment analysis, a Fake News
dataset for authenticity discernment, and a Movie Genres dataset for
multi-label classification.The proposed framework integrates human annotation
with the output of LLMs, depending on the model uncertainty levels. This
strategy achieves an optimal balance between cost efficiency and classification
performance. The empirical results show a substantial decrease in the costs
associated with data annotation while either maintaining or improving model
accuracy. | cs.CL | Enhancing Text Classification through LLM-Driven Active Learning and Human Annotation |
2024-06-17T21:37:09Z | http://arxiv.org/pdf/2406.12109v1 | http://arxiv.org/abs/2406.12109v1 | Almog Gueta, Amir Feder, Zorik Gekhman, Ariel Goldstein, Roi Reichart | 2024-06-17T21:37:09Z | cs.CL, cs.CE | This study empirically tests the $\textit{Narrative Economics}$ hypothesis,
which posits that narratives (ideas that are spread virally and affect public
beliefs) can influence economic fluctuations. We introduce two curated datasets
containing posts from X (formerly Twitter) which capture economy-related
narratives (Data will be shared upon paper acceptance). Employing Natural
Language Processing (NLP) methods, we extract and summarize narratives from the
tweets. We test their predictive power for $\textit{macroeconomic}$ forecasting
by incorporating the tweets' or the extracted narratives' representations in
downstream financial prediction tasks. Our work highlights the challenges in
improving macroeconomic models with narrative data, paving the way for the
research community to realistically address this important challenge. From a
scientific perspective, our investigation offers valuable insights and NLP
tools for narrative extraction and summarization using Large Language Models
(LLMs), contributing to future research on the role of narratives in economics. | cs.CL | Can LLMs Learn Macroeconomic Narratives from Social Media? |
2024-06-17T21:33:01Z | http://arxiv.org/pdf/2406.12104v1 | http://arxiv.org/abs/2406.12104v1 | Karime Maamari, Amine Mhedhbi | 2024-06-17T21:33:01Z | cs.CL, cs.AI, cs.DB, cs.LG | Recent advancements in Text-to-SQL have pushed database management systems
towards greater democratization of data access. Today's language models are at
the core of these advancements. They enable impressive Text-to-SQL generation
as experienced in the development of Distyl AI's Analytics Insight Engine. Its
early deployment with enterprise customers has highlighted three core
challenges. First, data analysts expect support with authoring SQL queries of
very high complexity. Second, requests are ad-hoc and, as such, require low
latency. Finally, generation requires an understanding of domain-specific
terminology and practices.
The design and implementation of our Text-to-SQL generation pipeline, powered
by large language models, tackles these challenges. The core tenants of our
approach rely on external knowledge that we extract in a pre-processing phase,
on retrieving the appropriate external knowledge at query generation time, and
on decomposing SQL query generation following a hierarchical CTE-based
structure. Finally, an adaptation framework leverages feedback to update the
external knowledge, in turn improving query generation over time. We give an
overview of our end-to-end approach and highlight the operators generating SQL
during inference. | cs.CL | End-to-end Text-to-SQL Generation within an Analytics Insight Engine |
2024-06-17T21:15:12Z | http://arxiv.org/pdf/2406.12094v1 | http://arxiv.org/abs/2406.12094v1 | Asma Ghandeharioun, Ann Yuan, Marius Guerard, Emily Reif, Michael A. Lepori, Lucas Dixon | 2024-06-17T21:15:12Z | cs.CL, cs.AI | Despite investments in improving model safety, studies show that misaligned
