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
|
|---|---|---|---|---|---|---|---|---|
2024-05-11T04:23:53Z | http://arxiv.org/pdf/2405.06906v1 | http://arxiv.org/abs/2405.06906v1 | Guangyuan Jiang, Matthias Hofer, Jiayuan Mao, Lionel Wong, Joshua B. Tenenbaum, Roger P. Levy | 2024-05-11T04:23:53Z | cs.CL | One hallmark of human language is its combinatoriality -- reusing a
relatively small inventory of building blocks to create a far larger inventory
of increasingly complex structures. In this paper, we explore the idea that
combinatoriality in language reflects a human inductive bias toward
representational efficiency in symbol systems. We develop a computational
framework for discovering structure in a writing system. Built on top of
state-of-the-art library learning and program synthesis techniques, our
computational framework discovers known linguistic structures in the Chinese
writing system and reveals how the system evolves towards simplification under
pressures for representational efficiency. We demonstrate how a library
learning approach, utilizing learned abstractions and compression, may help
reveal the fundamental computational principles that underlie the creation of
combinatorial structures in human cognition, and offer broader insights into
the evolution of efficient communication systems. | cs.CL | Finding structure in logographic writing with library learning |
2024-05-11T03:56:23Z | http://arxiv.org/pdf/2405.08014v1 | http://arxiv.org/abs/2405.08014v1 | Jinwei Lin | 2024-05-11T03:56:23Z | cs.RO, cs.CL | Front-following is more technically difficult to implement than the other two
human following technologies, but front-following technology is more practical
and can be applied in more areas to solve more practical problems.
Front-following technology has many advantages not found in back-following and
side-by-side technologies. In this paper, we will discuss basic and significant
principles and general design idea of this technology. Besides, various of
novel and special useful methods will be presented and provided. We use enough
beautiful figures to display our novel design idea. Our research result is open
source in 2018, and this paper is just to expand the research result
propagation granularity. Abundant magic design idea are included in this paper,
more idea and analyzing can sear and see other paper naming with a start of
Robot Design System with Jinwei Lin, the only author of this series papers. | cs.RO | Robot Detection System 1: Front-Following |
2024-05-11T03:06:08Z | http://arxiv.org/pdf/2405.06890v1 | http://arxiv.org/abs/2405.06890v1 | Yong Guan, Xiaozhi Wang, Lei Hou, Juanzi Li, Jeff Pan, Jiaoyan Chen, Freddy Lecue | 2024-05-11T03:06:08Z | cs.CL, cs.AI | Event relation extraction (ERE) is a critical and fundamental challenge for
natural language processing. Existing work mainly focuses on directly modeling
the entire document, which cannot effectively handle long-range dependencies
and information redundancy. To address these issues, we propose a cluster-aware
compression method for improving event relation extraction (TacoERE), which
explores a compression-then-extraction paradigm. Specifically, we first
introduce document clustering for modeling event dependencies. It splits the
document into intra- and inter-clusters, where intra-clusters aim to enhance
the relations within the same cluster, while inter-clusters attempt to model
the related events at arbitrary distances. Secondly, we utilize cluster
summarization to simplify and highlight important text content of clusters for
mitigating information redundancy and event distance. We have conducted
extensive experiments on both pre-trained language models, such as RoBERTa, and
large language models, such as ChatGPT and GPT-4, on three ERE datasets, i.e.,
MAVEN-ERE, EventStoryLine and HiEve. Experimental results demonstrate that
TacoERE is an effective method for ERE. | cs.CL | TacoERE: Cluster-aware Compression for Event Relation Extraction |
2024-05-11T02:55:11Z | http://arxiv.org/pdf/2405.06886v1 | http://arxiv.org/abs/2405.06886v1 | Yong Guan, Dingxiao Liu, Jinchen Ma, Hao Peng, Xiaozhi Wang, Lei Hou, Ru Li | 2024-05-11T02:55:11Z | cs.IR, cs.AI, cs.CL | Generative document retrieval, an emerging paradigm in information retrieval,
learns to build connections between documents and identifiers within a single
model, garnering significant attention. However, there are still two
challenges: (1) neglecting inner-content correlation during document
representation; (2) lacking explicit semantic structure during identifier
construction. Nonetheless, events have enriched relations and well-defined
taxonomy, which could facilitate addressing the above two challenges. Inspired
by this, we propose Event GDR, an event-centric generative document retrieval
model, integrating event knowledge into this task. Specifically, we utilize an
exchange-then-reflection method based on multi-agents for event knowledge
extraction. For document representation, we employ events and relations to
model the document to guarantee the comprehensiveness and inner-content
correlation. For identifier construction, we map the events to well-defined
event taxonomy to construct the identifiers with explicit semantic structure.
Our method achieves significant improvement over the baselines on two datasets,
and also hopes to provide insights for future research. | cs.IR | Event GDR: Event-Centric Generative Document Retrieval |
2024-05-10T23:24:18Z | http://arxiv.org/pdf/2405.12999v1 | http://arxiv.org/abs/2405.12999v1 | Julius Heitkoetter, Michael Gerovitch, Laker Newhouse | 2024-05-10T23:24:18Z | cs.CL, cs.AI | The trustworthiness of highly capable language models is put at risk when
they are able to produce deceptive outputs. Moreover, when models are
vulnerable to deception it undermines reliability. In this paper, we introduce
a method to investigate complex, model-on-model deceptive scenarios. We create
a dataset of over 10,000 misleading explanations by asking Llama-2 7B, 13B,
70B, and GPT-3.5 to justify the wrong answer for questions in the MMLU. We find
that, when models read these explanations, they are all significantly deceived.
