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
|
|---|---|---|---|---|---|---|---|---|
2024-06-20T08:12:49Z | http://arxiv.org/pdf/2406.14091v1 | http://arxiv.org/abs/2406.14091v1 | Dohyun Lee, Daniel Rim, Minseok Choi, Jaegul Choo | 2024-06-20T08:12:49Z | cs.CL | Although language models (LMs) demonstrate exceptional capabilities on
various tasks, they are potentially vulnerable to extraction attacks, which
represent a significant privacy risk. To mitigate the privacy concerns of LMs,
machine unlearning has emerged as an important research area, which is utilized
to induce the LM to selectively forget about some of its training data. While
completely retraining the model will guarantee successful unlearning and
privacy assurance, it is impractical for LMs, as it would be time-consuming and
resource-intensive. Prior works efficiently unlearn the target token sequences,
but upon subsequent iterations, the LM displays significant degradation in
performance. In this work, we propose Privacy Protection via Optimal Parameters
(POP), a novel unlearning method that effectively forgets the target token
sequences from the pretrained LM by applying optimal gradient updates to the
parameters. Inspired by the gradient derivation of complete retraining, we
approximate the optimal training objective that successfully unlearns the
target sequence while retaining the knowledge from the rest of the training
data. Experimental results demonstrate that POP exhibits remarkable retention
performance post-unlearning across 9 classification and 4 dialogue benchmarks,
outperforming the state-of-the-art by a large margin. Furthermore, we introduce
Remnant Memorization Accuracy that quantifies privacy risks based on token
likelihood and validate its effectiveness through both qualitative and
quantitative analyses. | cs.CL | Protecting Privacy Through Approximating Optimal Parameters for Sequence Unlearning in Language Models |
2024-06-20T08:04:07Z | http://arxiv.org/pdf/2406.14088v1 | http://arxiv.org/abs/2406.14088v1 | Zhiyu Mei, Wei Fu, Kaiwei Li, Guangju Wang, Huanchen Zhang, Yi Wu | 2024-06-20T08:04:07Z | cs.DC, cs.AI, cs.CL, cs.LG | Reinforcement Learning from Human Feedback (RLHF) stands as a pivotal
technique in empowering large language model (LLM) applications. Since RLHF
involves diverse computational workloads and intricate dependencies among
multiple LLMs, directly adopting parallelization techniques from supervised
training can result in sub-optimal performance. To overcome this limitation, we
propose a novel approach named parameter ReaLlocation, which dynamically
redistributes LLM parameters in the cluster and adapts parallelization
strategies during training. Building upon this idea, we introduce ReaLHF, a
pioneering system capable of automatically discovering and running efficient
execution plans for RLHF training given the desired algorithmic and hardware
configurations. ReaLHF formulates the execution plan for RLHF as an augmented
dataflow graph. Based on this formulation, ReaLHF employs a tailored search
algorithm with a lightweight cost estimator to discover an efficient execution
plan. Subsequently, the runtime engine deploys the selected plan by effectively
parallelizing computations and redistributing parameters. We evaluate ReaLHF on
the LLaMA-2 models with up to $4\times70$ billion parameters and 128 GPUs. The
experiment results showcase ReaLHF's substantial speedups of $2.0-10.6\times$
compared to baselines. Furthermore, the execution plans generated by ReaLHF
exhibit an average of $26\%$ performance improvement over heuristic approaches
based on Megatron-LM. The source code of ReaLHF is publicly available at
https://github.com/openpsi-project/ReaLHF . | cs.DC | ReaLHF: Optimized RLHF Training for Large Language Models through Parameter Reallocation |
2024-06-20T07:50:37Z | http://arxiv.org/pdf/2406.14075v1 | http://arxiv.org/abs/2406.14075v1 | Yi-Fan Lu, Xian-Ling Mao, Bo Wang, Xiao Liu, Heyan Huang | 2024-06-20T07:50:37Z | cs.CL | It is crucial to utilize events to understand a specific domain. There are
lots of research on event extraction in many domains such as news, finance and
biology domain. However, scientific domain still lacks event extraction
research, including comprehensive datasets and corresponding methods. Compared
to other domains, scientific domain presents two characteristics: denser
nuggets and more complex events. To solve the above problem, considering these
two characteristics, we first construct SciEvents, a large-scale multi-event
document-level dataset with a schema tailored for scientific domain. It has
2,508 documents and 24,381 events under refined annotation and quality control.
Then, we propose EXCEEDS, a novel end-to-end scientific event extraction
framework by storing dense nuggets in a grid matrix and simplifying complex
event extraction into a dot construction and connection task. Experimental
results demonstrate state-of-the-art performances of EXCEEDS on SciEvents.
Additionally, we release SciEvents and EXCEEDS on GitHub. | cs.CL | EXCEEDS: Extracting Complex Events as Connecting the Dots to Graphs in Scientific Domain |
2024-06-20T07:17:09Z | http://arxiv.org/pdf/2406.14051v1 | http://arxiv.org/abs/2406.14051v1 | Nidhir Bhavsar, Jonathan Jordan, Sherzod Hakimov, David Schlangen | 2024-06-20T07:17:09Z | cs.CL, cs.AI | What makes a good Large Language Model (LLM)? That it performs well on the
relevant benchmarks -- which hopefully measure, with some validity, the
presence of capabilities that are also challenged in real application. But what
makes the model perform well? What gives a model its abilities? We take a
recently introduced type of benchmark that is meant to challenge capabilities
in a goal-directed, agentive context through self-play of conversational games,
and analyse how performance develops as a function of model characteristics
like number of parameters, or type of training. We find that while there is a
clear relationship between number of parameters and performance, there is still
a wide spread of performance points within a given size bracket, which is to be
accounted for by training parameters such as fine-tuning data quality and
method. From a more practical angle, we also find a certain degree of
unpredictability about performance across access methods, possible due to
unexposed sampling parameters, and a, very welcome, performance stability
against at least moderate weight quantisation during inference. | cs.CL | How Many Parameters Does it Take to Change a Light Bulb? Evaluating Performance in Self-Play of Conversational Games as a Function of Model Characteristics |
2024-06-20T07:13:25Z | http://arxiv.org/pdf/2406.14048v1 | http://arxiv.org/abs/2406.14048v1 | Daniil Khomsky, Narek Maloyan, Bulat Nutfullin | 2024-06-20T07:13:25Z | cs.CL | Large language models play a crucial role in modern natural language
processing technologies. However, their extensive use also introduces potential
security risks, such as the possibility of black-box attacks. These attacks can
embed hidden malicious features into the model, leading to adverse consequences
during its deployment.
This paper investigates methods for black-box attacks on large language
models with a three-tiered defense mechanism. It analyzes the challenges and
significance of these attacks, highlighting their potential implications for
language processing system security. Existing attack and defense methods are
examined, evaluating their effectiveness and applicability across various
scenarios.
Special attention is given to the detection algorithm for black-box attacks,
identifying hazardous vulnerabilities in language models and retrieving
sensitive information. This research presents a methodology for vulnerability
detection and the development of defensive strategies against black-box attacks
on large language models. | cs.CL | Prompt Injection Attacks in Defended Systems |
2024-06-20T07:06:58Z | http://arxiv.org/pdf/2406.14043v1 | http://arxiv.org/abs/2406.14043v1 | Yueqing Liang, Liangwei Yang, Chen Wang, Xiongxiao Xu, Philip S. Yu, Kai Shu | 2024-06-20T07:06:58Z | cs.IR, cs.CL | With the emergence of large language models (LLMs) and their ability to
perform a variety of tasks, their application in recommender systems (RecSys)
has shown promise. However, we are facing significant challenges when deploying
LLMs into RecSys, such as limited prompt length, unstructured item information,
and un-constrained generation of recommendations, leading to sub-optimal
performance. To address these issues, we propose a novel method using a
taxonomy dictionary. This method provides a systematic framework for
categorizing and organizing items, improving the clarity and structure of item
information. By incorporating the taxonomy dictionary into LLM prompts, we
achieve efficient token utilization and controlled feature generation, leading
to more accurate and contextually relevant recommendations. Our Taxonomy-guided
Recommendation (TaxRec) approach features a two-step process: one-time taxonomy
categorization and LLM-based recommendation, enabling zero-shot recommendations
without the need for domain-specific fine-tuning. Experimental results
demonstrate TaxRec significantly enhances recommendation quality compared to
traditional zero-shot approaches, showcasing its efficacy as personal
recommender with LLMs. Code is available at
https://github.com/yueqingliang1/TaxRec. | cs.IR | Taxonomy-Guided Zero-Shot Recommendations with LLMs |
2024-06-20T06:59:46Z | http://arxiv.org/pdf/2406.14039v1 | http://arxiv.org/abs/2406.14039v1 | Ying Zhang, Matthieu Petit Guillaume, Aurélien Krauth, Manel Labidi | 2024-06-20T06:59:46Z | cs.AI, cs.CE, cs.CL, cs.NE | CryptoGPT: a 7B model competing with GPT-4 in a specific task -- The Impact
of Automatic Annotation and Strategic Fine-Tuning via QLoRAIn this article, we
present a method aimed at refining a dedicated LLM of reasonable quality with
limited resources in an industrial setting via CryptoGPT. It is an LLM designed
for financial news analysis for the cryptocurrency market in real-time. This
project was launched in an industrial context. This model allows not only for
the classification of financial information but also for providing
comprehensive analysis. We refined different LLMs of the same size such as
Mistral-7B and LLama-7B using semi-automatic annotation and compared them with
various LLMs such as GPT-3.5 and GPT-4. Our goal is to find a balance among
several needs: 1. Protecting data (by avoiding their transfer to external
servers), 2. Limiting annotation cost and time, 3. Controlling the model's size
(to manage deployment costs), and 4. Maintaining better analysis quality. | cs.AI | CryptoGPT: a 7B model rivaling GPT-4 in the task of analyzing and classifying real-time financial news |
2024-06-20T06:56:35Z | http://arxiv.org/pdf/2406.14036v1 | http://arxiv.org/abs/2406.14036v1 | Jiuxiang Gu, Yingyu Liang, Zhenmei Shi, Zhao Song, Chiwun Yang | 2024-06-20T06:56:35Z | cs.LG, cs.AI, cs.CL | Prompting and contextual-based fine-tuning methods, which we call Prefix
Learning, have been proposed to enhance the performance of language models on
various downstream tasks that can match full parameter fine-tuning. There
remains a limited theoretical understanding of how these methods work. In this
paper, we aim to relieve this limitation by studying the learning ability of
Prefix Learning from the perspective of prefix length. In particular, we
approximate the infinite-long Prefix Learning optimization process by the
Neural Tangent Kernel (NTK) technique. We formulate and solve it as a learning
problem of the infinite-long prefix in a one-layer attention network. Our
results confirm the over-parameterization property and arbitrary small loss
convergence guarantee of the infinite-long Prefix Learning in attention. To the
implementation end, we propose our NTK-Attention method, which is "equivalent"
to attention computation with arbitrary prefix length efficiently. Its time
complexity mainly depends on the sub-quadratic of input length (without
prefix), and our method only requires $d^2 + d$ extra parameters for
representation, where $d$ is the feature dimension. In addition, we conducted
experiments that compare our NTK-Attention with full parameters fine-tuning,
LoRA, and P-Tuning V2 methods across vision or natural language datasets. The
results indicate our approach may be a promising
parameter-efficient-fine-tuning method since it has demonstrated superior
performance in numerous scenarios. Our code can be found at
\url{https://github.com/ChristianYang37/chiwun/tree/main/src/NTK-Attention}. | cs.LG | Toward Infinite-Long Prefix in Transformer |
2024-06-20T06:56:19Z | http://arxiv.org/pdf/2406.14035v1 | http://arxiv.org/abs/2406.14035v1 | Sherzod Hakimov, Yerkezhan Abdullayeva, Kushal Koshti, Antonia Schmidt, Yan Weiser, Anne Beyer, David Schlangen | 2024-06-20T06:56:19Z | cs.CL, cs.AI | While the situation has improved for text-only models, it again seems to be
the case currently that multimodal (text and image) models develop faster than
ways to evaluate them. In this paper, we bring a recently developed evaluation
paradigm from text models to multimodal models, namely evaluation through the
goal-oriented game (self) play, complementing reference-based and
preference-based evaluation. Specifically, we define games that challenge a
model's capability to represent a situation from visual information and align
such representations through dialogue. We find that the largest closed models
perform rather well on the games that we define, while even the best
open-weight models struggle with them. On further analysis, we find that the
exceptional deep captioning capabilities of the largest models drive some of
the performance. There is still room to grow for both kinds of models, ensuring
the continued relevance of the benchmark. | cs.CL | Two Giraffes in a Dirt Field: Using Game Play to Investigate Situation Modelling in Large Multimodal Models |
2024-06-20T06:46:23Z | http://arxiv.org/pdf/2406.14026v1 | http://arxiv.org/abs/2406.14026v1 | Xisen Jin, Xiang Ren | 2024-06-20T06:46:23Z | cs.LG, cs.CL, stat.ML | Language models (LMs) are known to suffer from forgetting of previously
learned examples when fine-tuned, breaking stability of deployed LM systems.
Despite efforts on mitigating forgetting, few have investigated whether, and
how forgotten upstream examples are associated with newly learned tasks.
Insights on such associations enable efficient and targeted mitigation of
forgetting. In this paper, we empirically analyze forgetting that occurs in $N$
upstream examples while the model learns $M$ new tasks and visualize their
associations with a $M \times N$ matrix. We empirically demonstrate that the
degree of forgetting can often be approximated by simple multiplicative
contributions of the upstream examples and newly learned tasks. We also reveal
more complicated patterns where specific subsets of examples are forgotten with
statistics and visualization. Following our analysis, we predict forgetting
that happens on upstream examples when learning a new task with matrix
completion over the empirical associations, outperforming prior approaches that
rely on trainable LMs. Project website:
https://inklab.usc.edu/lm-forgetting-prediction/ | cs.LG | Demystifying Forgetting in Language Model Fine-Tuning with Statistical Analysis of Example Associations |
2024-06-20T06:42:27Z | http://arxiv.org/pdf/2406.14024v2 | http://arxiv.org/abs/2406.14024v2 | Bofei Gao, Zefan Cai, Runxin Xu, Peiyi Wang, Ce Zheng, Runji Lin, Keming Lu, Junyang Lin, Chang Zhou, Wen Xiao, Junjie Hu, Tianyu Liu, Baobao Chang | 2024-06-30T13:44:04Z | cs.CL | Mathematical verfier achieves success in mathematical reasoning tasks by
validating the correctness of solutions. However, existing verifiers are
trained with binary classification labels, which are not informative enough for
the model to accurately assess the solutions. To mitigate the aforementioned
insufficiency of binary labels, we introduce step-wise natural language
feedbacks as rationale labels (i.e., the correctness of the current step and
the explanations). In this paper, we propose \textbf{Math-Minos}, a natural
language feedback enhanced verifier by constructing automatically-generated
training data and a two-stage training paradigm for effective training and
efficient inference. Our experiments reveal that a small set (30k) of natural
language feedbacks can significantly boost the performance of the verifier by
the accuracy of 1.6\% (86.6\% $\rightarrow$ 88.2\%) on GSM8K and 0.8\% (37.8\%
$\rightarrow$ 38.6\%) on MATH. We have released our code and data for further
exploration. | cs.CL | LLM Critics Help Catch Bugs in Mathematics: Towards a Better Mathematical Verifier with Natural Language Feedback |
2024-06-20T06:42:08Z | http://arxiv.org/pdf/2406.14023v1 | http://arxiv.org/abs/2406.14023v1 | Yuchen Wen, Keping Bi, Wei Chen, Jiafeng Guo, Xueqi Cheng | 2024-06-20T06:42:08Z | cs.CL, cs.AI | As Large Language Models (LLMs) become an important way of information
seeking, there have been increasing concerns about the unethical content LLMs
may generate. In this paper, we conduct a rigorous evaluation of LLMs' implicit
bias towards certain groups by attacking them with carefully crafted
instructions to elicit biased responses. Our attack methodology is inspired by
psychometric principles in cognitive and social psychology. We propose three
attack approaches, i.e., Disguise, Deception, and Teaching, based on which we
built evaluation datasets for four common bias types. Each prompt attack has
bilingual versions. Extensive evaluation of representative LLMs shows that 1)
all three attack methods work effectively, especially the Deception attacks; 2)
GLM-3 performs the best in defending our attacks, compared to GPT-3.5 and
GPT-4; 3) LLMs could output content of other bias types when being taught with
one type of bias. Our methodology provides a rigorous and effective way of
evaluating LLMs' implicit bias and will benefit the assessments of LLMs'
potential ethical risks. | cs.CL | Evaluating Implicit Bias in Large Language Models by Attacking From a Psychometric Perspective |
2024-06-20T06:37:47Z | http://arxiv.org/pdf/2406.14022v1 | http://arxiv.org/abs/2406.14022v1 | Xiaolei Wang, Xinyu Tang, Wayne Xin Zhao, Ji-Rong Wen | 2024-06-20T06:37:47Z | cs.LG, cs.CL | The emergence of in-context learning (ICL) is potentially attributed to two
major abilities: task recognition (TR) for recognizing the task from
demonstrations and utilizing pre-trained priors, and task learning (TL) for
learning from demonstrations. However, relationships between the two abilities
and how such relationships affect the emergence of ICL is unclear. In this
paper, we take the first step by examining the pre-training dynamics of the
emergence of ICL. With carefully designed metrics, we find that these two
abilities are, in fact, competitive during pre-training. Moreover, we observe a
strong negative correlation between the competition and ICL performance.