capabilities remain latent in safety-tuned models. In this work, we shed light
on the mechanics of this phenomenon. First, we show that even when model
generations are safe, harmful content can persist in hidden representations and
can be extracted by decoding from earlier layers. Then, we show that whether
the model divulges such content depends significantly on its perception of who
it is talking to, which we refer to as user persona. In fact, we find
manipulating user persona to be even more effective for eliciting harmful
content than direct attempts to control model refusal. We study both natural
language prompting and activation steering as control methods and show that
activation steering is significantly more effective at bypassing safety
filters. We investigate why certain personas break model safeguards and find
that they enable the model to form more charitable interpretations of otherwise
dangerous queries. Finally, we show we can predict a persona's effect on
refusal given only the geometry of its steering vector. | cs.CL | Who's asking? User personas and the mechanics of latent misalignment |
2024-06-17T21:06:00Z | http://arxiv.org/pdf/2406.12091v2 | http://arxiv.org/abs/2406.12091v2 | Pankayaraj Pathmanathan, Souradip Chakraborty, Xiangyu Liu, Yongyuan Liang, Furong Huang | 2024-06-19T17:56:17Z | cs.LG, cs.CL, cs.CR | Recent advancements in Reinforcement Learning with Human Feedback (RLHF) have
significantly impacted the alignment of Large Language Models (LLMs). The
sensitivity of reinforcement learning algorithms such as Proximal Policy
Optimization (PPO) has led to new line work on Direct Policy Optimization
(DPO), which treats RLHF in a supervised learning framework. The increased
practical use of these RLHF methods warrants an analysis of their
vulnerabilities. In this work, we investigate the vulnerabilities of DPO to
poisoning attacks under different scenarios and compare the effectiveness of
preference poisoning, a first of its kind. We comprehensively analyze DPO's
vulnerabilities under different types of attacks, i.e., backdoor and
non-backdoor attacks, and different poisoning methods across a wide array of
language models, i.e., LLama 7B, Mistral 7B, and Gemma 7B. We find that unlike
PPO-based methods, which, when it comes to backdoor attacks, require at least
4\% of the data to be poisoned to elicit harmful behavior, we exploit the true
vulnerabilities of DPO more simply so we can poison the model with only as much
as 0.5\% of the data. We further investigate the potential reasons behind the
vulnerability and how well this vulnerability translates into backdoor vs
non-backdoor attacks. | cs.LG | Is poisoning a real threat to LLM alignment? Maybe more so than you think |
2024-06-17T20:49:35Z | http://arxiv.org/pdf/2406.12084v1 | http://arxiv.org/abs/2406.12084v1 | Yebowen Hu, Kaiqiang Song, Sangwoo Cho, Xiaoyang Wang, Wenlin Yao, Hassan Foroosh, Dong Yu, Fei Liu | 2024-06-17T20:49:35Z | cs.CL, cs.AI | Reasoning is most powerful when an LLM accurately aggregates relevant
information. We examine the critical role of information aggregation in
reasoning by requiring the LLM to analyze sports narratives. To succeed at this
task, an LLM must infer points from actions, identify related entities,
attribute points accurately to players and teams, and compile key statistics to
draw conclusions. We conduct comprehensive experiments with real NBA basketball
data and present SportsGen, a new method to synthesize game narratives. By
synthesizing data, we can rigorously evaluate LLMs' reasoning capabilities
under complex scenarios with varying narrative lengths and density of
information. Our findings show that most models, including GPT-4o, often fail
to accurately aggregate basketball scores due to frequent scoring patterns.
Open-source models like Llama-3 further suffer from significant score
hallucinations. Finally, the effectiveness of reasoning is influenced by
narrative complexity, information density, and domain-specific terms,
highlighting the challenges in analytical reasoning tasks. | cs.CL | When Reasoning Meets Information Aggregation: A Case Study with Sports Narratives |
2024-06-17T20:20:47Z | http://arxiv.org/pdf/2406.12074v1 | http://arxiv.org/abs/2406.12074v1 | Zihao He, Rebecca Dorn, Siyi Guo, Minh Duc Chu, Kristina Lerman | 2024-06-17T20:20:47Z | cs.CL | Social scientists use surveys to probe the opinions and beliefs of
populations, but these methods are slow, costly, and prone to biases. Recent
advances in large language models (LLMs) enable creating computational
representations or "digital twins" of populations that generate human-like
responses mimicking the population's language, styles, and attitudes. We
introduce Community-Cross-Instruct, an unsupervised framework for aligning LLMs
to online communities to elicit their beliefs. Given a corpus of a community's
online discussions, Community-Cross-Instruct automatically generates
instruction-output pairs by an advanced LLM to (1) finetune an foundational LLM
to faithfully represent that community, and (2) evaluate the alignment of the
finetuned model to the community. We demonstrate the method's utility in
accurately representing political and fitness communities on Reddit. Unlike
prior methods requiring human-authored instructions, Community-Cross-Instruct
generates instructions in a fully unsupervised manner, enhancing scalability
and generalization across domains. This work enables cost-effective and
automated surveying of diverse online communities. | cs.CL | COMMUNITY-CROSS-INSTRUCT: Unsupervised Instruction Generation for Aligning Large Language Models to Online Communities |
2024-06-17T20:14:16Z | http://arxiv.org/pdf/2406.12069v1 | http://arxiv.org/abs/2406.12069v1 | Marko Sterbentz, Cameron Barrie, Shubham Shahi, Abhratanu Dutta, Donna Hooshmand, Harper Pack, Kristian J. Hammond | 2024-06-17T20:14:16Z | cs.CL | Large language models (LLMs) are capable of producing documents, and
retrieval augmented generation (RAG) has shown itself to be a powerful method
for improving accuracy without sacrificing fluency. However, not all
information can be retrieved from text. We propose an approach that uses the
analysis of structured data to generate fact sets that are used to guide
generation in much the same way that retrieved documents are used in RAG. This
analytics augmented generation (AAG) approach supports the ability to utilize
standard analytic techniques to generate facts that are then converted to text
and passed to an LLM. We present a neurosymbolic platform, Satyrn that
leverages AAG to produce accurate, fluent, and coherent reports grounded in
large scale databases. In our experiments, we find that Satyrn generates
reports in which over 86% accurate claims while maintaining high levels of
fluency and coherence, even when using smaller language models such as
Mistral-7B, as compared to GPT-4 Code Interpreter in which just 57% of claims
are accurate. | cs.CL | Satyrn: A Platform for Analytics Augmented Generation |
2024-06-17T20:09:24Z | http://arxiv.org/pdf/2406.12066v2 | http://arxiv.org/abs/2406.12066v2 | Jack Gallifant, Shan Chen, Pedro Moreira, Nikolaj Munch, Mingye Gao, Jackson Pond, Leo Anthony Celi, Hugo Aerts, Thomas Hartvigsen, Danielle Bitterman | 2024-06-19T03:59:41Z | cs.CL | Medical knowledge is context-dependent and requires consistent reasoning
across various natural language expressions of semantically equivalent phrases.
This is particularly crucial for drug names, where patients often use brand
names like Advil or Tylenol instead of their generic equivalents. To study
this, we create a new robustness dataset, RABBITS, to evaluate performance
differences on medical benchmarks after swapping brand and generic drug names
using physician expert annotations.