Further analysis of common pre-training factors (i.e., model size, dataset
size, and data curriculum) demonstrates possible ways to manage the
competition. Based on these insights, we propose a simple yet effective method
to better integrate these two abilities for ICL at inference time. Through
adaptive ensemble learning, the performance of ICL can be significantly
boosted, enabling two small models to outperform a larger one with more than
twice the parameters. The code is available at
https://github.com/RUCAIBox/Competitive-ICL. | cs.LG | Investigating the Pre-Training Dynamics of In-Context Learning: Task Recognition vs. Task Learning |
2024-06-20T06:37:35Z | http://arxiv.org/pdf/2406.14021v1 | http://arxiv.org/abs/2406.14021v1 | Yongqiang Chen, Quanming Yao, Juzheng Zhang, James Cheng, Yatao Bian | 2024-06-20T06:37:35Z | cs.CL, cs.LG, q-bio.QM | Recently there has been a surge of interest in extending the success of large
language models (LLMs) to graph modality, such as social networks and
molecules. As LLMs are predominantly trained with 1D text data, most existing
approaches adopt a graph neural network to represent a graph as a series of
node tokens and feed these tokens to LLMs for graph-language alignment. Despite
achieving some successes, existing approaches have overlooked the hierarchical
structures that are inherent in graph data. Especially, in molecular graphs,
the high-order structural information contains rich semantics of molecular
functional groups, which encode crucial biochemical functionalities of the
molecules. We establish a simple benchmark showing that neglecting the
hierarchical information in graph tokenization will lead to subpar
graph-language alignment and severe hallucination in generated outputs. To
address this problem, we propose a novel strategy called HIerarchical GrapH
Tokenization (HIGHT). HIGHT employs a hierarchical graph tokenizer that
extracts and encodes the hierarchy of node, motif, and graph levels of
informative tokens to improve the graph perception of LLMs. HIGHT also adopts
an augmented graph-language supervised fine-tuning dataset, enriched with the
hierarchical graph information, to further enhance the graph-language
alignment. Extensive experiments on 7 molecule-centric benchmarks confirm the
effectiveness of HIGHT in reducing hallucination by 40%, as well as significant
improvements in various molecule-language downstream tasks. | cs.CL | HIGHT: Hierarchical Graph Tokenization for Graph-Language Alignment |
2024-06-20T06:02:04Z | http://arxiv.org/pdf/2406.14012v1 | http://arxiv.org/abs/2406.14012v1 | Navid Ayoobi, Sadat Shahriar, Arjun Mukherjee | 2024-06-20T06:02:04Z | cs.CL, cs.AI | LLMs offer valuable capabilities, yet they can be utilized by malicious users
to disseminate deceptive information and generate fake news. The growing
prevalence of LLMs poses difficulties in crafting detection approaches that
remain effective across various text domains. Additionally, the absence of
precautionary measures for AI-generated news on online social platforms is
concerning. Therefore, there is an urgent need to improve people's ability to
differentiate between news articles written by humans and those produced by
LLMs. By providing cues in human-written and LLM-generated news, we can help
individuals increase their skepticism towards fake LLM-generated news. This
paper aims to elucidate simple markers that help individuals distinguish
between articles penned by humans and those created by LLMs. To achieve this,
we initially collected a dataset comprising 39k news articles authored by
humans or generated by four distinct LLMs with varying degrees of fake. We then
devise a metric named Entropy-Shift Authorship Signature (ESAS) based on the
information theory and entropy principles. The proposed ESAS ranks terms or
entities, like POS tagging, within news articles based on their relevance in
discerning article authorship. We demonstrate the effectiveness of our metric
by showing the high accuracy attained by a basic method, i.e., TF-IDF combined
with logistic regression classifier, using a small set of terms with the
highest ESAS score. Consequently, we introduce and scrutinize these top
ESAS-ranked terms to aid individuals in strengthening their skepticism towards
LLM-generated fake news. | cs.CL | Seeing Through AI's Lens: Enhancing Human Skepticism Towards LLM-Generated Fake News |
2024-06-20T05:18:37Z | http://arxiv.org/pdf/2406.14005v2 | http://arxiv.org/abs/2406.14005v2 | Mandar Sharma, Nikhil Muralidhar, Shengzhe Xu, Raquib Bin Yousuf, Naren Ramakrishnan | 2024-06-21T12:41:17Z | cs.CL, cs.AI, cs.LG | The pretraining-fine-tuning paradigm has been the de facto strategy for
transfer learning in modern language modeling. With the understanding that task
adaptation in LMs is often a function of parameters shared across tasks, we
argue that a more surgical approach to regularization needs to exist for
smoother transfer learning. Towards this end, we investigate how the
pretraining loss landscape is affected by these task-sensitive parameters
through an information-theoretic lens. We then leverage the findings from our
investigations to devise a novel approach to dropout for improved model
regularization and better downstream generalization. This approach, named
guided dropout, is both task & architecture agnostic and adds no computational
overhead to the fine-tuning process. Through empirical evaluations, we showcase
that our approach to regularization yields consistently better performance,
even in scenarios of data paucity, compared to standardized baselines. | cs.CL | Information Guided Regularization for Fine-tuning Language Models |
2024-06-20T04:52:19Z | http://arxiv.org/pdf/2406.13997v1 | http://arxiv.org/abs/2406.13997v1 | Mahammed Kamruzzaman, Hieu Minh Nguyen, Gene Louis Kim | 2024-06-20T04:52:19Z | cs.CL, cs.CE | Many recent studies have investigated social biases in LLMs but brand bias
has received little attention. This research examines the biases exhibited by
LLMs towards different brands, a significant concern given the widespread use
of LLMs in affected use cases such as product recommendation and market
analysis. Biased models may perpetuate societal inequalities, unfairly favoring
established global brands while marginalizing local ones. Using a curated
dataset across four brand categories, we probe the behavior of LLMs in this
space. We find a consistent pattern of bias in this space -- both in terms of
disproportionately associating global brands with positive attributes and
disproportionately recommending luxury gifts for individuals in high-income
countries. We also find LLMs are subject to country-of-origin effects which may
boost local brand preference in LLM outputs in specific contexts. | cs.CL | "Global is Good, Local is Bad?": Understanding Brand Bias in LLMs |
2024-06-20T04:44:20Z | http://arxiv.org/pdf/2406.13993v1 | http://arxiv.org/abs/2406.13993v1 | Mahammed Kamruzzaman, Gene Louis Kim | 2024-06-20T04:44:20Z | cs.CL, cs.LG | Persona assignment has become a common strategy for customizing LLM use to
particular tasks and contexts. In this study, we explore how perceptions of
different nations change when LLMs are assigned specific nationality personas.
We assign 193 different nationality personas (e.g., an American person) to four
LLMs and examine how the LLM perceptions of countries change. We find that all
LLM-persona combinations tend to favor Western European nations, though
nation-personas push LLM behaviors to focus more on and view more favorably the
nation-persona's own region. Eastern European, Latin American, and African
nations are viewed more negatively by different nationality personas. Our study
provides insight into how biases and stereotypes are realized within LLMs when
adopting different national personas. In line with the "Blueprint for an AI
Bill of Rights", our findings underscore the critical need for developing
mechanisms to ensure LLMs uphold fairness and not over-generalize at a global
scale. | cs.CL | Exploring Changes in Nation Perception with Nationality-Assigned Personas in LLMs |
2024-06-20T04:35:59Z | http://arxiv.org/pdf/2406.13990v2 | http://arxiv.org/abs/2406.13990v2 | Qin Zhu, Qingyuan Cheng, Runyu Peng, Xiaonan Li, Tengxiao Liu, Ru Peng, Xipeng Qiu, Xuanjing Huang | 2024-06-23T16:46:00Z | cs.CL | The training process of large language models (LLMs) often involves varying
degrees of test data contamination. Although current LLMs are achieving
increasingly better performance on various benchmarks, their performance in
practical applications does not always match their benchmark results. Leakage
of benchmarks can prevent the accurate assessment of LLMs' true performance.
However, constructing new benchmarks is costly, labor-intensive and still
carries the risk of leakage. Therefore, in this paper, we ask the question, Can
we reuse these leaked benchmarks for LLM evaluation? We propose Inference-Time
Decontamination (ITD) to address this issue by detecting and rewriting leaked
samples without altering their difficulties. ITD can mitigate performance
inflation caused by memorizing leaked benchmarks. Our proof-of-concept
experiments demonstrate that ITD reduces inflated accuracy by 22.9% on GSM8K
and 19.0% on MMLU. On MMLU, using Inference-time Decontamination can lead to a
decrease in the results of Phi3 and Mistral by 6.7% and 3.6% respectively. We
hope that ITD can provide more truthful evaluation results for large language
models. | cs.CL | Inference-Time Decontamination: Reusing Leaked Benchmarks for Large Language Model Evaluation |
2024-06-20T03:50:23Z | http://arxiv.org/pdf/2406.13975v1 | http://arxiv.org/abs/2406.13975v1 | Zhongshen Zeng, Yinhong Liu, Yingjia Wan, Jingyao Li, Pengguang Chen, Jianbo Dai, Yuxuan Yao, Rongwu Xu, Zehan Qi, Wanru Zhao, Linling Shen, Jianqiao Lu, Haochen Tan, Yukang Chen, Hao Zhang, Zhan Shi, Bailin Wang, Zhijiang Guo, Jiaya Jia | 2024-06-20T03:50:23Z | cs.CL, cs.AI | Large language models (LLMs) have shown increasing capability in
problem-solving and decision-making, largely based on the step-by-step
chain-of-thought reasoning processes. However, it has been increasingly
challenging to evaluate the reasoning capability of LLMs. Concretely, existing
outcome-based benchmarks begin to saturate and become less sufficient to
monitor the progress. To this end, we present a process-based benchmark MR-BEN
that demands a meta reasoning skill, where LMs are asked to locate and analyse
potential errors in automatically generated reasoning steps. MR-BEN is a
comprehensive benchmark comprising 5,975 questions collected from human
experts, covering various subjects such as physics, chemistry, logic, coding,
and more. Through our designed metrics for assessing meta-reasoning on this
benchmark, we identify interesting limitations and weaknesses of current LLMs
(open-source and closed-source models). For example, open-source models are
seemingly comparable to GPT-4 on outcome-based benchmarks, but they lag far
behind on our benchmark, revealing the underlying reasoning capability gap
between them. Our dataset and codes are available on
https://randolph-zeng.github.io/Mr-Ben.github.io/. | cs.CL | MR-BEN: A Comprehensive Meta-Reasoning Benchmark for Large Language Models |
2024-06-20T03:02:38Z | http://arxiv.org/pdf/2406.13960v1 | http://arxiv.org/abs/2406.13960v1 | Yi Cheng, Wenge Liu, Kaishuai Xu, Wenjun Hou, Yi Ouyang, Chak Tou Leong, Xian Wu, Yefeng Zheng | 2024-06-20T03:02:38Z | cs.CL, cs.AI | Previous research on persona-based dialogue agents typically preset the
agent's persona before deployment, which remains static thereafter. In this
paper, we take a step further and explore a new paradigm called Self-evolving
Personalized Dialogue Agents (SPDA), where the agent continuously evolves
during the conversation to better align with the user's anticipation by
dynamically adapting its persona. This paradigm could enable better
personalization for each user, but also introduce unique challenges, which
mainly lie in the process of persona adaptation. Two key issues include how to
achieve persona alignment with the user and how to ensure smooth transition in
the adaptation process. To address them, we propose a novel framework that
refines the persona at hierarchical levels to progressively align better with
the user in a controllable way. Experiments show that integrating the personas
adapted by our framework consistently enhances personalization and overall
dialogue performance across various base systems. | cs.CL | Evolving to be Your Soulmate: Personalized Dialogue Agents with Dynamically Adapted Personas |
2024-06-20T02:32:16Z | http://arxiv.org/pdf/2406.13948v1 | http://arxiv.org/abs/2406.13948v1 | Jie Feng, Yuwei Du, Tianhui Liu, Siqi Guo, Yuming Lin, Yong Li | 2024-06-20T02:32:16Z | cs.AI, cs.CL, cs.LG | Large language models(LLMs) with powerful language generation and reasoning
capabilities have already achieved success in many domains, e.g., math and code
generation. However, due to the lacking of physical world's corpus and
knowledge during training, they usually fail to solve many real-life tasks in
the urban space. In this paper, we propose CityGPT, a systematic framework for
enhancing the capability of LLMs on understanding urban space and solving the
related urban tasks by building a city-scale world model in the model. First,
we construct a diverse instruction tuning dataset CityInstruction for injecting
urban knowledge and enhancing spatial reasoning capability effectively. By
using a mixture of CityInstruction and general instruction data, we fine-tune
various LLMs (e.g., ChatGLM3-6B, Qwen1.5 and LLama3 series) to enhance their
capability without sacrificing general abilities. To further validate the
effectiveness of proposed methods, we construct a comprehensive benchmark
CityEval to evaluate the capability of LLMs on diverse urban scenarios and
problems. Extensive evaluation results demonstrate that small LLMs trained with
CityInstruction can achieve competitive performance with commercial LLMs in the
comprehensive evaluation of CityEval. The source codes are openly accessible to
the research community via https://github.com/tsinghua-fib-lab/CityGPT. | cs.AI | CityGPT: Empowering Urban Spatial Cognition of Large Language Models |
2024-06-20T02:29:46Z | http://arxiv.org/pdf/2406.13947v1 | http://arxiv.org/abs/2406.13947v1 | Ya-Lun Li | 2024-06-20T02:29:46Z | cs.AI, cs.CL | Due to the rapid advancement of Large Language Model (LLM), the whole
community eagerly consumes any available text data in order to train the LLM.
Currently, large portion of the available text data are collected from
internet, which has been thought as a cheap source of the training data.
However, when people try to extend the LLM's capability to the personal related
domain, such as healthcare or education, the lack of public dataset in these
domains make the adaption of the LLM in such domains much slower. The reason of
lacking public available dataset in such domains is because they usually
contain personal sensitive information. In order to comply with privacy law,
the data in such domains need to be de-identified before any kind of
dissemination. It had been much research tried to address this problem for the
image or tabular data. However, there was limited research on the efficient and
general de-identification method for text data. Most of the method based on
human annotation or predefined category list. It usually can not be easily
adapted to specific domains. The goal of this proposal is to develop a text
de-identification framework, which can be easily adapted to the specific
domain, leverage the existing expert knowledge without further human
annotation. We propose an aspect-based utility-preserved de-identification
summarization framework, AspirinSum, by learning to align expert's aspect from
existing comment data, it can efficiently summarize the personal sensitive
document by extracting personal sensitive aspect related sub-sentence and
de-identify it by substituting it with similar aspect sub-sentence. We envision
that the de-identified text can then be used in data publishing, eventually
publishing our de-identified dataset for downstream task use. | cs.AI | AspirinSum: an Aspect-based utility-preserved de-identification Summarization framework |
2024-06-20T02:25:07Z | http://arxiv.org/pdf/2406.13945v1 | http://arxiv.org/abs/2406.13945v1 | Jie Feng, Jun Zhang, Junbo Yan, Xin Zhang, Tianjian Ouyang, Tianhui Liu, Yuwei Du, Siqi Guo, Yong Li | 2024-06-20T02:25:07Z | cs.AI, cs.CL, cs.LG | Large language models (LLMs) with powerful generalization ability has been
widely used in many domains. A systematic and reliable evaluation of LLMs is a
crucial step in their development and applications, especially for specific
professional fields. In the urban domain, there have been some early
explorations about the usability of LLMs, but a systematic and scalable
evaluation benchmark is still lacking. The challenge in constructing a
systematic evaluation benchmark for the urban domain lies in the diversity of
data and scenarios, as well as the complex and dynamic nature of cities. In
this paper, we propose CityBench, an interactive simulator based evaluation
platform, as the first systematic evaluation benchmark for the capability of
LLMs for urban domain. First, we build CitySim to integrate the multi-source
data and simulate fine-grained urban dynamics. Based on CitySim, we design 7
tasks in 2 categories of perception-understanding and decision-making group to
evaluate the capability of LLMs as city-scale world model for urban domain. Due
to the flexibility and ease-of-use of CitySim, our evaluation platform
CityBench can be easily extended to any city in the world. We evaluate 13
well-known LLMs including open source LLMs and commercial LLMs in 13 cities
around the world. Extensive experiments demonstrate the scalability and
effectiveness of proposed CityBench and shed lights for the future development
of LLMs in urban domain. The dataset, benchmark and source codes are openly
accessible to the research community via
https://github.com/tsinghua-fib-lab/CityBench | cs.AI | CityBench: Evaluating the Capabilities of Large Language Model as World Model |
2024-06-20T02:19:33Z | http://arxiv.org/pdf/2406.13940v1 | http://arxiv.org/abs/2406.13940v1 | Yongheng Zhang, Qiguang Chen, Min Li, Wanxiang Che, Libo Qin | 2024-06-20T02:19:33Z | cs.CL | Cross-lingual chain-of-thought can effectively complete reasoning tasks
across languages, which gains increasing attention. Recently, dominant
approaches in the literature improve cross-lingual alignment capabilities by
integrating reasoning knowledge from different languages. Despite achieving
excellent performance, current methods still have two main challenges: (1)
Manual language specification: They still highly rely on manually selecting the
languages to integrate, severely affecting their generalizability; (2) Static
weight allocation: Current methods simply integrate all languages equally. In
fact, different language reasoning paths should have different weights to
achieve better complementation and integration. Motivated by this, we introduce
an Automatic Cross-lingual Alignment Planning (AutoCAP) for zero-shot
chain-of-thought to address the above challenges. The core of AutoCAP consists
of two components: (1) Automatic Language Selection Prompting to guide LLMs to
select appropriate languages and (2) Automatic Weight Allocation Prompting to
automatically allocate alignment weight scores to each reasoning path.
Extensive experiments on several benchmarks reveal that AutoCAP achieves
state-of-the-art performance, surpassing previous methods that required manual
effort. | cs.CL | AutoCAP: Towards Automatic Cross-lingual Alignment Planning for Zero-shot Chain-of-Thought |
2024-06-20T02:02:53Z | http://arxiv.org/pdf/2406.13934v1 | http://arxiv.org/abs/2406.13934v1 | Kaishuai Xu, Yi Cheng, Wenjun Hou, Qiaoyu Tan, Wenjie Li | 2024-06-20T02:02:53Z | cs.CL, cs.AI | Medical dialogue systems have attracted significant attention for their
potential to act as medical assistants. Enabling these medical systems to
emulate clinicians' diagnostic reasoning process has been the long-standing
research focus. Previous studies rudimentarily realized the simulation of
clinicians' diagnostic process by fine-tuning language models on high-quality
dialogue datasets. Nonetheless, they overly focus on the outcomes of the
clinician's reasoning process while ignoring their internal thought processes
and alignment with clinician preferences. Our work aims to build a medical
dialogue system that aligns with clinicians' diagnostic reasoning processes. We
propose a novel framework, Emulation, designed to generate an appropriate
response that relies on abductive and deductive diagnostic reasoning analyses
and aligns with clinician preferences through thought process modeling.
Experimental results on two datasets confirm the efficacy of Emulation.
Crucially, our framework furnishes clear explanations for the generated
responses, enhancing its transparency in medical consultations. | cs.CL | Reasoning Like a Doctor: Improving Medical Dialogue Systems via Diagnostic Reasoning Process Alignment |
2024-06-20T01:53:25Z | http://arxiv.org/pdf/2406.13929v1 | http://arxiv.org/abs/2406.13929v1 | Jongyoon Song, Sangwon Yu, Sungroh Yoon | 2024-06-20T01:53:25Z | cs.CL, cs.AI, cs.LG | In this paper, we identify a new category of bias that induces
input-conflicting hallucinations, where large language models (LLMs) generate
responses inconsistent with the content of the input context. This issue we
have termed the false negative problem refers to the phenomenon where LLMs are
predisposed to return negative judgments when assessing the correctness of a
statement given the context. In experiments involving pairs of statements that
contain the same information but have contradictory factual directions, we
observe that LLMs exhibit a bias toward false negatives. Specifically, the
model presents greater overconfidence when responding with False. Furthermore,
we analyze the relationship between the false negative problem and context and
query rewriting and observe that both effectively tackle false negatives in
LLMs. | cs.CL | Large Language Models are Skeptics: False Negative Problem of Input-conflicting Hallucination |
2024-06-20T01:45:44Z | http://arxiv.org/pdf/2406.13925v1 | http://arxiv.org/abs/2406.13925v1 | Tao Zhang, Ziqian Zeng, Yuxiang Xiao, Huiping Zhuang, Cen Chen, James Foulds, Shimei Pan | 2024-06-20T01:45:44Z | cs.CL, cs.AI | Large Language Models (LLMs) are prone to generating content that exhibits
gender biases, raising significant ethical concerns. Alignment, the process of
fine-tuning LLMs to better align with desired behaviors, is recognized as an
effective approach to mitigate gender biases. Although proprietary LLMs have
made significant strides in mitigating gender bias, their alignment datasets
are not publicly available. The commonly used and publicly available alignment
dataset, HH-RLHF, still exhibits gender bias to some extent. There is a lack of
publicly available alignment datasets specifically designed to address gender
bias. Hence, we developed a new dataset named GenderAlign, aiming at mitigating
a comprehensive set of gender biases in LLMs. This dataset comprises 8k
single-turn dialogues, each paired with a "chosen" and a "rejected" response.
Compared to the "rejected" responses, the "chosen" responses demonstrate lower
levels of gender bias and higher quality. Furthermore, we categorized the
gender biases in the "rejected" responses of GenderAlign into 4 principal
categories. The experimental results show the effectiveness of GenderAlign in
reducing gender bias in LLMs. | cs.CL | GenderAlign: An Alignment Dataset for Mitigating Gender Bias in Large Language Models |
2024-06-20T01:43:08Z | http://arxiv.org/pdf/2406.13923v1 | http://arxiv.org/abs/2406.13923v1 | Junjie Wang, Yin Zhang, Yatai Ji, Yuxiang Zhang, Chunyang Jiang, Yubo Wang, Kang Zhu, Zekun Wang, Tiezhen Wang, Wenhao Huang, Jie Fu, Bei Chen, Qunshu Lin, Minghao Liu, Ge Zhang, Wenhu Chen | 2024-06-20T01:43:08Z | cs.AI, cs.CL, cs.CV, cs.MM | Recent advancements in Large Multimodal Models (LMMs) have leveraged
extensive multimodal datasets to enhance capabilities in complex
knowledge-driven tasks. However, persistent challenges in perceptual and
reasoning errors limit their efficacy, particularly in interpreting intricate
visual data and deducing multimodal relationships. Addressing these issues, we
introduce a novel dataset format, PIN (Paired and INterleaved multimodal
documents), designed to significantly improve both the depth and breadth of
multimodal training. The PIN format is built on three foundational principles:
knowledge intensity, scalability, and support for diverse training modalities.
This innovative format combines markdown files and comprehensive images to
enrich training data with a dense knowledge structure and versatile training
strategies. We present PIN-14M, an open-source dataset comprising 14 million
samples derived from a diverse range of Chinese and English sources, tailored
to include complex web and scientific content. This dataset is constructed
meticulously to ensure data quality and ethical integrity, aiming to facilitate
advanced training strategies and improve model robustness against common
multimodal training pitfalls. Our initial results, forming the basis of this
technical report, suggest significant potential for the PIN format in refining
LMM performance, with plans for future expansions and detailed evaluations of
its impact on model capabilities. | cs.AI | PIN: A Knowledge-Intensive Dataset for Paired and Interleaved Multimodal Documents |
2024-06-20T01:24:30Z | http://arxiv.org/pdf/2407.02511v1 | http://arxiv.org/abs/2407.02511v1 | Silin Meng, Yiwei Wang, Cheng-Fu Yang, Nanyun Peng, Kai-Wei Chang | 2024-06-20T01:24:30Z | cs.RO, cs.AI, cs.CL | Path planning is a fundamental scientific problem in robotics and autonomous
navigation, requiring the derivation of efficient routes from starting to
destination points while avoiding obstacles. Traditional algorithms like A* and
its variants are capable of ensuring path validity but suffer from significant
computational and memory inefficiencies as the state space grows. Conversely,
large language models (LLMs) excel in broader environmental analysis through
contextual understanding, providing global insights into environments. However,
they fall short in detailed spatial and temporal reasoning, often leading to
invalid or inefficient routes. In this work, we propose LLM-A*, an new LLM
based route planning method that synergistically combines the precise
pathfinding capabilities of A* with the global reasoning capability of LLMs.
This hybrid approach aims to enhance pathfinding efficiency in terms of time
and space complexity while maintaining the integrity of path validity,
especially in large-scale scenarios. By integrating the strengths of both
methodologies, LLM-A* addresses the computational and memory limitations of
conventional algorithms without compromising on the validity required for
effective pathfinding. | cs.RO | LLM-A*: Large Language Model Enhanced Incremental Heuristic Search on Path Planning |
2024-06-20T00:28:33Z | http://arxiv.org/pdf/2406.13905v1 | http://arxiv.org/abs/2406.13905v1 | Mohamed Elaraby, Diane Litman, Xiang Lorraine Li, Ahmed Magooda | 2024-06-20T00:28:33Z | cs.CL | Generating free-text rationales is among the emergent capabilities of Large
Language Models (LLMs). These rationales have been found to enhance LLM
performance across various NLP tasks. Recently, there has been growing interest
in using these rationales to provide insights for various important downstream
tasks. In this paper, we analyze generated free-text rationales in tasks with
subjective answers, emphasizing the importance of rationalization in such
scenarios. We focus on pairwise argument ranking, a highly subjective task with
significant potential for real-world applications, such as debate assistance.
We evaluate the persuasiveness of rationales generated by nine LLMs to support
their subjective choices. Our findings suggest that open-source LLMs,
particularly Llama2-70B-chat, are capable of providing highly persuasive
rationalizations, surpassing even GPT models. Additionally, our experiments
show that rationale persuasiveness can be improved by controlling its
parameters through prompting or through self-refinement. | cs.CL | Persuasiveness of Generated Free-Text Rationales in Subjective Decisions: A Case Study on Pairwise Argument Ranking |
2024-06-20T00:25:43Z | http://arxiv.org/pdf/2406.13903v1 | http://arxiv.org/abs/2406.13903v1 | Hamdireza Rouzegar, Masoud Makrehchi | 2024-06-20T00:25:43Z | cs.CL, cs.AI, cs.LG, 68T50, I.2.7 | This study investigates how LLMs, specifically GPT-3.5 and GPT-4, can develop
tailored questions for Grade 9 math, aligning with active learning principles.
By utilizing an iterative method, these models adjust questions based on
difficulty and content, responding to feedback from a simulated 'student'
model. A novel aspect of the research involved using GPT-4 as a 'teacher' to
create complex questions, with GPT-3.5 as the 'student' responding to these
challenges. This setup mirrors active learning, promoting deeper engagement.
The findings demonstrate GPT-4's superior ability to generate precise,
challenging questions and notable improvements in GPT-3.5's ability to handle
more complex problems after receiving instruction from GPT-4. These results
underscore the potential of LLMs to mimic and enhance active learning
scenarios, offering a promising path for AI in customized education. This
research contributes to understanding how AI can support personalized learning
experiences, highlighting the need for further exploration in various
educational contexts | cs.CL | Generative AI for Enhancing Active Learning in Education: A Comparative Study of GPT-3.5 and GPT-4 in Crafting Customized Test Questions |
2024-06-20T00:05:10Z | http://arxiv.org/pdf/2406.13898v1 | http://arxiv.org/abs/2406.13898v1 | Huthaifa I. Ashqar, Taqwa I. Alhadidi, Mohammed Elhenawy, Nour O. Khanfar | 2024-06-20T00:05:10Z | cs.CV, cs.CL, cs.CY | The integration of thermal imaging data with Multimodal Large Language Models
(MLLMs) constitutes an exciting opportunity for improving the safety and
functionality of autonomous driving systems and many Intelligent Transportation
Systems (ITS) applications. This study investigates whether MLLMs can
understand complex images from RGB and thermal cameras and detect objects
directly. Our goals were to 1) assess the ability of the MLLM to learn from
information from various sets, 2) detect objects and identify elements in
thermal cameras, 3) determine whether two independent modality images show the
same scene, and 4) learn all objects using different modalities. The findings
showed that both GPT-4 and Gemini were effective in detecting and classifying
objects in thermal images. Similarly, the Mean Absolute Percentage Error (MAPE)
for pedestrian classification was 70.39% and 81.48%, respectively. Moreover,
the MAPE for bike, car, and motorcycle detection were 78.4%, 55.81%, and
96.15%, respectively. Gemini produced MAPE of 66.53%, 59.35% and 78.18%
respectively. This finding further demonstrates that MLLM can identify thermal
images and can be employed in advanced imaging automation technologies for ITS
applications. | cs.CV | The Use of Multimodal Large Language Models to Detect Objects from Thermal Images: Transportation Applications |
2024-06-19T23:49:56Z | http://arxiv.org/pdf/2406.13893v1 | http://arxiv.org/abs/2406.13893v1 | Pablo Gamallo, Pablo Rodríguez, Iria de-Dios-Flores, Susana Sotelo, Silvia Paniagua, Daniel Bardanca, José Ramom Pichel, Marcos Garcia | 2024-06-19T23:49:56Z | cs.CL | Large language models (LLMs) have transformed natural language processing.
Yet, their predominantly English-centric training has led to biases and
performance disparities across languages. This imbalance marginalizes
minoritized languages, making equitable access to NLP technologies more
difficult for languages with lower resources, such as Galician. We present the
first two generative LLMs focused on Galician to bridge this gap. These models,
freely available as open-source resources, were trained using a GPT
architecture with 1.3B parameters on a corpus of 2.1B words. Leveraging
continual pretraining, we adapt to Galician two existing LLMs trained on larger
corpora, thus mitigating the data constraints that would arise if the training
were performed from scratch. The models were evaluated using human judgments
and task-based datasets from standardized benchmarks. These evaluations reveal
a promising performance, underscoring the importance of linguistic diversity in
generative models. | cs.CL | Open Generative Large Language Models for Galician |
2024-06-19T23:47:59Z | http://arxiv.org/pdf/2406.13892v1 | http://arxiv.org/abs/2406.13892v1 | Honghua Zhang, Po-Nien Kung, Masahiro Yoshida, Guy Van den Broeck, Nanyun Peng | 2024-06-19T23:47:59Z | cs.CL | Despite the success of Large Language Models (LLMs) on various tasks
following human instructions, controlling model generation at inference time
poses a persistent challenge. In this paper, we introduce Ctrl-G, an adaptable
framework that facilitates tractable and flexible control of LLM generation to
reliably follow logical constraints. Ctrl-G combines any production-ready LLM
with a Hidden Markov Model, enabling LLM outputs to adhere to logical
constraints represented as deterministic finite automata. We show that Ctrl-G,
when applied to a TULU2-7B model, outperforms GPT3.5 and GPT4 on the task of
interactive text editing: specifically, for the task of generating text
insertions/continuations following logical constraints, Ctrl-G achieves over
30% higher satisfaction rate in human evaluation compared to GPT4. When applied
to medium-size language models (e.g., GPT2-large), Ctrl-G also beats its
counterparts for constrained generation by large margins on standard
benchmarks. Additionally, as a proof-of-concept study, we experiment Ctrl-G on
the Grade School Math benchmark to assist LLM reasoning, foreshadowing the
application of Ctrl-G, as well as other constrained generation approaches,
beyond traditional language generation tasks. | cs.CL | Adaptable Logical Control for Large Language Models |
2024-06-19T23:44:25Z | http://arxiv.org/pdf/2406.13890v1 | http://arxiv.org/abs/2406.13890v1 | Weixiang Yan, Haitian Liu, Tengxiao Wu, Qian Chen, Wen Wang, Haoyuan Chai, Jiayi Wang, Weishan Zhao, Yixin Zhang, Renjun Zhang, Li Zhu | 2024-06-19T23:44:25Z | cs.CL, cs.AI | LLMs have achieved significant performance progress in various NLP
applications. However, LLMs still struggle to meet the strict requirements for
accuracy and reliability in the medical field and face many challenges in
clinical applications. Existing clinical diagnostic evaluation benchmarks for
evaluating medical agents powered by LLMs have severe limitations. Firstly,
most existing medical evaluation benchmarks face the risk of data leakage or
contamination. Secondly, existing benchmarks often neglect the characteristics
of multiple departments and specializations in modern medical practice.
Thirdly, existing evaluation methods are limited to multiple-choice questions,
which do not align with the real-world diagnostic scenarios. Lastly, existing
evaluation methods lack comprehensive evaluations of end-to-end real clinical
scenarios. These limitations in benchmarks in turn obstruct advancements of
LLMs and agents for medicine. To address these limitations, we introduce
ClinicalLab, a comprehensive clinical diagnosis agent alignment suite.
ClinicalLab includes ClinicalBench, an end-to-end multi-departmental clinical
diagnostic evaluation benchmark for evaluating medical agents and LLMs.
ClinicalBench is based on real cases that cover 24 departments and 150
diseases. ClinicalLab also includes four novel metrics (ClinicalMetrics) for
evaluating the effectiveness of LLMs in clinical diagnostic tasks. We evaluate
17 LLMs and find that their performance varies significantly across different
departments. Based on these findings, in ClinicalLab, we propose ClinicalAgent,
an end-to-end clinical agent that aligns with real-world clinical diagnostic
practices. We systematically investigate the performance and applicable
scenarios of variants of ClinicalAgent on ClinicalBench. Our findings
demonstrate the importance of aligning with modern medical practices in
designing medical agents. | cs.CL | ClinicalLab: Aligning Agents for Multi-Departmental Clinical Diagnostics in the Real World |
2024-06-19T23:30:01Z | http://arxiv.org/pdf/2406.13885v1 | http://arxiv.org/abs/2406.13885v1 | Hang Li, Tianlong Xu, Jiliang Tang, Qingsong Wen | 2024-06-19T23:30:01Z | cs.CL, cs.AI | Knowledge tagging for questions plays a crucial role in contemporary
intelligent educational applications, including learning progress diagnosis,
practice question recommendations, and course content organization.