We assess both open-source and API-based LLMs on MedQA and MedMCQA, revealing
a consistent performance drop ranging from 1-10\%. Furthermore, we identify a
potential source of this fragility as the contamination of test data in widely
used pre-training datasets. All code is accessible at
https://github.com/BittermanLab/RABBITS, and a HuggingFace leaderboard is
available at https://huggingface.co/spaces/AIM-Harvard/rabbits-leaderboard. | cs.CL | Language Models are Surprisingly Fragile to Drug Names in Biomedical Benchmarks |
2024-06-17T20:00:04Z | http://arxiv.org/pdf/2406.12060v1 | http://arxiv.org/abs/2406.12060v1 | Ukyo Honda, Tatsushi Oka, Peinan Zhang, Masato Mita | 2024-06-17T20:00:04Z | cs.CL, cs.LG | Recent models for natural language understanding are inclined to exploit
simple patterns in datasets, commonly known as shortcuts. These shortcuts hinge
on spurious correlations between labels and latent features existing in the
training data. At inference time, shortcut-dependent models are likely to
generate erroneous predictions under distribution shifts, particularly when
some latent features are no longer correlated with the labels. To avoid this,
previous studies have trained models to eliminate the reliance on shortcuts. In
this study, we explore a different direction: pessimistically aggregating the
predictions of a mixture-of-experts, assuming each expert captures relatively
different latent features. The experimental results demonstrate that our
post-hoc control over the experts significantly enhances the model's robustness
to the distribution shift in shortcuts. Besides, we show that our approach has
some practical advantages. We also analyze our model and provide results to
support the assumption. | cs.CL | Not Eliminate but Aggregate: Post-Hoc Control over Mixture-of-Experts to Address Shortcut Shifts in Natural Language Understanding |
2024-06-17T19:50:40Z | http://arxiv.org/pdf/2406.12058v3 | http://arxiv.org/abs/2406.12058v3 | Seyedali Mohammadi, Edward Raff, Jinendra Malekar, Vedant Palit, Francis Ferraro, Manas Gaur | 2024-06-28T04:08:12Z | cs.AI, cs.CL | Language Models (LMs) are being proposed for mental health applications where
the heightened risk of adverse outcomes means predictive performance may not be
a sufficient litmus test of a model's utility in clinical practice. A model
that can be trusted for practice should have a correspondence between
explanation and clinical determination, yet no prior research has examined the
attention fidelity of these models and their effect on ground truth
explanations. We introduce an evaluation design that focuses on the robustness
and explainability of LMs in identifying Wellness Dimensions (WD). We focus on
two mental health and well-being datasets: (a) Multi-label Classification-based
MultiWD, and (b) WellXplain for evaluating attention mechanism veracity against
expert-labeled explanations. The labels are based on Halbert Dunn's theory of
wellness, which gives grounding to our evaluation. We reveal four surprising
results about LMs/LLMs: (1) Despite their human-like capabilities, GPT-3.5/4
lag behind RoBERTa, and MedAlpaca, a fine-tuned LLM fails to deliver any
remarkable improvements in performance or explanations. (2) Re-examining LMs'
predictions based on a confidence-oriented loss function reveals a significant
performance drop. (3) Across all LMs/LLMs, the alignment between attention and
explanations remains low, with LLMs scoring a dismal 0.0. (4) Most mental
health-specific LMs/LLMs overlook domain-specific knowledge and undervalue
explanations, causing these discrepancies. This study highlights the need for
further research into their consistency and explanations in mental health and
well-being. | cs.AI | WellDunn: On the Robustness and Explainability of Language Models and Large Language Models in Identifying Wellness Dimensions |
2024-06-17T19:46:05Z | http://arxiv.org/pdf/2406.12053v1 | http://arxiv.org/abs/2406.12053v1 | Mohammad Beigi, Ying Shen, Runing Yang, Zihao Lin, Qifan Wang, Ankith Mohan, Jianfeng He, Ming Jin, Chang-Tien Lu, Lifu Huang | 2024-06-17T19:46:05Z | cs.CL | Despite their vast capabilities, Large Language Models (LLMs) often struggle
with generating reliable outputs, frequently producing high-confidence
inaccuracies known as hallucinations. Addressing this challenge, our research
introduces InternalInspector, a novel framework designed to enhance confidence
estimation in LLMs by leveraging contrastive learning on internal states
including attention states, feed-forward states, and activation states of all