Traditionally, these annotations are always conducted by pedagogical experts,
as the task requires not only a strong semantic understanding of both question
stems and knowledge definitions but also deep insights into connecting
question-solving logic with corresponding knowledge concepts. With the recent
emergence of advanced text encoding algorithms, such as pre-trained language
models, many researchers have developed automatic knowledge tagging systems
based on calculating the semantic similarity between the knowledge and question
embeddings. In this paper, we explore automating the task using Large Language
Models (LLMs), in response to the inability of prior encoding-based methods to
deal with the hard cases which involve strong domain knowledge and complicated
concept definitions. By showing the strong performance of zero- and few-shot
results over math questions knowledge tagging tasks, we demonstrate LLMs' great
potential in conquering the challenges faced by prior methods. Furthermore, by
proposing a reinforcement learning-based demonstration retriever, we
successfully exploit the great potential of different-sized LLMs in achieving
better performance results while keeping the in-context demonstration usage
efficiency high. | cs.CL | Knowledge Tagging System on Math Questions via LLMs with Flexible Demonstration Retriever |
2024-06-19T21:45:20Z | http://arxiv.org/pdf/2406.13862v1 | http://arxiv.org/abs/2406.13862v1 | Haochen Liu, Song Wang, Yaochen Zhu, Yushun Dong, Jundong Li | 2024-06-19T21:45:20Z | cs.CL, cs.AI | Large Language Models (LLMs) have shown unprecedented performance in various
real-world applications. However, they are known to generate factually
inaccurate outputs, a.k.a. the hallucination problem. In recent years,
incorporating external knowledge extracted from Knowledge Graphs (KGs) has
become a promising strategy to improve the factual accuracy of LLM-generated
outputs. Nevertheless, most existing explorations rely on LLMs themselves to
perform KG knowledge extraction, which is highly inflexible as LLMs can only
provide binary judgment on whether a certain knowledge (e.g., a knowledge path
in KG) should be used. In addition, LLMs tend to pick only knowledge with
direct semantic relationship with the input text, while potentially useful
knowledge with indirect semantics can be ignored. In this work, we propose a
principled framework KELP with three stages to handle the above problems.
Specifically, KELP is able to achieve finer granularity of flexible knowledge
extraction by generating scores for knowledge paths with input texts via latent
semantic matching. Meanwhile, knowledge paths with indirect semantic
relationships with the input text can also be considered via trained encoding
between the selected paths in KG and the input text. Experiments on real-world
datasets validate the effectiveness of KELP. | cs.CL | Knowledge Graph-Enhanced Large Language Models via Path Selection |
2024-06-19T21:40:35Z | http://arxiv.org/pdf/2406.15507v1 | http://arxiv.org/abs/2406.15507v1 | Haochen Liu, Song Wang, Chen Chen, Jundong Li | 2024-06-19T21:40:35Z | cs.CL, cs.LG | Few-shot Knowledge Graph (KG) Relational Reasoning aims to predict unseen
triplets (i.e., query triplets) for rare relations in KGs, given only several
triplets of these relations as references (i.e., support triplets). This task
has gained significant traction due to the widespread use of knowledge graphs
in various natural language processing applications. Previous approaches have
utilized meta-training methods and manually constructed meta-relation sets to
tackle this task. Recent efforts have focused on edge-mask-based methods, which
exploit the structure of the contextualized graphs of target triplets (i.e., a
subgraph containing relevant triplets in the KG). However, existing
edge-mask-based methods have limitations in extracting insufficient information
from KG and are highly influenced by spurious information in KG. To overcome
these challenges, we propose SAFER (Subgraph Adaptation for Few-shot Relational
Reasoning), a novel approach that effectively adapts the information in
contextualized graphs to various subgraphs generated from support and query
triplets to perform the prediction. Specifically, SAFER enables the extraction
of more comprehensive information from support triplets while minimizing the
impact of spurious information when predicting query triplets. Experimental
results on three prevalent datasets demonstrate the superiority of our proposed
framework SAFER. | cs.CL | Few-shot Knowledge Graph Relational Reasoning via Subgraph Adaptation |
2024-06-19T21:36:40Z | http://arxiv.org/pdf/2406.13858v1 | http://arxiv.org/abs/2406.13858v1 | Yuval Shalev, Amir Feder, Ariel Goldstein | 2024-06-19T21:36:40Z | cs.CL | Large language models (LLMs) have shown an impressive ability to perform
tasks believed to require thought processes. When the model does not document
an explicit thought process, it becomes difficult to understand the processes
occurring within its hidden layers and to determine if these processes can be
referred to as reasoning. We introduce a novel and interpretable analysis of
internal multi-hop reasoning processes in LLMs. We demonstrate that the
prediction process for compositional reasoning questions can be modeled using a
simple linear transformation between two semantic category spaces. We show that
during inference, the middle layers of the network generate highly
interpretable embeddings that represent a set of potential intermediate answers
for the multi-hop question. We use statistical analyses to show that a
corresponding subset of tokens is activated in the model's output, implying the
existence of parallel reasoning paths. These observations hold true even when
the model lacks the necessary knowledge to solve the task. Our findings can
help uncover the strategies that LLMs use to solve reasoning tasks, offering
insights into the types of thought processes that can emerge from artificial
intelligence. Finally, we also discuss the implication of cognitive modeling of
these results. | cs.CL | Distributional reasoning in LLMs: Parallel reasoning processes in multi-hop reasoning |
2024-06-19T21:19:37Z | http://arxiv.org/pdf/2406.13846v1 | http://arxiv.org/abs/2406.13846v1 | Kyoka Ono, Simon A. Lee | 2024-06-19T21:19:37Z | cs.CL, cs.LG | Recent research has explored how Language Models (LMs) can be used for
feature representation and prediction in tabular machine learning tasks. This
involves employing text serialization and supervised fine-tuning (SFT)
techniques. Despite the simplicity of these techniques, significant gaps remain
in our understanding of the applicability and reliability of LMs in this
context. Our study assesses how emerging LM technologies compare with
traditional paradigms in tabular machine learning and evaluates the feasibility
of adopting similar approaches with these advanced technologies. At the data
level, we investigate various methods of data representation and curation of
serialized tabular data, exploring their impact on prediction performance. At
the classification level, we examine whether text serialization combined with
LMs enhances performance on tabular datasets (e.g. class imbalance,
distribution shift, biases, and high dimensionality), and assess whether this
method represents a state-of-the-art (SOTA) approach for addressing tabular
machine learning challenges. Our findings reveal current pre-trained models
should not replace conventional approaches. | cs.CL | Text Serialization and Their Relationship with the Conventional Paradigms of Tabular Machine Learning |
2024-06-19T21:11:01Z | http://arxiv.org/pdf/2406.13842v1 | http://arxiv.org/abs/2406.13842v1 | Alexander Blatt, Aravind Krishnan, Dietrich Klakow | 2024-06-19T21:11:01Z | cs.CL, cs.SD, eess.AS | Utilizing air-traffic control (ATC) data for downstream natural-language
processing tasks requires preprocessing steps. Key steps are the transcription
of the data via automatic speech recognition (ASR) and speaker diarization,
respectively speaker role detection (SRD) to divide the transcripts into pilot
and air-traffic controller (ATCO) transcripts. While traditional approaches
take on these tasks separately, we propose a transformer-based joint ASR-SRD
system that solves both tasks jointly while relying on a standard ASR
architecture. We compare this joint system against two cascaded approaches for
ASR and SRD on multiple ATC datasets. Our study shows in which cases our joint
system can outperform the two traditional approaches and in which cases the
other architectures are preferable. We additionally evaluate how acoustic and
lexical differences influence all architectures and show how to overcome them
for our joint architecture. | cs.CL | Joint vs Sequential Speaker-Role Detection and Automatic Speech Recognition for Air-traffic Control |
2024-06-19T21:07:35Z | http://arxiv.org/pdf/2406.13840v1 | http://arxiv.org/abs/2406.13840v1 | Davit Abrahamyan, Fatemeh H. Fard | 2024-06-19T21:07:35Z | cs.AI, cs.CL | Developers spend much time finding information that is relevant to their
questions. Stack Overflow has been the leading resource, and with the advent of
Large Language Models (LLMs), generative models such as ChatGPT are used
frequently. However, there is a catch in using each one separately. Searching
for answers is time-consuming and tedious, as shown by the many tools developed
by researchers to address this issue. On the other, using LLMs is not reliable,
as they might produce irrelevant or unreliable answers (i.e., hallucination).
In this work, we present StackRAG, a retrieval-augmented Multiagent generation
tool based on LLMs that combines the two worlds: aggregating the knowledge from
SO to enhance the reliability of the generated answers. Initial evaluations
show that the generated answers are correct, accurate, relevant, and useful. | cs.AI | StackRAG Agent: Improving Developer Answers with Retrieval-Augmented Generation |
2024-06-19T20:47:36Z | http://arxiv.org/pdf/2406.13828v1 | http://arxiv.org/abs/2406.13828v1 | Tanawan Premsri, Parisa Kordjamshidi | 2024-06-19T20:47:36Z | cs.CL | Recent research shows that more data and larger models can provide more
accurate solutions to natural language problems requiring reasoning. However,
models can easily fail to provide solutions in unobserved complex input
compositions due to not achieving the level of abstraction required for
generalizability. To alleviate this issue, we propose training the language
models with neuro-symbolic techniques that can exploit the logical rules of
reasoning as constraints and provide additional supervision sources to the
model. Training models to adhere to the regulations of reasoning pushes them to
make more effective abstractions needed for generalizability and transfer
learning. We focus on a challenging problem of spatial reasoning over text. Our
results on various benchmarks using multiple language models confirm our
hypothesis of effective domain transfer based on neuro-symbolic training. | cs.CL | Neuro-symbolic Training for Reasoning over Spatial Language |
2024-06-19T20:47:23Z | http://arxiv.org/pdf/2406.13827v2 | http://arxiv.org/abs/2406.13827v2 | Lucy Horowitz, Ryan Hathaway | 2024-06-27T22:31:07Z | cs.CL | In this paper, we fine-tuned three pre-trained BERT models on the task of
"definition extraction" from mathematical English written in LaTeX. This is
presented as a binary classification problem, where either a sentence contains
a definition of a mathematical term or it does not. We used two original data
sets, "Chicago" and "TAC," to fine-tune and test these models. We also tested
on WFMALL, a dataset presented by Vanetik and Litvak in 2021 and compared the
performance of our models to theirs. We found that a high-performance
Sentence-BERT transformer model performed best based on overall accuracy,
recall, and precision metrics, achieving comparable results to the earlier
models with less computational effort. | cs.CL | Fine-Tuning BERTs for Definition Extraction from Mathematical Text |
2024-06-19T20:31:34Z | http://arxiv.org/pdf/2406.13820v1 | http://arxiv.org/abs/2406.13820v1 | Julia Mendelsohn, Maya Vijan, Dallas Card, Ceren Budak | 2024-06-19T20:31:34Z | cs.CL | Social media enables activists to directly communicate with the public and
provides a space for movement leaders, participants, bystanders, and opponents
to collectively construct and contest narratives. Focusing on Twitter messages
from social movements surrounding three issues in 2018-2019 (guns, immigration,
and LGBTQ rights), we create a codebook, annotated dataset, and computational
models to detect diagnostic (problem identification and attribution),
prognostic (proposed solutions and tactics), and motivational (calls to action)
framing strategies. We conduct an in-depth unsupervised linguistic analysis of
each framing strategy, and uncover cross-movement similarities in associations
between framing and linguistic features such as pronouns and deontic modal
verbs. Finally, we compare framing strategies across issues and other social,
cultural, and interactional contexts. For example, we show that diagnostic
framing is more common in replies than original broadcast posts, and that
social movement organizations focus much more on prognostic and motivational
framing than journalists and ordinary citizens. | cs.CL | Framing Social Movements on Social Media: Unpacking Diagnostic, Prognostic, and Motivational Strategies |
2024-06-19T20:14:14Z | http://arxiv.org/pdf/2406.13807v2 | http://arxiv.org/abs/2406.13807v2 | Alessandro Suglia, Claudio Greco, Katie Baker, Jose L. Part, Ioannis Papaioannou, Arash Eshghi, Ioannis Konstas, Oliver Lemon | 2024-06-21T09:53:41Z | cs.CV, cs.AI, cs.CL | AI personal assistants deployed via robots or wearables require embodied
understanding to collaborate with humans effectively. However, current
Vision-Language Models (VLMs) primarily focus on third-person view videos,
neglecting the richness of egocentric perceptual experience. To address this
gap, we propose three key contributions. First, we introduce the Egocentric
Video Understanding Dataset (EVUD) for training VLMs on video captioning and
question answering tasks specific to egocentric videos. Second, we present
AlanaVLM, a 7B parameter VLM trained using parameter-efficient methods on EVUD.
Finally, we evaluate AlanaVLM's capabilities on OpenEQA, a challenging
benchmark for embodied video question answering. Our model achieves
state-of-the-art performance, outperforming open-source models including strong
Socratic models using GPT-4 as a planner by 3.6%. Additionally, we outperform
Claude 3 and Gemini Pro Vision 1.0 and showcase competitive results compared to
Gemini Pro 1.5 and GPT-4V, even surpassing the latter in spatial reasoning.
This research paves the way for building efficient VLMs that can be deployed in
robots or wearables, leveraging embodied video understanding to collaborate
seamlessly with humans in everyday tasks, contributing to the next generation
of Embodied AI. | cs.CV | AlanaVLM: A Multimodal Embodied AI Foundation Model for Egocentric Video Understanding |
2024-06-19T20:13:42Z | http://arxiv.org/pdf/2406.13805v1 | http://arxiv.org/abs/2406.13805v1 | Yufang Hou, Alessandra Pascale, Javier Carnerero-Cano, Tigran Tchrakian, Radu Marinescu, Elizabeth Daly, Inkit Padhi, Prasanna Sattigeri | 2024-06-19T20:13:42Z | cs.CL, cs.AI, cs.LG | Retrieval-augmented generation (RAG) has emerged as a promising solution to
mitigate the limitations of large language models (LLMs), such as
hallucinations and outdated information. However, it remains unclear how LLMs
handle knowledge conflicts arising from different augmented retrieved passages,
especially when these passages originate from the same source and have equal
trustworthiness. In this work, we conduct a comprehensive evaluation of
LLM-generated answers to questions that have varying answers based on
contradictory passages from Wikipedia, a dataset widely regarded as a
high-quality pre-training resource for most LLMs. Specifically, we introduce
WikiContradict, a benchmark consisting of 253 high-quality, human-annotated
instances designed to assess LLM performance when augmented with retrieved
passages containing real-world knowledge conflicts. We benchmark a diverse
range of both closed and open-source LLMs under different QA scenarios,
including RAG with a single passage, and RAG with 2 contradictory passages.
Through rigorous human evaluations on a subset of WikiContradict instances
involving 5 LLMs and over 3,500 judgements, we shed light on the behaviour and
limitations of these models. For instance, when provided with two passages
containing contradictory facts, all models struggle to generate answers that
accurately reflect the conflicting nature of the context, especially for
implicit conflicts requiring reasoning. Since human evaluation is costly, we
also introduce an automated model that estimates LLM performance using a strong
open-source language model, achieving an F-score of 0.8. Using this automated
metric, we evaluate more than 1,500 answers from seven LLMs across all
WikiContradict instances. To facilitate future work, we release WikiContradict
on: https://ibm.biz/wikicontradict. | cs.CL | WikiContradict: A Benchmark for Evaluating LLMs on Real-World Knowledge Conflicts from Wikipedia |
2024-06-19T20:07:37Z | http://arxiv.org/pdf/2406.13803v1 | http://arxiv.org/abs/2406.13803v1 | Sam Musker, Alex Duchnowski, Raphaël Millière, Ellie Pavlick | 2024-06-19T20:07:37Z | cs.CL | Analogical reasoning is considered core to human learning and cognition.
Recent studies have compared the analogical reasoning abilities of human
subjects and Large Language Models (LLMs) on abstract symbol manipulation
tasks, such as letter string analogies. However, these studies largely neglect
analogical reasoning over semantically meaningful symbols, such as natural
language words. This ability to draw analogies that link language to
non-linguistic domains, which we term semantic structure-mapping, is thought to
play a crucial role in language acquisition and broader cognitive development.
We test human subjects and LLMs on analogical reasoning tasks that require the
transfer of semantic structure and content from one domain to another. Advanced
LLMs match human performance across many task variations. However, humans and
LLMs respond differently to certain task variations and semantic distractors.
Overall, our data suggest that LLMs are approaching human-level performance on
these important cognitive tasks, but are not yet entirely human like. | cs.CL | Semantic Structure-Mapping in LLM and Human Analogical Reasoning |
2024-06-19T19:35:14Z | http://arxiv.org/pdf/2406.13791v2 | http://arxiv.org/abs/2406.13791v2 | Amelie Gyrard, Seyedali Mohammadi, Manas Gaur, Antonio Kung | 2024-06-29T15:29:56Z | cs.AI, cs.CL, cs.CY, cs.LG | Sustainable Development Goals (SDGs) give the UN a road map for development
with Agenda 2030 as a target. SDG3 "Good Health and Well-Being" ensures healthy
lives and promotes well-being for all ages. Digital technologies can support
SDG3. Burnout and even depression could be reduced by encouraging better
preventive health. Due to the lack of patient knowledge and focus to take care
of their health, it is necessary to help patients before it is too late. New
trends such as positive psychology and mindfulness are highly encouraged in the
USA. Digital Twin (DT) can help with the continuous monitoring of emotion using
physiological signals (e.g., collected via wearables). Digital twins facilitate
monitoring and provide constant health insight to improve quality of life and
well-being with better personalization. Healthcare DT challenges are
standardizing data formats, communication protocols, and data exchange
mechanisms. To achieve those data integration and knowledge challenges, we
designed the Mental Health Knowledge Graph (ontology and dataset) to boost
mental health. The Knowledge Graph (KG) acquires knowledge from ontology-based
mental health projects classified within the LOV4IoT ontology catalog (Emotion,
Depression, and Mental Health). Furthermore, the KG is mapped to standards
(e.g., ontologies) when possible. Standards from ETSI SmartM2M, ITU/WHO, ISO,
W3C, NIST, and IEEE are relevant to mental health. | cs.AI | IoT-Based Preventive Mental Health Using Knowledge Graphs and Standards for Better Well-Being |
2024-06-19T19:11:22Z | http://arxiv.org/pdf/2406.13781v1 | http://arxiv.org/abs/2406.13781v1 | Tan M. Nguyen, Tam Nguyen, Nhat Ho, Andrea L. Bertozzi, Richard G. Baraniuk, Stanley J. Osher | 2024-06-19T19:11:22Z | cs.LG, cs.AI, cs.CL, cs.CV, stat.ML | Self-attention is key to the remarkable success of transformers in sequence
modeling tasks including many applications in natural language processing and
computer vision. Like neural network layers, these attention mechanisms are
often developed by heuristics and experience. To provide a principled framework
for constructing attention layers in transformers, we show that the
self-attention corresponds to the support vector expansion derived from a
support vector regression problem, whose primal formulation has the form of a
neural network layer. Using our framework, we derive popular attention layers
used in practice and propose two new attentions: 1) the Batch Normalized
Attention (Attention-BN) derived from the batch normalization layer and 2) the
Attention with Scaled Head (Attention-SH) derived from using less training data
to fit the SVR model. We empirically demonstrate the advantages of the
Attention-BN and Attention-SH in reducing head redundancy, increasing the
model's accuracy, and improving the model's efficiency in a variety of
practical applications including image and time-series classification. | cs.LG | A Primal-Dual Framework for Transformers and Neural Networks |
2024-06-19T19:06:36Z | http://arxiv.org/pdf/2406.13779v1 | http://arxiv.org/abs/2406.13779v1 | Tianchi Cai, Zhiwen Tan, Xierui Song, Tao Sun, Jiyan Jiang, Yunqi Xu, Yinger Zhang, Jinjie Gu | 2024-06-19T19:06:36Z | cs.CL | Retrieval Augmented Generation (RAG) has become prevalent in
question-answering (QA) tasks due to its ability of utilizing search engine to
enhance the quality of long-form question-answering (LFQA). Despite the
emergence of various open source methods and web-enhanced commercial systems
such as Bing Chat, two critical problems remain unsolved, i.e., the lack of
factuality and clear logic in the generated long-form answers. In this paper,
we remedy these issues via a systematic study on answer generation in
web-enhanced LFQA. Specifically, we first propose a novel outline-enhanced
generator to achieve clear logic in the generation of multifaceted answers and
construct two datasets accordingly. Then we propose a factuality optimization
method based on a carefully designed doubly fine-grained RLHF framework, which
contains automatic evaluation and reward modeling in different levels of
granularity. Our generic framework comprises conventional fine-grained RLHF
methods as special cases. Extensive experiments verify the superiority of our
proposed \textit{Factuality-optimized RAG (FoRAG)} method on both English and
Chinese benchmarks. In particular, when applying our method to Llama2-7B-chat,
the derived model FoRAG-L-7B outperforms WebGPT-175B in terms of three commonly
used metrics (i.e., coherence, helpfulness, and factuality), while the number
of parameters is much smaller (only 1/24 of that of WebGPT-175B). Our datasets
and models are made publicly available for better reproducibility:
https://huggingface.co/forag. | cs.CL | FoRAG: Factuality-optimized Retrieval Augmented Generation for Web-enhanced Long-form Question Answering |
2024-06-19T19:02:44Z | http://arxiv.org/pdf/2406.13777v1 | http://arxiv.org/abs/2406.13777v1 | Shruthi K. Hiremath, Thomas Ploetz | 2024-06-19T19:02:44Z | cs.LG, cs.CL | Human Activity Recognition is a time-series analysis problem. A popular
analysis procedure used by the community assumes an optimal window length to
design recognition pipelines. However, in the scenario of smart homes, where
activities are of varying duration and frequency, the assumption of a constant
sized window does not hold. Additionally, previous works have shown these
activities to be made up of building blocks. We focus on identifying these
underlying building blocks--structural constructs, with the use of large
language models. Identifying these constructs can be beneficial especially in
recognizing short-duration and infrequent activities. We also propose the
development of an activity recognition procedure that uses these building
blocks to model activities, thus helping the downstream task of activity
monitoring in smart homes. | cs.LG | Game of LLMs: Discovering Structural Constructs in Activities using Large Language Models |
2024-06-19T18:38:11Z | http://arxiv.org/pdf/2406.13770v1 | http://arxiv.org/abs/2406.13770v1 | Stefan K. Nielsen, Laziz U. Abdullaev, Rachel Teo, Tan M. Nguyen | 2024-06-19T18:38:11Z | cs.LG, cs.AI, cs.CL, cs.CV, stat.ML | Pairwise dot-product self-attention is key to the success of transformers
that achieve state-of-the-art performance across a variety of applications in
language and vision. This dot-product self-attention computes attention weights
among the input tokens using Euclidean distance, which makes the model prone to
representation collapse and vulnerable to contaminated samples. In this paper,
we propose using a Mahalanobis distance metric for computing the attention
weights to stretch the underlying feature space in directions of high
contextual relevance. In particular, we define a hyper-ellipsoidal neighborhood
around each query to increase the attention weights of the tokens lying in the
contextually important directions. We term this novel class of attention
Elliptical Attention. Our Elliptical Attention provides two benefits: 1)
reducing representation collapse and 2) enhancing the model's robustness as the
Elliptical Attention pays more attention to contextually relevant information
rather than focusing on some small subset of informative features. We
empirically demonstrate the advantages of Elliptical Attention over the
baseline dot-product attention and state-of-the-art attention methods on
various practical tasks, including object classification, image segmentation,
and language modeling across different data modalities. | cs.LG | Elliptical Attention |
2024-06-19T18:26:19Z | http://arxiv.org/pdf/2406.13764v1 | http://arxiv.org/abs/2406.13764v1 | Yuan Yang, Siheng Xiong, Ali Payani, Ehsan Shareghi, Faramarz Fekri | 2024-06-19T18:26:19Z | cs.CL | Large Language Models (LLMs) have shown superior capability to solve
reasoning problems with programs. While being a promising direction, most of
such frameworks are trained and evaluated in settings with a prior knowledge of
task requirements. However, as LLMs become more capable, it is necessary to
assess their reasoning abilities in more realistic scenarios where many
real-world problems are open-ended with ambiguous scope, and often require
multiple formalisms to solve. To investigate this, we introduce the task of
reasoning in the wild, where an LLM is tasked to solve a reasoning problem of
unknown type by identifying the subproblems and their corresponding formalisms,
and writing a program to solve each subproblem, guided by a tactic. We create a
large tactic-guided trajectory dataset containing detailed solutions to a
diverse set of reasoning problems, ranging from well-defined single-form
reasoning (e.g., math, logic), to ambiguous and hybrid ones (e.g., commonsense,
combined math and logic). This allows us to test various aspects of LLMs
reasoning at the fine-grained level such as the selection and execution of
tactics, and the tendency to take undesired shortcuts. In experiments, we
highlight that existing LLMs fail significantly on problems with ambiguous and
mixed scope, revealing critical limitations and overfitting issues (e.g.
accuracy on GSM8K drops by at least 50\%). We further show the potential of
finetuning a local LLM on the tactic-guided trajectories in achieving better
performance. Project repo is available at
github.com/gblackout/Reason-in-the-Wild | cs.CL | Can LLMs Reason in the Wild with Programs? |
2024-06-19T18:22:32Z | http://arxiv.org/pdf/2406.13762v1 | http://arxiv.org/abs/2406.13762v1 | Rachel S. Y. Teo, Tan M. Nguyen | 2024-06-19T18:22:32Z | cs.LG, cs.AI, cs.CL, cs.CV, stat.ML | The remarkable success of transformers in sequence modeling tasks, spanning
various applications in natural language processing and computer vision, is
attributed to the critical role of self-attention. Similar to the development
of most deep learning models, the construction of these attention mechanisms
rely on heuristics and experience. In our work, we derive self-attention from
kernel principal component analysis (kernel PCA) and show that self-attention
projects its query vectors onto the principal component axes of its key matrix
in a feature space. We then formulate the exact formula for the value matrix in
self-attention, theoretically and empirically demonstrating that this value
matrix captures the eigenvectors of the Gram matrix of the key vectors in
self-attention. Leveraging our kernel PCA framework, we propose Attention with
Robust Principal Components (RPC-Attention), a novel class of robust attention
that is resilient to data contamination. We empirically demonstrate the
advantages of RPC-Attention over softmax attention on the ImageNet-1K object
classification, WikiText-103 language modeling, and ADE20K image segmentation
task. | cs.LG | Unveiling the Hidden Structure of Self-Attention via Kernel Principal Component Analysis |
2024-06-19T18:01:08Z | http://arxiv.org/pdf/2406.13748v1 | http://arxiv.org/abs/2406.13748v1 | Taiming Lu, Philipp Koehn | 2024-06-19T18:01:08Z | cs.CL, cs.LG | This paper investigates the propagation of harmful information in
multilingual large language models (LLMs) and evaluates the efficacy of various
unlearning methods. We demonstrate that fake information, regardless of the
language it is in, once introduced into these models through training data, can
spread across different languages, compromising the integrity and reliability
of the generated content. Our findings reveal that standard unlearning
techniques, which typically focus on English data, are insufficient in
mitigating the spread of harmful content in multilingual contexts and could
inadvertently reinforce harmful content across languages. We show that only by
addressing harmful responses in both English and the original language of the
harmful data can we effectively eliminate generations for all languages. This
underscores the critical need for comprehensive unlearning strategies that
consider the multilingual nature of modern LLMs to enhance their safety and
reliability across diverse linguistic landscapes. | cs.CL | Every Language Counts: Learn and Unlearn in Multilingual LLMs |
2024-06-19T18:00:07Z | http://arxiv.org/pdf/2406.13743v2 | http://arxiv.org/abs/2406.13743v2 | Baiqi Li, Zhiqiu Lin, Deepak Pathak, Jiayao Li, Yixin Fei, Kewen Wu, Tiffany Ling, Xide Xia, Pengchuan Zhang, Graham Neubig, Deva Ramanan | 2024-06-21T19:09:36Z | cs.CV, cs.AI, cs.CL, cs.LG, cs.MM | While text-to-visual models now produce photo-realistic images and videos,
they struggle with compositional text prompts involving attributes,
relationships, and higher-order reasoning such as logic and comparison. In this
work, we conduct an extensive human study on GenAI-Bench to evaluate the
performance of leading image and video generation models in various aspects of
compositional text-to-visual generation. We also compare automated evaluation
metrics against our collected human ratings and find that VQAScore -- a metric
measuring the likelihood that a VQA model views an image as accurately
depicting the prompt -- significantly outperforms previous metrics such as
CLIPScore. In addition, VQAScore can improve generation in a black-box manner
(without finetuning) via simply ranking a few (3 to 9) candidate images.