layers. Unlike existing methods that primarily focus on the final activation
state, InternalInspector conducts a comprehensive analysis across all internal
states of every layer to accurately identify both correct and incorrect
prediction processes. By benchmarking InternalInspector against existing
confidence estimation methods across various natural language understanding and
generation tasks, including factual question answering, commonsense reasoning,
and reading comprehension, InternalInspector achieves significantly higher
accuracy in aligning the estimated confidence scores with the correctness of
the LLM's predictions and lower calibration error. Furthermore,
InternalInspector excels at HaluEval, a hallucination detection benchmark,
outperforming other internal-based confidence estimation methods in this task. | cs.CL | InternalInspector $I^2$: Robust Confidence Estimation in LLMs through Internal States |
2024-06-17T19:45:21Z | http://arxiv.org/pdf/2406.12052v1 | http://arxiv.org/abs/2406.12052v1 | Yi Fang, Dongzhe Fan, Sirui Ding, Ninghao Liu, Qiaoyu Tan | 2024-06-17T19:45:21Z | cs.CL, cs.AI, cs.IR, cs.LG | Representation learning on text-attributed graphs (TAGs), where nodes are
represented by textual descriptions, is crucial for textual and relational
knowledge systems and recommendation systems. Currently, state-of-the-art
embedding methods for TAGs primarily focus on fine-tuning language models
(e.g., BERT) using structure-aware training signals. While effective, these
methods are tailored for individual TAG and cannot generalize across various
graph scenarios. Given the shared textual space, leveraging multiple TAGs for
joint fine-tuning, aligning text and graph structure from different aspects,
would be more beneficial. Motivated by this, we introduce a novel Unified Graph
Language Model (UniGLM) framework, the first graph embedding model that
generalizes well to both in-domain and cross-domain TAGs. Specifically, UniGLM
is trained over multiple TAGs with different domains and scales using
self-supervised contrastive learning. UniGLM includes an adaptive positive
sample selection technique for identifying structurally similar nodes and a
lazy contrastive module that is devised to accelerate training by minimizing
repetitive encoding calculations. Extensive empirical results across 9
benchmark TAGs demonstrate UniGLM's efficacy against leading embedding
baselines in terms of generalization (various downstream tasks and backbones)
and transfer learning (in and out of domain scenarios). The code is available
at https://github.com/NYUSHCS/UniGLM. | cs.CL | UniGLM: Training One Unified Language Model for Text-Attributed Graphs |
2024-06-17T19:42:22Z | http://arxiv.org/pdf/2406.12050v1 | http://arxiv.org/abs/2406.12050v1 | Zhihan Zhang, Zhenwen Liang, Wenhao Yu, Dian Yu, Mengzhao Jia, Dong Yu, Meng Jiang | 2024-06-17T19:42:22Z | cs.CL | Supervised fine-tuning enhances the problem-solving abilities of language
models across various mathematical reasoning tasks. To maximize such benefits,
existing research focuses on broadening the training set with various data
augmentation techniques, which is effective for standard single-round
question-answering settings. Our work introduces a novel technique aimed at
cultivating a deeper understanding of the training problems at hand, enhancing
performance not only in standard settings but also in more complex scenarios
that require reflective thinking. Specifically, we propose reflective
augmentation, a method that embeds problem reflection into each training
instance. It trains the model to consider alternative perspectives and engage
with abstractions and analogies, thereby fostering a thorough comprehension
through reflective reasoning. Extensive experiments validate the achievement of
our aim, underscoring the unique advantages of our method and its complementary
nature relative to existing augmentation techniques. | cs.CL | Learn Beyond The Answer: Training Language Models with Reflection for Mathematical Reasoning |
2024-06-17T19:33:08Z | http://arxiv.org/pdf/2406.12045v1 | http://arxiv.org/abs/2406.12045v1 | Shunyu Yao, Noah Shinn, Pedram Razavi, Karthik Narasimhan | 2024-06-17T19:33:08Z | cs.AI, cs.CL | Existing benchmarks do not test language agents on their interaction with
human users or ability to follow domain-specific rules, both of which are vital
for deploying them in real world applications. We propose $\tau$-bench, a
benchmark emulating dynamic conversations between a user (simulated by language