Ranking by VQAScore is 2x to 3x more effective than other scoring methods like
PickScore, HPSv2, and ImageReward at improving human alignment ratings for
DALL-E 3 and Stable Diffusion, especially on compositional prompts that require
advanced visio-linguistic reasoning. We will release a new GenAI-Rank benchmark
with over 40,000 human ratings to evaluate scoring metrics on ranking images
generated from the same prompt. Lastly, we discuss promising areas for
improvement in VQAScore, such as addressing fine-grained visual details. We
will release all human ratings (over 80,000) to facilitate scientific
benchmarking of both generative models and automated metrics. | cs.CV | GenAI-Bench: Evaluating and Improving Compositional Text-to-Visual Generation |
2024-06-19T17:24:36Z | http://arxiv.org/pdf/2406.13720v1 | http://arxiv.org/abs/2406.13720v1 | Md Ibrahim Ibne Alam, Parikshit Ram, Soham Dan, Horst Samulowitz, Koushik Kar | 2024-06-19T17:24:36Z | cs.CL, cs.LG | Large Language Models (LLMs) have been observed to perform well on a wide
range of downstream tasks when fine-tuned on domain-specific data. However,
such data may not be readily available in many applications, motivating
zero-shot or few-shot approaches using domain-adjacent models. While several
fine-tuned models for various tasks are available, finding an appropriate
domain-adjacent model for a given task is often not straight forward. In this
paper, we study DAFT-E, a framework that utilizes an Ensemble of
Domain-Adjacent Fine-Tuned Foundation Models for few-shot problems. We show
that for zero-shot problems, this ensembling method provides an accuracy
performance close to that of the single best model. With few-shot problems,
this performance improves further, at which point DEFT-E can outperform any
single domain-adjacent model while requiring much less data for domain-specific
fine-tuning. | cs.CL | On the Utility of Domain-Adjacent Fine-Tuned Model Ensembles for Few-shot Problems |
2024-06-19T17:20:28Z | http://arxiv.org/pdf/2406.13718v1 | http://arxiv.org/abs/2406.13718v1 | Niyati Bafna, Kenton Murray, David Yarowsky | 2024-06-19T17:20:28Z | cs.CL | While large language models exhibit certain cross-lingual generalization
capabilities, they suffer from performance degradation (PD) on unseen
closely-related languages (CRLs) and dialects relative to their high-resource
language neighbour (HRLN). However, we currently lack a fundamental
understanding of what kinds of linguistic distances contribute to PD, and to
what extent. Furthermore, studies of cross-lingual generalization are
confounded by unknown quantities of CRL language traces in the training data,
and by the frequent lack of availability of evaluation data in lower-resource
related languages and dialects. To address these issues, we model phonological,
morphological, and lexical distance as Bayesian noise processes to synthesize
artificial languages that are controllably distant from the HRLN. We analyse PD
as a function of underlying noise parameters, offering insights on model
robustness to isolated and composed linguistic phenomena, and the impact of
task and HRL characteristics on PD. We calculate parameter posteriors on real
CRL-HRLN pair data and show that they follow computed trends of artificial
languages, demonstrating the viability of our noisers. Our framework offers a
cheap solution to estimating task performance on an unseen CRL given HRLN
performance using its posteriors, as well as for diagnosing observed PD on a
CRL in terms of its linguistic distances from its HRLN, and opens doors to
principled methods of mitigating performance degradation. | cs.CL | Evaluating Large Language Models along Dimensions of Language Variation: A Systematik Invesdigatiom uv Cross-lingual Generalization |
2024-06-19T17:14:41Z | http://arxiv.org/pdf/2406.13714v1 | http://arxiv.org/abs/2406.13714v1 | Vansh Nagpal, Siva Likitha Valluru, Kausik Lakkaraju, Biplav Srivastava | 2024-06-19T17:14:41Z | cs.AI, cs.CL | A common, yet regular, decision made by people, whether healthy or with any
health condition, is to decide what to have in meals like breakfast, lunch, and
dinner, consisting of a combination of foods for appetizer, main course, side
dishes, desserts, and beverages. However, often this decision is seen as a
trade-off between nutritious choices (e.g., low salt and sugar) or convenience
(e.g., inexpensive, fast to prepare/obtain, taste better). In this preliminary
work, we present a data-driven approach for the novel meal recommendation
problem that can explore and balance choices for both considerations while also
reasoning about a food's constituents and cooking process. Beyond the problem
formulation, our contributions also include a goodness measure, a recipe
conversion method from text to the recently introduced multimodal rich recipe
representation (R3) format, and learning methods using contextual bandits that
show promising results. | cs.AI | BEACON: Balancing Convenience and Nutrition in Meals With Long-Term Group Recommendations and Reasoning on Multimodal Recipes |
2024-06-19T17:11:51Z | http://arxiv.org/pdf/2406.13713v1 | http://arxiv.org/abs/2406.13713v1 | Mahaman Sanoussi Yahaya Alassan, Jessica López Espejel, Merieme Bouhandi, Walid Dahhane, El Hassane Ettifouri | 2024-06-19T17:11:51Z | cs.CL | In the rapidly evolving landscape of Natural Language Processing (NLP), Large
Language Models (LLMs) have demonstrated remarkable capabilities in tasks such
as question answering (QA). However, the accessibility and practicality of
utilizing these models for industrial applications pose significant challenges,
particularly concerning cost-effectiveness, inference speed, and resource
efficiency. This paper presents a comprehensive benchmarking study comparing
open-source LLMs with their non-open-source counterparts on the task of
question answering. Our objective is to identify open-source alternatives
capable of delivering comparable performance to proprietary models while being
lightweight in terms of resource requirements and suitable for Central
Processing Unit (CPU)-based inference. Through rigorous evaluation across
various metrics including accuracy, inference speed, and resource consumption,
we aim to provide insights into selecting efficient LLMs for real-world
applications. Our findings shed light on viable open-source alternatives that
offer acceptable performance and efficiency, addressing the pressing need for
accessible and efficient NLP solutions in industry settings. | cs.CL | Benchmarking Open-Source Language Models for Efficient Question Answering in Industrial Applications |
2024-06-19T16:58:32Z | http://arxiv.org/pdf/2406.13706v1 | http://arxiv.org/abs/2406.13706v1 | Natalie Grace Brigham, Chongjiu Gao, Tadayoshi Kohno, Franziska Roesner, Niloofar Mireshghallah | 2024-06-19T16:58:32Z | cs.CL, cs.AI, cs.CY | Journalists are among the many users of large language models (LLMs). To
better understand the journalist-AI interactions, we conduct a study of LLM
usage by two news agencies through browsing the WildChat dataset, identifying
candidate interactions, and verifying them by matching to online published
articles. Our analysis uncovers instances where journalists provide sensitive
material such as confidential correspondence with sources or articles from
other agencies to the LLM as stimuli and prompt it to generate articles, and
publish these machine-generated articles with limited intervention (median
output-publication ROUGE-L of 0.62). Based on our findings, we call for further
research into what constitutes responsible use of AI, and the establishment of
clear guidelines and best practices on using LLMs in a journalistic context. | cs.CL | Breaking News: Case Studies of Generative AI's Use in Journalism |
2024-06-19T16:52:22Z | http://arxiv.org/pdf/2406.13698v1 | http://arxiv.org/abs/2406.13698v1 | Shun Wang, Ge Zhang, Han Wu, Tyler Loakman, Wenhao Huang, Chenghua Lin | 2024-06-19T16:52:22Z | cs.CL | Machine Translation (MT) has developed rapidly since the release of Large
Language Models and current MT evaluation is performed through comparison with
reference human translations or by predicting quality scores from human-labeled
data. However, these mainstream evaluation methods mainly focus on fluency and
factual reliability, whilst paying little attention to figurative quality. In
this paper, we investigate the figurative quality of MT and propose a set of
human evaluation metrics focused on the translation of figurative language. We
additionally present a multilingual parallel metaphor corpus generated by
post-editing. Our evaluation protocol is designed to estimate four aspects of
MT: Metaphorical Equivalence, Emotion, Authenticity, and Quality. In doing so,
we observe that translations of figurative expressions display different traits
from literal ones. | cs.CL | MMTE: Corpus and Metrics for Evaluating Machine Translation Quality of Metaphorical Language |
2024-06-19T16:47:44Z | http://arxiv.org/pdf/2407.00071v1 | http://arxiv.org/abs/2407.00071v1 | Mert Esencan, Tarun Advaith Kumar, Ata Akbari Asanjan, P. Aaron Lott, Masoud Mohseni, Can Unlu, Davide Venturelli, Alan Ho | 2024-06-19T16:47:44Z | cs.AI, cs.CL, cs.ET, cs.LG | Recent Large Language Models (LLMs) have demonstrated impressive capabilities
at tasks that require human intelligence and are a significant step towards
human-like artificial intelligence (AI). Yet the performance of LLMs at
reasoning tasks have been subpar and the reasoning capability of LLMs is a
matter of significant debate. While it has been shown that the choice of the
prompting technique to the LLM can alter its performance on a multitude of
tasks, including reasoning, the best performing techniques require human-made
prompts with the knowledge of the tasks at hand. We introduce a framework for
what we call Combinatorial Reasoning (CR), a fully-automated prompting method,
where reasons are sampled from an LLM pipeline and mapped into a Quadratic
Unconstrained Binary Optimization (QUBO) problem. The framework investigates
whether QUBO solutions can be profitably used to select a useful subset of the
reasons to construct a Chain-of-Thought style prompt. We explore the
acceleration of CR with specialized solvers. We also investigate the
performance of simpler zero-shot strategies such as linear majority rule or
random selection of reasons. Our preliminary study indicates that coupling a
combinatorial solver to generative AI pipelines is an interesting avenue for AI
reasoning and elucidates design principles for future CR methods. | cs.AI | Combinatorial Reasoning: Selecting Reasons in Generative AI Pipelines via Combinatorial Optimization |
2024-06-19T16:43:56Z | http://arxiv.org/pdf/2406.15504v1 | http://arxiv.org/abs/2406.15504v1 | Zipeng Liu, Likang Wu, Ming He, Zhong Guan, Hongke Zhao, Nan Feng | 2024-06-19T16:43:56Z | cs.CL, cs.LG | Significant efforts have been directed toward integrating powerful Large
Language Models (LLMs) with diverse modalities, particularly focusing on the
fusion of vision, language, and audio data. However, the graph-structured data,
inherently rich in structural and domain-specific knowledge, have not yet been
gracefully adapted to LLMs. Existing methods either describe the graph with raw
text, suffering the loss of graph structural information, or feed Graph Neural
Network (GNN) embeddings directly into LLM at the cost of losing semantic
representation. To bridge this gap, we introduce an innovative, end-to-end
modality-aligning framework, equipped with a pretrained Dual-Residual Vector
Quantized-Variational AutoEncoder (Dr.E). This framework is specifically
designed to facilitate token-level alignment with LLMs, enabling an effective
translation of the intrinsic `language' of graphs into comprehensible natural
language. Our experimental evaluations on standard GNN node classification
tasks demonstrate competitive performance against other state-of-the-art
approaches. Additionally, our framework ensures interpretability, efficiency,
and robustness, with its effectiveness further validated under both fine-tuning
and few-shot settings. This study marks the first successful endeavor to
achieve token-level alignment between GNNs and LLMs. | cs.CL | Dr.E Bridges Graphs with Large Language Models through Words |
2024-06-19T16:42:57Z | http://arxiv.org/pdf/2406.13692v1 | http://arxiv.org/abs/2406.13692v1 | Di Wu, Jia-Chen Gu, Fan Yin, Nanyun Peng, Kai-Wei Chang | 2024-06-19T16:42:57Z | cs.CL | Retrieval-augmented language models (RALMs) have shown strong performance and
wide applicability in knowledge-intensive tasks. However, there are significant
trustworthiness concerns as RALMs are prone to generating unfaithful outputs,
including baseless information or contradictions with the retrieved context.
This paper proposes SynCheck, a lightweight monitor that leverages fine-grained
decoding dynamics including sequence likelihood, uncertainty quantification,
context influence, and semantic alignment to synchronously detect unfaithful
sentences. By integrating efficiently measurable and complementary signals,
SynCheck enables accurate and immediate feedback and intervention, achieving
0.85 AUROC in detecting faithfulness errors across six long-form
retrieval-augmented generation tasks, improving prior best method by 4%.
Leveraging SynCheck, we further introduce FOD, a faithfulness-oriented decoding
algorithm guided by beam search for long-form retrieval-augmented generation.
Empirical results demonstrate that FOD outperforms traditional strategies such
as abstention, reranking, or contrastive decoding significantly in terms of
faithfulness, achieving over 10% improvement across six datasets. | cs.CL | Synchronous Faithfulness Monitoring for Trustworthy Retrieval-Augmented Generation |
2024-06-19T16:30:58Z | http://arxiv.org/pdf/2406.13677v1 | http://arxiv.org/abs/2406.13677v1 | Erik Derner, Sara Sansalvador de la Fuente, Yoan Gutiérrez, Paloma Moreda, Nuria Oliver | 2024-06-19T16:30:58Z | cs.CL, cs.CY | Gender bias in text corpora used in various natural language processing (NLP)
contexts, such as for training large language models (LLMs), can lead to the
perpetuation and amplification of societal inequalities. This is particularly
pronounced in gendered languages like Spanish or French, where grammatical
structures inherently encode gender, making the bias analysis more challenging.
Existing methods designed for English are inadequate for this task due to the
intrinsic linguistic differences between English and gendered languages. This
paper introduces a novel methodology that leverages the contextual
understanding capabilities of LLMs to quantitatively analyze gender
representation in Spanish corpora. By utilizing LLMs to identify and classify
gendered nouns and pronouns in relation to their reference to human entities,
our approach provides a nuanced analysis of gender biases. We empirically
validate our method on four widely-used benchmark datasets, uncovering
significant gender disparities with a male-to-female ratio ranging from 4:1 to
6:1. These findings demonstrate the value of our methodology for bias
quantification in gendered languages and suggest its application in NLP,
contributing to the development of more equitable language technologies. | cs.CL | Leveraging Large Language Models to Measure Gender Bias in Gendered Languages |
2024-06-19T16:10:26Z | http://arxiv.org/pdf/2406.13663v2 | http://arxiv.org/abs/2406.13663v2 | Jirui Qi, Gabriele Sarti, Raquel Fernández, Arianna Bisazza | 2024-07-01T12:39:26Z | cs.CL, cs.AI, cs.LG | Ensuring the verifiability of model answers is a fundamental challenge for
retrieval-augmented generation (RAG) in the question answering (QA) domain.
Recently, self-citation prompting was proposed to make large language models
(LLMs) generate citations to supporting documents along with their answers.
However, self-citing LLMs often struggle to match the required format, refer to
non-existent sources, and fail to faithfully reflect LLMs' context usage
throughout the generation. In this work, we present MIRAGE --Model
Internals-based RAG Explanations -- a plug-and-play approach using model
internals for faithful answer attribution in RAG applications. MIRAGE detects
context-sensitive answer tokens and pairs them with retrieved documents
contributing to their prediction via saliency methods. We evaluate our proposed
approach on a multilingual extractive QA dataset, finding high agreement with
human answer attribution. On open-ended QA, MIRAGE achieves citation quality
and efficiency comparable to self-citation while also allowing for a
finer-grained control of attribution parameters. Our qualitative evaluation
highlights the faithfulness of MIRAGE's attributions and underscores the
promising application of model internals for RAG answer attribution. | cs.CL | Model Internals-based Answer Attribution for Trustworthy Retrieval-Augmented Generation |
2024-06-19T16:09:58Z | http://arxiv.org/pdf/2406.13662v1 | http://arxiv.org/abs/2406.13662v1 | Yue Huang, Jingyu Tang, Dongping Chen, Bingda Tang, Yao Wan, Lichao Sun, Xiangliang Zhang | 2024-06-19T16:09:58Z | cs.CL | Recently, Large Language Models (LLMs) have garnered significant attention
for their exceptional natural language processing capabilities. However,
concerns about their trustworthiness remain unresolved, particularly in
addressing "jailbreaking" attacks on aligned LLMs. Previous research
predominantly relies on scenarios with white-box LLMs or specific and fixed
prompt templates, which are often impractical and lack broad applicability. In
this paper, we introduce a straightforward and novel method, named
ObscurePrompt, for jailbreaking LLMs, inspired by the observed fragile
alignments in Out-of-Distribution (OOD) data. Specifically, we first formulate
the decision boundary in the jailbreaking process and then explore how obscure
text affects LLM's ethical decision boundary. ObscurePrompt starts with
constructing a base prompt that integrates well-known jailbreaking techniques.
Powerful LLMs are then utilized to obscure the original prompt through
iterative transformations, aiming to bolster the attack's robustness.
Comprehensive experiments show that our approach substantially improves upon
previous methods in terms of attack effectiveness, maintaining efficacy against
two prevalent defense mechanisms. We believe that our work can offer fresh
insights for future research on enhancing LLM alignment. | cs.CL | ObscurePrompt: Jailbreaking Large Language Models via Obscure Input |
2024-06-19T16:06:21Z | http://arxiv.org/pdf/2406.13660v1 | http://arxiv.org/abs/2406.13660v1 | Lily H. Zhang, Rajesh Ranganath, Arya Tafvizi | 2024-06-19T16:06:21Z | cs.CL, cs.AI | Generative models of language exhibit impressive capabilities but still place
non-negligible probability mass over undesirable outputs. In this work, we
address the task of updating a model to avoid unwanted outputs while minimally
changing model behavior otherwise, a challenge we refer to as a minimal
targeted update. We first formalize the notion of a minimal targeted update and
propose a method to achieve such updates using negative examples from a model's
generations. Our proposed Targeted Negative Training (TNT) results in updates
that keep the new distribution close to the original, unlike existing losses
for negative signal which push down probability but do not control what the
updated distribution will be. In experiments, we demonstrate that TNT yields a
better trade-off between reducing unwanted behavior and maintaining model
generation behavior than baselines, paving the way towards a modeling paradigm
based on iterative training updates that constrain models from generating
undesirable outputs while preserving their impressive capabilities. | cs.CL | Towards Minimal Targeted Updates of Language Models with Targeted Negative Training |
2024-06-19T15:28:29Z | http://arxiv.org/pdf/2406.13632v2 | http://arxiv.org/abs/2406.13632v2 | Arie Cattan, Alon Jacovi, Alex Fabrikant, Jonathan Herzig, Roee Aharoni, Hannah Rashkin, Dror Marcus, Avinatan Hassidim, Yossi Matias, Idan Szpektor, Avi Caciularu | 2024-06-23T07:19:22Z | cs.CL | Despite recent advancements in Large Language Models (LLMs), their
performance on tasks involving long contexts remains sub-optimal. In-Context
Learning (ICL) with few-shot examples may be an appealing solution to enhance
LLM performance in this scenario; However, naively adding ICL examples with
long context introduces challenges, including substantial token overhead added
for each few-shot example and context mismatch between the demonstrations and
the target query. In this work, we propose to automatically generate few-shot
examples for long context QA tasks by recycling contexts. Specifically, given a
long input context (1-3k tokens) and a query, we generate additional
query-output pairs from the given context as few-shot examples, while
introducing the context only once. This ensures that the demonstrations are
leveraging the same context as the target query while only adding a small
number of tokens to the prompt. We further enhance each demonstration by
instructing the model to explicitly identify the relevant paragraphs before the
answer, which improves performance while providing fine-grained attribution to
the answer source. We apply our method on multiple LLMs and obtain substantial
improvements (+23\% on average across models) on various QA datasets with long
context, especially when the answer lies within the middle of the context.
Surprisingly, despite introducing only single-hop ICL examples, LLMs also
successfully generalize to multi-hop long-context QA using our approach. | cs.CL | Can Few-shot Work in Long-Context? Recycling the Context to Generate Demonstrations |
2024-06-19T15:25:29Z | http://arxiv.org/pdf/2406.13629v1 | http://arxiv.org/abs/2406.13629v1 | Zhepei Wei, Wei-Lin Chen, Yu Meng | 2024-06-19T15:25:29Z | cs.CL, cs.LG | Retrieval-augmented generation (RAG) has shown promising potential to enhance
the accuracy and factuality of language models (LMs). However, imperfect
retrievers or noisy corpora can introduce misleading or even erroneous
information to the retrieved contents, posing a significant challenge to the
generation quality. Existing RAG methods typically address this challenge by
directly predicting final answers despite potentially noisy inputs, resulting
in an implicit denoising process that is difficult to interpret and verify. On
the other hand, the acquisition of explicit denoising supervision is often
costly, involving significant human efforts. In this work, we propose
InstructRAG, where LMs explicitly learn the denoising process through
self-synthesized rationales -- First, we instruct the LM to explain how the
ground-truth answer is derived from retrieved documents. Then, these rationales
can be used either as demonstrations for in-context learning of explicit
denoising or as supervised fine-tuning data to train the model. Compared to
standard RAG approaches, InstructRAG requires no additional supervision, allows
for easier verification of the predicted answers, and effectively improves
generation accuracy. Experiments show InstructRAG consistently outperforms
existing RAG methods in both training-free and trainable scenarios, achieving a
relative improvement of 8.3% over the best baseline method on average across
five knowledge-intensive benchmarks. Extensive analysis indicates that
InstructRAG scales well with increased numbers of retrieved documents and
consistently exhibits robust denoising ability even in out-of-domain datasets,
demonstrating strong generalizability. | cs.CL | InstructRAG: Instructing Retrieval-Augmented Generation with Explicit Denoising |
2024-06-19T15:20:19Z | http://arxiv.org/pdf/2406.13626v1 | http://arxiv.org/abs/2406.13626v1 | Kangtong Mo, Wenyan Liu, Xuanzhen Xu, Chang Yu, Yuelin Zou, Fangqing Xia | 2024-06-19T15:20:19Z | cs.CL, cs.AI | In this study, we explore the application of sentiment analysis on financial
news headlines to understand investor sentiment. By leveraging Natural Language
Processing (NLP) and Large Language Models (LLM), we analyze sentiment from the
perspective of retail investors. The FinancialPhraseBank dataset, which
contains categorized sentiments of financial news headlines, serves as the
basis for our analysis. We fine-tuned several models, including
distilbert-base-uncased, Llama, and gemma-7b, to evaluate their effectiveness
in sentiment classification. Our experiments demonstrate that the fine-tuned
gemma-7b model outperforms others, achieving the highest precision, recall, and
F1 score. Specifically, the gemma-7b model showed significant improvements in
accuracy after fine-tuning, indicating its robustness in capturing the nuances
of financial sentiment. This model can be instrumental in providing market
insights, risk management, and aiding investment decisions by accurately
predicting the sentiment of financial news. The results highlight the potential
of advanced LLMs in transforming how we analyze and interpret financial
information, offering a powerful tool for stakeholders in the financial
industry. | cs.CL | Fine-Tuning Gemma-7B for Enhanced Sentiment Analysis of Financial News Headlines |
2024-06-19T15:17:10Z | http://arxiv.org/pdf/2406.13621v1 | http://arxiv.org/abs/2406.13621v1 | Guy Yariv, Idan Schwartz, Yossi Adi, Sagie Benaim | 2024-06-19T15:17:10Z | cs.CL, cs.CV, cs.LG | Commonsense reasoning is fundamentally based on multimodal knowledge.