models) and a language agent provided with domain-specific API tools and policy
guidelines. We employ an efficient and faithful evaluation process that
compares the database state at the end of a conversation with the annotated
goal state. We also propose a new metric (pass^k) to evaluate the reliability
of agent behavior over multiple trials. Our experiments show that even
state-of-the-art function calling agents (like gpt-4o) succeed on <50% of the
tasks, and are quite inconsistent (pass^8 <25% in retail). Our findings point
to the need for methods that can improve the ability of agents to act
consistently and follow rules reliably. | cs.AI | $τ$-bench: A Benchmark for Tool-Agent-User Interaction in Real-World Domains |
2024-06-17T19:11:40Z | http://arxiv.org/pdf/2406.12038v1 | http://arxiv.org/abs/2406.12038v1 | Karuna Bhaila, Minh-Hao Van, Xintao Wu | 2024-06-17T19:11:40Z | cs.CL, cs.AI, cs.LG | The widespread popularity of Large Language Models (LLMs), partly due to
their unique ability to perform in-context learning, has also brought to light
the importance of ethical and safety considerations when deploying these
pre-trained models. In this work, we focus on investigating machine unlearning
for LLMs motivated by data protection regulations. In contrast to the growing
literature on fine-tuning methods to achieve unlearning, we focus on a
comparatively lightweight alternative called soft prompting to realize the
unlearning of a subset of training data. With losses designed to enforce
forgetting as well as utility preservation, our framework \textbf{S}oft
\textbf{P}rompting for \textbf{U}n\textbf{l}earning (SPUL) learns prompt tokens
that can be appended to an arbitrary query to induce unlearning of specific
examples at inference time without updating LLM parameters. We conduct a
rigorous evaluation of the proposed method and our results indicate that SPUL
can significantly improve the trade-off between utility and forgetting in the
context of text classification with LLMs. We further validate our method using
multiple LLMs to highlight the scalability of our framework and provide
detailed insights into the choice of hyperparameters and the influence of the
size of unlearning data. Our implementation is available at
\url{https://github.com/karuna-bhaila/llm_unlearning}. | cs.CL | Soft Prompting for Unlearning in Large Language Models |
2024-06-17T19:07:21Z | http://arxiv.org/pdf/2406.12036v4 | http://arxiv.org/abs/2406.12036v4 | Nikhil Khandekar, Qiao Jin, Guangzhi Xiong, Soren Dunn, Serina S Applebaum, Zain Anwar, Maame Sarfo-Gyamfi, Conrad W Safranek, Abid A Anwar, Andrew Zhang, Aidan Gilson, Maxwell B Singer, Amisha Dave, Andrew Taylor, Aidong Zhang, Qingyu Chen, Zhiyong Lu | 2024-06-30T15:12:10Z | cs.CL, cs.AI | As opposed to evaluating computation and logic-based reasoning, current
benchmarks for evaluating large language models (LLMs) in medicine are
primarily focused on question-answering involving domain knowledge and
descriptive reasoning. While such qualitative capabilities are vital to medical
diagnosis, in real-world scenarios, doctors frequently use clinical calculators
that follow quantitative equations and rule-based reasoning paradigms for
evidence-based decision support. To this end, we propose MedCalc-Bench, a
first-of-its-kind dataset focused on evaluating the medical calculation
capability of LLMs. MedCalc-Bench contains an evaluation set of over 1000
manually reviewed instances from 55 different medical calculation tasks. Each
instance in MedCalc-Bench consists of a patient note, a question requesting to
compute a specific medical value, a ground truth answer, and a step-by-step
explanation showing how the answer is obtained. While our evaluation results
show the potential of LLMs in this area, none of them are effective enough for
clinical settings. Common issues include extracting the incorrect entities, not
using the correct equation or rules for a calculation task, or incorrectly
performing the arithmetic for the computation. We hope our study highlights the
quantitative knowledge and reasoning gaps in LLMs within medical settings,
encouraging future improvements of LLMs for various clinical calculation tasks. | cs.CL | MedCalc-Bench: Evaluating Large Language Models for Medical Calculations |
2024-06-17T19:06:54Z | http://arxiv.org/pdf/2406.12034v1 | http://arxiv.org/abs/2406.12034v1 | Junmo Kang, Leonid Karlinsky, Hongyin Luo, Zhen Wang, Jacob Hansen, James Glass, David Cox, Rameswar Panda, Rogerio Feris, Alan Ritter | 2024-06-17T19:06:54Z | cs.CL, cs.LG | We present Self-MoE, an approach that transforms a monolithic LLM into a