However, existing large language models (LLMs) are primarily trained using
textual data only, limiting their ability to incorporate essential visual
information. In contrast, Visual Language Models, which excel at
visually-oriented tasks, often fail at non-visual tasks such as basic
commonsense reasoning. This divergence highlights a critical challenge - the
integration of robust visual understanding with foundational text-based
language reasoning. To this end, we introduce a method aimed at enhancing LLMs'
visual commonsense. Specifically, our method generates multiple images based on
the input text prompt and integrates these into the model's decision-making
process by mixing their prediction probabilities. To facilitate multimodal
grounded language modeling, we employ a late-fusion layer that combines the
projected visual features with the output of a pre-trained LLM conditioned on
text only. This late-fusion layer enables predictions based on comprehensive
image-text knowledge as well as text only when this is required. We evaluate
our approach using several visual commonsense reasoning tasks together with
traditional NLP tasks, including common sense reasoning and reading
comprehension. Our experimental results demonstrate significant superiority
over existing baselines. When applied to recent state-of-the-art LLMs (e.g.,
Llama3), we observe improvements not only in visual common sense but also in
traditional NLP benchmarks. Code and models are available under
https://github.com/guyyariv/vLMIG. | cs.CL | Improving Visual Commonsense in Language Models via Multiple Image Generation |
2024-06-19T15:14:55Z | http://arxiv.org/pdf/2406.13618v1 | http://arxiv.org/abs/2406.13618v1 | Xiangfeng Wang, Zaiyi Chen, Zheyong Xie, Tong Xu, Yongyi He, Enhong Chen | 2024-06-19T15:14:55Z | cs.CL | With the rising popularity of Transformer-based large language models (LLMs),
reducing their high inference costs has become a significant research focus.
One effective approach is to compress the long input contexts. Existing methods
typically leverage the self-attention mechanism of the LLM itself for context
compression. While these methods have achieved notable results, the compression
process still involves quadratic time complexity, which limits their
applicability. To mitigate this limitation, we propose the In-Context Former
(IC-Former). Unlike previous methods, IC-Former does not depend on the target
LLMs. Instead, it leverages the cross-attention mechanism and a small number of
learnable digest tokens to directly condense information from the contextual
word embeddings. This approach significantly reduces inference time, which
achieves linear growth in time complexity within the compression range.
Experimental results indicate that our method requires only 1/32 of the
floating-point operations of the baseline during compression and improves
processing speed by 68 to 112 times while achieving over 90% of the baseline
performance on evaluation metrics. Overall, our model effectively reduces
compression costs and makes real-time compression scenarios feasible. | cs.CL | In-Context Former: Lightning-fast Compressing Context for Large Language Model |
2024-06-19T15:13:07Z | http://arxiv.org/pdf/2406.13617v1 | http://arxiv.org/abs/2406.13617v1 | Wenjie Li, Tianyu Sun, Kun Qian, Wenhong Wang | 2024-06-19T15:13:07Z | cs.CL, cs.AI | The advent of large language models (LLMs) has significantly advanced various
fields, including natural language processing and automated dialogue systems.
This paper explores the application of LLMs in psychological counseling,
addressing the increasing demand for mental health services. We present a
method for instruction tuning LLMs with specialized prompts to enhance their
performance in providing empathetic, relevant, and supportive responses. Our
approach involves developing a comprehensive dataset of counseling-specific
prompts, refining them through feedback from professional counselors, and
conducting rigorous evaluations using both automatic metrics and human
assessments. The results demonstrate that our instruction-tuned model
outperforms several baseline LLMs, highlighting its potential as a scalable and
accessible tool for mental health support. | cs.CL | Optimizing Psychological Counseling with Instruction-Tuned Large Language Models |
2024-06-19T14:12:05Z | http://arxiv.org/pdf/2406.13578v1 | http://arxiv.org/abs/2406.13578v1 | Han-Cheng Yu, Yu-An Shih, Kin-Man Law, Kai-Yu Hsieh, Yu-Chen Cheng, Hsin-Chih Ho, Zih-An Lin, Wen-Chuan Hsu, Yao-Chung Fan | 2024-06-19T14:12:05Z | cs.CL | In this paper, we tackle the task of distractor generation (DG) for
multiple-choice questions. Our study introduces two key designs. First, we
propose \textit{retrieval augmented pretraining}, which involves refining the
language model pretraining to align it more closely with the downstream task of
DG. Second, we explore the integration of knowledge graphs to enhance the
performance of DG. Through experiments with benchmarking datasets, we show that
our models significantly outperform the state-of-the-art results. Our
best-performing model advances the F1@3 score from 14.80 to 16.47 in MCQ
dataset and from 15.92 to 16.50 in Sciq dataset. | cs.CL | Enhancing Distractor Generation for Multiple-Choice Questions with Retrieval Augmented Pretraining and Knowledge Graph Integration |
2024-06-19T13:48:19Z | http://arxiv.org/pdf/2406.13560v1 | http://arxiv.org/abs/2406.13560v1 | Jindřich Libovický, Jindřich Helcl | 2024-06-19T13:48:19Z | cs.CL | We present three innovations in tokenization and subword segmentation. First,
we propose to use unsupervised morphological analysis with Morfessor as
pre-tokenization. Second, we present an algebraic method for obtaining subword
embeddings grounded in a word embedding space. Based on that, we design a novel
subword segmentation algorithm that uses the embeddings, ensuring that the
procedure considers lexical meaning. Third, we introduce an efficient
segmentation algorithm based on a subword bigram model that can be initialized
with the lexically aware segmentation method to avoid using Morfessor and large
embedding tables at inference time. We evaluate the proposed approaches using
two intrinsic metrics and measure their performance on two downstream tasks:
part-of-speech tagging and machine translation. Our experiments show
significant improvements in the morphological plausibility of the segmentation
when evaluated using segmentation precision on morpheme boundaries and improved
R\'enyi efficiency in 8 languages. Although the proposed tokenization methods
do not have a large impact on automatic translation quality, we observe
consistent performance gains in the arguably more morphological task of
part-of-speech tagging. | cs.CL | Lexically Grounded Subword Segmentation |
2024-06-19T13:45:21Z | http://arxiv.org/pdf/2406.13556v1 | http://arxiv.org/abs/2406.13556v1 | Yi Zhou, Danushka Bollegala, Jose Camacho-Collados | 2024-06-19T13:45:21Z | cs.CL | Social biases such as gender or racial biases have been reported in language
models (LMs), including Masked Language Models (MLMs). Given that MLMs are
continuously trained with increasing amounts of additional data collected over
time, an important yet unanswered question is how the social biases encoded
with MLMs vary over time. In particular, the number of social media users
continues to grow at an exponential rate, and it is a valid concern for the
MLMs trained specifically on social media data whether their social biases (if
any) would also amplify over time. To empirically analyse this problem, we use
a series of MLMs pretrained on chronologically ordered temporal snapshots of
corpora. Our analysis reveals that, although social biases are present in all
MLMs, most types of social bias remain relatively stable over time (with a few
exceptions). To further understand the mechanisms that influence social biases
in MLMs, we analyse the temporal corpora used to train the MLMs. Our findings
show that some demographic groups, such as male, obtain higher preference over
the other, such as female on the training corpora constantly. | cs.CL | Evaluating Short-Term Temporal Fluctuations of Social Biases in Social Media Data and Masked Language Models |
2024-06-19T13:44:56Z | http://arxiv.org/pdf/2406.13555v1 | http://arxiv.org/abs/2406.13555v1 | Minchong Li, Feng Zhou, Xiaohui Song | 2024-06-19T13:44:56Z | cs.CL, cs.AI | In recent years, large language models (LLMs) have shown exceptional
capabilities across various natural language processing (NLP) tasks. However,
such impressive performance often comes with the trade-off of an increased
parameter size, posing significant challenges for widespread deployment.
Knowledge distillation (KD) provides a solution by transferring knowledge from
a large teacher model to a smaller student model. In this paper, we explore the
task-specific distillation of LLMs at the logit level. Our investigation
reveals that the logits of fine-tuned LLMs exhibit a more extreme long-tail
distribution than those from vision models, with hidden "noise" in the long
tail affecting distillation performance. Furthermore, existing logits
distillation methods often struggle to effectively utilize the internal ranking
information from the logits. To address these, we propose the Bi-directional
Logits Difference (BiLD) loss. The BiLD loss filters out the long-tail noise by
utilizing only top-$k$ teacher and student logits, and leverages the internal
logits ranking information by constructing logits differences. To evaluate BiLD
loss, we conduct comprehensive experiments on 13 datasets using two types of
LLMs. Our results show that the BiLD loss, with only the top-8 logits,
outperforms supervised fine-tuning (SFT), vanilla KL loss, and five other
distillation methods from both NLP and CV fields. | cs.CL | BiLD: Bi-directional Logits Difference Loss for Large Language Model Distillation |
2024-06-19T13:43:27Z | http://arxiv.org/pdf/2406.13553v1 | http://arxiv.org/abs/2406.13553v1 | Mateusz Grzyb, Mateusz Krzyziński, Bartłomiej Sobieski, Mikołaj Spytek, Bartosz Pieliński, Daniel Dan, Anna Wróblewska | 2024-06-19T13:43:27Z | cs.CL, cs.IR | This project explores the application of Natural Language Processing (NLP)
techniques to analyse United Nations General Assembly (UNGA) speeches. Using
NLP allows for the efficient processing and analysis of large volumes of
textual data, enabling the extraction of semantic patterns, sentiment analysis,
and topic modelling. Our goal is to deliver a comprehensive dataset and a tool
(interface with descriptive statistics and automatically extracted topics) from
which political scientists can derive insights into international relations and
have the opportunity to have a nuanced understanding of global diplomatic
discourse. | cs.CL | Mining United Nations General Assembly Debates |
2024-06-19T13:38:34Z | http://arxiv.org/pdf/2406.13551v1 | http://arxiv.org/abs/2406.13551v1 | Omkar Dige, Diljot Singh, Tsz Fung Yau, Qixuan Zhang, Borna Bolandraftar, Xiaodan Zhu, Faiza Khan Khattak | 2024-06-19T13:38:34Z | cs.CL, cs.AI | Mitigating bias in language models (LMs) has become a critical problem due to
the widespread deployment of LMs. Numerous approaches revolve around data
pre-processing and fine-tuning of language models, tasks that can be both
time-consuming and computationally demanding. Consequently, there is a growing
interest in machine unlearning techniques given their capacity to induce the
forgetting of undesired behaviors of the existing pre-trained or fine-tuned
models with lower computational cost. In this work, we explore two unlearning
methods, (1) Partitioned Contrastive Gradient Unlearning (PCGU) applied on
decoder models and (2) Negation via Task Vector, to reduce social biases in
state-of-the-art and open-source LMs such as LLaMA-2 and OPT. We also implement
distributed PCGU for large models. It is empirically shown, through
quantitative and qualitative analyses, that negation via Task Vector method
outperforms PCGU in debiasing with minimum deterioration in performance and
perplexity of the models. On LLaMA-27B, negation via Task Vector reduces the
bias score by 11.8% | cs.CL | Mitigating Social Biases in Language Models through Unlearning |
2024-06-19T13:29:53Z | http://arxiv.org/pdf/2406.13542v1 | http://arxiv.org/abs/2406.13542v1 | Guanting Dong, Keming Lu, Chengpeng Li, Tingyu Xia, Bowen Yu, Chang Zhou, Jingren Zhou | 2024-06-19T13:29:53Z | cs.CL, cs.AI, cs.LG | One core capability of large language models (LLMs) is to follow natural
language instructions. However, the issue of automatically constructing
high-quality training data to enhance the complex instruction-following
abilities of LLMs without manual annotation remains unresolved. In this paper,
we introduce AutoIF, the first scalable and reliable method for automatically
generating instruction-following training data. AutoIF transforms the
validation of instruction-following data quality into code verification,
requiring LLMs to generate instructions, the corresponding code to check the
correctness of the instruction responses, and unit test samples to verify the
code's correctness. Then, execution feedback-based rejection sampling can
generate data for Supervised Fine-Tuning (SFT) and Reinforcement Learning from
Human Feedback (RLHF) training. AutoIF achieves significant improvements across
three training algorithms, SFT, Offline DPO, and Online DPO, when applied to
the top open-source LLMs, Qwen2 and LLaMA3, in self-alignment and
strong-to-weak distillation settings. Our code is publicly available at
https://github.com/QwenLM/AutoIF. | cs.CL | Self-play with Execution Feedback: Improving Instruction-following Capabilities of Large Language Models |
2024-06-19T12:47:34Z | http://arxiv.org/pdf/2406.13502v1 | http://arxiv.org/abs/2406.13502v1 | Jean Seo, Minha Kang, Sungjoo Byun, Sangah Lee | 2024-06-19T12:47:34Z | cs.CL, cs.SD, eess.AS | This study addresses the widening gap in Automatic Speech Recognition (ASR)
research between high resource and extremely low resource languages, with a
particular focus on Manchu, a critically endangered language. Manchu
exemplifies the challenges faced by marginalized linguistic communities in
accessing state-of-the-art technologies. In a pioneering effort, we introduce
the first-ever Manchu ASR model ManWav, leveraging Wav2Vec2-XLSR-53. The
results of the first Manchu ASR is promising, especially when trained with our
augmented data. Wav2Vec2-XLSR-53 fine-tuned with augmented data demonstrates a
0.02 drop in CER and 0.13 drop in WER compared to the same base model
fine-tuned with original data. | cs.CL | ManWav: The First Manchu ASR Model |
2024-06-19T11:57:42Z | http://arxiv.org/pdf/2406.13476v3 | http://arxiv.org/abs/2406.13476v3 | Roman Koshkin, Katsuhito Sudoh, Satoshi Nakamura | 2024-06-25T04:45:19Z | cs.CL | The advent of transformers has fueled progress in machine translation. More
recently large language models (LLMs) have come to the spotlight thanks to
their generality and strong performance in a wide range of language tasks,
including translation. Here we show that open-source LLMs perform on par with
or better than some state-of-the-art baselines in simultaneous machine
translation (SiMT) tasks, zero-shot. We also demonstrate that injection of
minimal background information, which is easy with an LLM, brings further
performance gains, especially on challenging technical subject-matter. This
highlights LLMs' potential for building next generation of massively
multilingual, context-aware and terminologically accurate SiMT systems that
require no resource-intensive training or fine-tuning. | cs.CL | LLMs Are Zero-Shot Context-Aware Simultaneous Translators |
2024-06-19T11:50:09Z | http://arxiv.org/pdf/2406.13469v1 | http://arxiv.org/abs/2406.13469v1 | Dan Saattrup Nielsen, Kenneth Enevoldsen, Peter Schneider-Kamp | 2024-06-19T11:50:09Z | cs.CL, cs.AI, cs.LG, I.2.7 | This paper explores the performance of encoder and decoder language models on
multilingual Natural Language Understanding (NLU) tasks, with a broad focus on
Germanic languages. Building upon the ScandEval benchmark, which initially was
restricted to evaluating encoder models, we extend the evaluation framework to
include decoder models. We introduce a method for evaluating decoder models on
NLU tasks and apply it to the languages Danish, Swedish, Norwegian, Icelandic,
Faroese, German, Dutch, and English. Through a series of experiments and
analyses, we address key research questions regarding the comparative
performance of encoder and decoder models, the impact of NLU task types, and
the variation across language resources. Our findings reveal that decoder
models can achieve significantly better NLU performance than encoder models,
with nuances observed across different tasks and languages. Additionally, we
investigate the correlation between decoders and task performance via a UMAP
analysis, shedding light on the unique capabilities of decoder and encoder
models. This study contributes to a deeper understanding of language model
paradigms in NLU tasks and provides valuable insights for model selection and
evaluation in multilingual settings. | cs.CL | Encoder vs Decoder: Comparative Analysis of Encoder and Decoder Language Models on Multilingual NLU Tasks |
2024-06-19T11:09:16Z | http://arxiv.org/pdf/2406.13444v2 | http://arxiv.org/abs/2406.13444v2 | Xueqing Wu, Zongyu Lin, Songyan Zhao, Te-Lin Wu, Pan Lu, Nanyun Peng, Kai-Wei Chang | 2024-06-27T17:09:24Z | cs.CL, cs.CV | Visual programs are executable code generated by large language models to
address visual reasoning problems. They decompose complex questions into
multiple reasoning steps and invoke specialized models for each step to solve
the problems. However, these programs are prone to logic errors, with our
preliminary evaluation showing that 58% of the total errors are caused by
program logic errors. Debugging complex visual programs remains a major
bottleneck for visual reasoning. To address this, we introduce VDebugger, a
novel critic-refiner framework trained to localize and debug visual programs by
tracking execution step by step. VDebugger identifies and corrects program
errors leveraging detailed execution feedback, improving interpretability and
accuracy. The training data is generated through an automated pipeline that
injects errors into correct visual programs using a novel mask-best decoding
technique. Evaluations on six datasets demonstrate VDebugger's effectiveness,
showing performance improvements of up to 3.2% in downstream task accuracy.