compositional, modular system of self-specialized experts, named MiXSE (MiXture
of Self-specialized Experts). Our approach leverages self-specialization, which
constructs expert modules using self-generated synthetic data, each equipped
with a shared base LLM and incorporating self-optimized routing. This allows
for dynamic and capability-specific handling of various target tasks, enhancing
overall capabilities, without extensive human-labeled data and added
parameters. Our empirical results reveal that specializing LLMs may exhibit
potential trade-offs in performances on non-specialized tasks. On the other
hand, our Self-MoE demonstrates substantial improvements over the base LLM
across diverse benchmarks such as knowledge, reasoning, math, and coding. It
also consistently outperforms other methods, including instance merging and
weight merging, while offering better flexibility and interpretability by
design with semantic experts and routing. Our findings highlight the critical
role of modularity and the potential of self-improvement in achieving
efficient, scalable, and adaptable systems. | cs.CL | Self-MoE: Towards Compositional Large Language Models with Self-Specialized Experts |
2024-06-17T19:05:32Z | http://arxiv.org/pdf/2406.12033v2 | http://arxiv.org/abs/2406.12033v2 | Yuqing Wang, Yun Zhao, Sara Alessandra Keller, Anne de Hond, Marieke M. van Buchem, Malvika Pillai, Tina Hernandez-Boussard | 2024-06-19T18:28:22Z | cs.CL | The advancement of large language models (LLMs) has demonstrated strong
capabilities across various applications, including mental health analysis.
However, existing studies have focused on predictive performance, leaving the
critical issue of fairness underexplored, posing significant risks to
vulnerable populations. Despite acknowledging potential biases, previous works
have lacked thorough investigations into these biases and their impacts. To
address this gap, we systematically evaluate biases across seven social factors
(e.g., gender, age, religion) using ten LLMs with different prompting methods
on eight diverse mental health datasets. Our results show that GPT-4 achieves
the best overall balance in performance and fairness among LLMs, although it
still lags behind domain-specific models like MentalRoBERTa in some cases.
Additionally, our tailored fairness-aware prompts can effectively mitigate bias
in mental health predictions, highlighting the great potential for fair
analysis in this field. | cs.CL | Unveiling and Mitigating Bias in Mental Health Analysis with Large Language Models |
2024-06-17T18:58:20Z | http://arxiv.org/pdf/2406.12031v1 | http://arxiv.org/abs/2406.12031v1 | Josh Gardner, Juan C. Perdomo, Ludwig Schmidt | 2024-06-17T18:58:20Z | cs.LG, cs.AI, cs.CL | Tabular data -- structured, heterogeneous, spreadsheet-style data with rows
and columns -- is widely used in practice across many domains. However, while
recent foundation models have reduced the need for developing task-specific
datasets and predictors in domains such as language modeling and computer
vision, this transfer learning paradigm has not had similar impact in the
tabular domain. In this work, we seek to narrow this gap and present TabuLa-8B,
a language model for tabular prediction. We define a process for extracting a
large, high-quality training dataset from the TabLib corpus, proposing methods
for tabular data filtering and quality control. Using the resulting dataset,
which comprises over 1.6B rows from 3.1M unique tables, we fine-tune a Llama
3-8B large language model (LLM) for tabular data prediction (classification and
binned regression) using a novel packing and attention scheme for tabular
prediction. Through evaluation across a test suite of 329 datasets, we find
that TabuLa-8B has zero-shot accuracy on unseen tables that is over 15
percentage points (pp) higher than random guessing, a feat that is not possible
with existing state-of-the-art tabular prediction models (e.g. XGBoost,
TabPFN). In the few-shot setting (1-32 shots), without any fine-tuning on the
target datasets, TabuLa-8B is 5-15 pp more accurate than XGBoost and TabPFN
models that are explicitly trained on equal, or even up to 16x more data. We
release our model, code, and data along with the publication of this paper. | cs.LG | Large Scale Transfer Learning for Tabular Data via Language Modeling |
Subsets and Splits
No saved queries yet
Save your SQL queries to embed, download, and access them later. Queries will appear here once saved.