Further studies show VDebugger's ability to generalize to unseen tasks,
bringing a notable improvement of 2.3% on the unseen COVR task. Code, data and
models are made publicly available at https://github.com/shirley-wu/vdebugger/ | cs.CL | VDebugger: Harnessing Execution Feedback for Debugging Visual Programs |
2024-06-19T11:08:56Z | http://arxiv.org/pdf/2406.13443v1 | http://arxiv.org/abs/2406.13443v1 | Muchen Yang, Moxin Li, Yongle Li, Zijun Chen, Chongming Gao, Junqi Zhang, Yangyang Li, Fuli Feng | 2024-06-19T11:08:56Z | cs.CL | Gradient-free prompt optimization methods have made significant strides in
enhancing the performance of closed-source Large Language Models (LLMs) across
a wide range of tasks. However, existing approaches make light of the
importance of high-quality prompt initialization and the identification of
effective optimization directions, thus resulting in substantial optimization
steps to obtain satisfactory performance. In this light, we aim to accelerate
prompt optimization process to tackle the challenge of low convergence rate. We
propose a dual-phase approach which starts with generating high-quality initial
prompts by adopting a well-designed meta-instruction to delve into
task-specific information, and iteratively optimize the prompts at the sentence
level, leveraging previous tuning experience to expand prompt candidates and
accept effective ones. Extensive experiments on eight datasets demonstrate the
effectiveness of our proposed method, achieving a consistent accuracy gain over
baselines with less than five optimization steps. | cs.CL | Dual-Phase Accelerated Prompt Optimization |
2024-06-19T10:59:48Z | http://arxiv.org/pdf/2406.13439v1 | http://arxiv.org/abs/2406.13439v1 | Sumanth Doddapaneni, Mohammed Safi Ur Rahman Khan, Sshubam Verma, Mitesh M. Khapra | 2024-06-19T10:59:48Z | cs.CL | Large Language Models (LLMs) are increasingly relied upon to evaluate text
outputs of other LLMs, thereby influencing leaderboards and development
decisions. However, concerns persist over the accuracy of these assessments and
the potential for misleading conclusions. In this work, we investigate the
effectiveness of LLMs as evaluators for text generation tasks. We propose FBI,
a novel framework designed to examine the proficiency of Evaluator LLMs in
assessing four critical abilities in other LLMs: factual accuracy, instruction
following, coherence in long-form writing, and reasoning proficiency. By
introducing targeted perturbations in answers generated by LLMs, that clearly
impact one of these key capabilities, we test whether an Evaluator LLM can
detect these quality drops. By creating a total of 2400 perturbed answers
covering 22 perturbation categories, we conduct a comprehensive study using
different evaluation strategies on five prominent LLMs commonly used as
evaluators in the literature. Our findings reveal significant shortcomings in
current Evaluator LLMs, which failed to identify quality drops in over 50\% of
cases on average. Single-answer and pairwise evaluations demonstrated notable
limitations, whereas reference-based evaluations showed comparatively better
performance. These results underscore the unreliable nature of current
Evaluator LLMs and advocate for cautious implementation in practical
applications. Code and data are available at https://github.com/AI4Bharat/FBI. | cs.CL | Finding Blind Spots in Evaluator LLMs with Interpretable Checklists |
2024-06-19T10:45:12Z | http://arxiv.org/pdf/2406.13431v2 | http://arxiv.org/abs/2406.13431v2 | Vrunda N. Sukhadia, Shammur Absar Chowdhury | 2024-06-24T15:31:59Z | cs.CL, cs.SD, eess.AS | Children's speech recognition is considered a low-resource task mainly due to
the lack of publicly available data. There are several reasons for such data
scarcity, including expensive data collection and annotation processes, and
data privacy, among others. Transforming speech signals into discrete tokens
that do not carry sensitive information but capture both linguistic and
acoustic information could be a solution for privacy concerns. In this study,
we investigate the integration of discrete speech tokens into children's speech
recognition systems as input without significantly degrading the ASR
performance. Additionally, we explored single-view and multi-view strategies
for creating these discrete labels. Furthermore, we tested the models for
generalization capabilities with unseen domain and nativity dataset. Results
reveal that the discrete token ASR for children achieves nearly equivalent
performance with an approximate 83% reduction in parameters. | cs.CL | Children's Speech Recognition through Discrete Token Enhancement |
2024-06-19T10:11:37Z | http://arxiv.org/pdf/2406.13415v1 | http://arxiv.org/abs/2406.13415v1 | Matéo Mahaut, Laura Aina, Paula Czarnowska, Momchil Hardalov, Thomas Müller, Lluís Màrquez | 2024-06-19T10:11:37Z | cs.CL, cs.LG | Large Language Models (LLMs) tend to be unreliable in the factuality of their
answers. To address this problem, NLP researchers have proposed a range of
techniques to estimate LLM's confidence over facts. However, due to the lack of
a systematic comparison, it is not clear how the different methods compare to
one another. To fill this gap, we present a survey and empirical comparison of
estimators of factual confidence. We define an experimental framework allowing
for fair comparison, covering both fact-verification and question answering.
Our experiments across a series of LLMs indicate that trained hidden-state
probes provide the most reliable confidence estimates, albeit at the expense of
requiring access to weights and training data. We also conduct a deeper
assessment of factual confidence by measuring the consistency of model behavior
under meaning-preserving variations in the input. We find that the confidence
of LLMs is often unstable across semantically equivalent inputs, suggesting
that there is much room for improvement of the stability of models' parametric
knowledge. Our code is available at
(https://github.com/amazon-science/factual-confidence-of-llms). | cs.CL | Factual Confidence of LLMs: on Reliability and Robustness of Current Estimators |
2024-06-19T09:57:19Z | http://arxiv.org/pdf/2406.13408v1 | http://arxiv.org/abs/2406.13408v1 | Jipeng Cen, Jiaxin Liu, Zhixu Li, Jingjing Wang | 2024-06-19T09:57:19Z | cs.CL | While fine-tuned large language models (LLMs) excel in generating
grammatically valid SQL in Text-to-SQL parsing, they often struggle to ensure
semantic accuracy in queries, leading to user confusion and diminished system
usability. To tackle this challenge, we introduce SQLFixAgent, an innovative
multi-agent collaborative framework designed for detecting and repairing
erroneous SQL. Our framework comprises a core agent, SQLRefiner, alongside two
auxiliary agents: SQLReviewer and QueryCrafter. The SQLReviewer agent employs
the rubber duck debugging method to identify potential semantic mismatches
between SQL statement and user query. If the error is detected, the
QueryCrafter agent generates multiple SQL statements as candidate repairs using
a fine-tuned SQLTool. Subsequently, leveraging similar repair retrieval and
failure memory reflexion, the SQLRefiner agent selects the most fitting SQL
statement from the candidates as the final repair. We evaluated our proposed
framework on five Text-to-SQL benchmarks. The experimental results show that
our method consistently enhances the performance of the baseline model,
specifically achieving an execution accuracy improvement of over 3\% on the
Bird benchmark. Our framework also has a higher token efficiency compared to
other advanced methods, making it more competitive. | cs.CL | SQLFixAgent: Towards Semantic-Accurate SQL Generation via Multi-Agent Collaboration |
2024-06-19T09:38:59Z | http://arxiv.org/pdf/2406.13397v1 | http://arxiv.org/abs/2406.13397v1 | Julian Schnitzler, Xanh Ho, Jiahao Huang, Florian Boudin, Saku Sugawara, Akiko Aizawa | 2024-06-19T09:38:59Z | cs.CL | Most existing multi-hop datasets are extractive answer datasets, where the
answers to the questions can be extracted directly from the provided context.
This often leads models to use heuristics or shortcuts instead of performing
true multi-hop reasoning. In this paper, we propose a new multi-hop dataset,
MoreHopQA, which shifts from extractive to generative answers. Our dataset is
created by utilizing three existing multi-hop datasets: HotpotQA,
2WikiMultihopQA, and MuSiQue. Instead of relying solely on factual reasoning,
we enhance the existing multi-hop questions by adding another layer of
questioning that involves one, two, or all three of the following types of
reasoning: commonsense, arithmetic, and symbolic. Our dataset is created
through a semi-automated process, resulting in a dataset with 1,118 samples
that have undergone human verification. We then use our dataset to evaluate
five different large language models: Mistral 7B, Gemma 7B, Llama 3 (8B and
70B), and GPT-4. We also design various cases to analyze the reasoning steps in
the question-answering process. Our results show that models perform well on
initial multi-hop questions but struggle with our extended questions,
indicating that our dataset is more challenging than previous ones. Our
analysis of question decomposition reveals that although models can correctly
answer questions, only a portion - 38.7% for GPT-4 and 33.4% for Llama3-70B -
achieve perfect reasoning, where all corresponding sub-questions are answered
correctly. Evaluation code and data are available at
https://github.com/Alab-NII/morehopqa | cs.CL | MoreHopQA: More Than Multi-hop Reasoning |
2024-06-19T09:23:53Z | http://arxiv.org/pdf/2406.13381v1 | http://arxiv.org/abs/2406.13381v1 | Xinming Hou, Mingming Yang, Wenxiang Jiao, Xing Wang, Zhaopeng Tu, Wayne Xin Zhao | 2024-06-19T09:23:53Z | cs.CL | Existing LLMs exhibit remarkable performance on various NLP tasks, but still
struggle with complex real-world tasks, even equipped with advanced strategies
like CoT and ReAct. In this work, we propose the CoAct framework, which
transfers the hierarchical planning and collaboration patterns in human society
to LLM systems. Specifically, our CoAct framework involves two agents: (1) A
global planning agent, to comprehend the problem scope, formulate macro-level
plans and provide detailed sub-task descriptions to local execution agents,
which serves as the initial rendition of a global plan. (2) A local execution
agent, to operate within the multi-tier task execution structure, focusing on
detailed execution and implementation of specific tasks within the global plan.
Experimental results on the WebArena benchmark show that CoAct can re-arrange
the process trajectory when facing failures, and achieves superior performance
over baseline methods on long-horizon web tasks. Code is available at
https://github.com/xmhou2002/CoAct. | cs.CL | CoAct: A Global-Local Hierarchy for Autonomous Agent Collaboration |
2024-06-19T09:16:14Z | http://arxiv.org/pdf/2406.13375v1 | http://arxiv.org/abs/2406.13375v1 | Yilong Xu, Jinhua Gao, Xiaoming Yu, Baolong Bi, Huawei Shen, Xueqi Cheng | 2024-06-19T09:16:14Z | cs.CL | Large Language Models (LLMs) can enhance the credibility and verifiability by
generating text with citations. However, existing tasks and evaluation methods
are predominantly limited to sentence-level statement, neglecting the
significance of positional fine-grained citations that can appear anywhere
within sentences. To facilitate further exploration of the fine-grained
citation generation, we propose ALiiCE, the first automatic evaluation
framework for this task. Our framework first parses the sentence claim into
atomic claims via dependency analysis and then calculates citation quality at
the atomic claim level. ALiiCE introduces three novel metrics for positional
fined-grained citation quality assessment, including positional fine-grained
citation recall and precision, and coefficient of variation of citation
positions. We evaluate the positional fine-grained citation generation
performance of several LLMs on two long-form QA datasets. Our experiments and
analyses demonstrate the effectiveness and reasonableness of ALiiCE. The
results also indicate that existing LLMs still struggle to provide positional
fine-grained citations. | cs.CL | ALiiCE: Evaluating Positional Fine-grained Citation Generation |
2024-06-19T09:09:11Z | http://arxiv.org/pdf/2406.13363v1 | http://arxiv.org/abs/2406.13363v1 | Ryoma Kumon, Daiki Matsuoka, Hitomi Yanaka | 2024-06-19T09:09:11Z | cs.CL | Compositional generalization refers to the ability to generalize to novel
combinations of previously observed words and syntactic structures. Since it is
regarded as a desired property of neural models, recent work has assessed
compositional generalization in machine translation as well as semantic
parsing. However, previous evaluations with machine translation have focused
mostly on lexical generalization (i.e., generalization to unseen combinations
of known words). Thus, it remains unclear to what extent models can translate
sentences that require structural generalization (i.e., generalization to
different sorts of syntactic structures). To address this question, we
construct SGET, a machine translation dataset covering various types of
compositional generalization with control of words and sentence structures. We
evaluate neural machine translation models on SGET and show that they struggle
more in structural generalization than in lexical generalization. We also find
different performance trends in semantic parsing and machine translation, which
indicates the importance of evaluations across various tasks. | cs.CL | Evaluating Structural Generalization in Neural Machine Translation |
2024-06-19T09:07:31Z | http://arxiv.org/pdf/2406.13362v1 | http://arxiv.org/abs/2406.13362v1 | Haowen Hou, Peigen Zeng, Fei Ma, Fei Richard Yu | 2024-06-19T09:07:31Z | cs.CV, cs.CL, cs.LG | Visual Language Models (VLMs) have rapidly progressed with the recent success
of large language models. However, there have been few attempts to incorporate
efficient linear Recurrent Neural Networks (RNNs) architectures into VLMs. In
this study, we introduce VisualRWKV, the first application of a linear RNN
model to multimodal learning tasks, leveraging the pre-trained RWKV language
model. We propose a data-dependent recurrence and sandwich prompts to enhance
our modeling capabilities, along with a 2D image scanning mechanism to enrich
the processing of visual sequences. Extensive experiments demonstrate that
VisualRWKV achieves competitive performance compared to Transformer-based
models like LLaVA-1.5 on various benchmarks. To facilitate further research and
analysis, we have made the checkpoints and the associated code publicly
accessible at the following GitHub repository:
\href{https://github.com/howard-hou/VisualRWKV}{https://github.com/howard-hou/VisualRWKV}. | cs.CV | VisualRWKV: Exploring Recurrent Neural Networks for Visual Language Models |
2024-06-19T09:06:24Z | http://arxiv.org/pdf/2406.13361v1 | http://arxiv.org/abs/2406.13361v1 | Zhuoran Li, Chunming Hu, Junfan Chen, Zhijun Chen, Xiaohui Guo, Richong Zhang | 2024-06-19T09:06:24Z | cs.CL, cs.LG | Code-switching is a data augmentation scheme mixing words from multiple
languages into source lingual text. It has achieved considerable generalization
performance of cross-lingual transfer tasks by aligning cross-lingual
contextual word representations. However, uncontrolled and over-replaced
code-switching would augment dirty samples to model training. In other words,
the excessive code-switching text samples will negatively hurt the models'
cross-lingual transferability. To this end, we propose a Progressive
Code-Switching (PCS) method to gradually generate moderately difficult
code-switching examples for the model to discriminate from easy to hard. The
idea is to incorporate progressively the preceding learned multilingual
knowledge using easier code-switching data to guide model optimization on
succeeding harder code-switching data. Specifically, we first design a
difficulty measurer to measure the impact of replacing each word in a sentence
based on the word relevance score. Then a code-switcher generates the
code-switching data of increasing difficulty via a controllable temperature
variable. In addition, a training scheduler decides when to sample harder
code-switching data for model training. Experiments show our model achieves
state-of-the-art results on three different zero-shot cross-lingual transfer
tasks across ten languages. | cs.CL | Improving Zero-Shot Cross-Lingual Transfer via Progressive Code-Switching |
2024-06-19T09:04:43Z | http://arxiv.org/pdf/2406.13357v1 | http://arxiv.org/abs/2406.13357v1 | Boyong Wu, Chao Yan, Haoran Pu | 2024-06-19T09:04:43Z | cs.CL, cs.SD, eess.AS | By leveraging the power of Large Language Models(LLMs) and speech foundation
models, state of the art speech-text bimodal works can achieve challenging
tasks like spoken translation(ST) and question answering(SQA) altogether with
much simpler architectures. In this paper, we utilize the capability of Whisper
encoder and pre-trained Yi-6B. Empirical results reveal that modal alignment
can be achieved with one layer module and hundred hours of speech-text
multitask corpus. We further swap the Yi-6B with human preferences aligned
version of Yi-6B-Chat during inference, and discover that the alignment
capability is applicable as well. In addition, the alignment subspace revealed
by singular value decomposition(SVD) also implies linear alignment subspace is
sparse, which leaves the possibility to concatenate other features like
voice-print or video to expand modality. | cs.CL | Transferable speech-to-text large language model alignment module |
Subsets and Splits