Worryingly, models of all capabilities are successful at misleading others,
while more capable models are only slightly better at resisting deception. We
recommend the development of techniques to detect and defend against deception. | cs.CL | An Assessment of Model-On-Model Deception |
2024-05-10T21:39:09Z | http://arxiv.org/pdf/2405.06818v1 | http://arxiv.org/abs/2405.06818v1 | Sheriff Issaka, Zhaoyi Zhang, Mihir Heda, Keyi Wang, Yinka Ajibola, Ryan DeMar, Xuefeng Du | 2024-05-10T21:39:09Z | cs.CL | Despite comprising one-third of global languages, African languages are
critically underrepresented in Artificial Intelligence (AI), threatening
linguistic diversity and cultural heritage. Ghanaian languages, in particular,
face an alarming decline, with documented extinction and several at risk. This
study pioneers a comprehensive survey of Natural Language Processing (NLP)
research focused on Ghanaian languages, identifying methodologies, datasets,
and techniques employed. Additionally, we create a detailed roadmap outlining
challenges, best practices, and future directions, aiming to improve
accessibility for researchers. This work serves as a foundational resource for
Ghanaian NLP research and underscores the critical need for integrating global
linguistic diversity into AI development. | cs.CL | The Ghanaian NLP Landscape: A First Look |
2024-05-10T20:45:40Z | http://arxiv.org/pdf/2405.06808v1 | http://arxiv.org/abs/2405.06808v1 | Zhiyu Cao, Zachary Feinstein | 2024-05-10T20:45:40Z | q-fin.RM, cs.AI, cs.CL | This study explores the innovative use of Large Language Models (LLMs) as
analytical tools for interpreting complex financial regulations. The primary
objective is to design effective prompts that guide LLMs in distilling verbose
and intricate regulatory texts, such as the Basel III capital requirement
regulations, into a concise mathematical framework that can be subsequently
translated into actionable code. This novel approach aims to streamline the
implementation of regulatory mandates within the financial reporting and risk
management systems of global banking institutions. A case study was conducted
to assess the performance of various LLMs, demonstrating that GPT-4 outperforms
other models in processing and collecting necessary information, as well as
executing mathematical calculations. The case study utilized numerical
simulations with asset holdings -- including fixed income, equities, currency
pairs, and commodities -- to demonstrate how LLMs can effectively implement the
Basel III capital adequacy requirements. | q-fin.RM | Large Language Model in Financial Regulatory Interpretation |
2024-05-10T20:45:34Z | http://arxiv.org/pdf/2405.06807v1 | http://arxiv.org/abs/2405.06807v1 | Ngoc Phuoc An Vo, Brent Paulovicks, Vadim Sheinin | 2024-05-10T20:45:34Z | cs.CL, cs.SE | Given recent advancement of Large Language Models (LLMs), the task of
translating from natural language prompts to different programming languages
(code generation) attracts immense attention for wide application in different
domains. Specially code generation for Bash (NL2Bash) is widely used to
generate Bash scripts for automating different tasks, such as performance
monitoring, compilation, system administration, system diagnostics, etc.
Besides code generation, validating synthetic code is critical before using
them for any application. Different methods for code validation are proposed,
both direct (execution evaluation) and indirect validations (i.e. exact/partial
match, BLEU score). Among these, Execution-based Evaluation (EE) can validate
the predicted code by comparing the execution output of model prediction and
expected output in system. However, designing and implementing such an
execution-based evaluation system for NL2Bash is not a trivial task. In this
paper, we present a machinery for execution-based evaluation for NL2Bash. We
create a set of 50 prompts to evaluate some popular LLMs for NL2Bash. We also
analyze several advantages and challenges of EE such as syntactically different
yet semantically equivalent Bash scripts generated by different LLMs, or
syntactically correct but semantically incorrect Bash scripts, and how we
capture and process them correctly. | cs.CL | Tackling Execution-Based Evaluation for NL2Bash |
2024-05-10T20:29:25Z | http://arxiv.org/pdf/2405.06802v1 | http://arxiv.org/abs/2405.06802v1 | Raul Salles de Padua, Imran Qureshi | 2024-05-10T20:29:25Z | cs.CL, cs.AI | Patient hand-off and triage are two fundamental problems in health care.
Often doctors must painstakingly summarize complex findings to efficiently
communicate with specialists and quickly make decisions on which patients have
the most urgent cases. In pursuit of these challenges, we present (1) a model
with state-of-art radiology report summarization performance using (2) a novel
method for augmenting medical data, and (3) an analysis of the model
limitations and radiology knowledge gain. We also provide a data processing
pipeline for future models developed on the the MIMIC CXR dataset. Our best
performing model was a fine-tuned BERT-to-BERT encoder-decoder with 58.75/100
ROUGE-L F1, which outperformed specialized checkpoints with more sophisticated
attention mechanisms. We investigate these aspects in this work. | cs.CL | Summarizing Radiology Reports Findings into Impressions |
2024-05-10T20:23:46Z | http://arxiv.org/pdf/2405.06800v2 | http://arxiv.org/abs/2405.06800v2 | Rohan Ajwani, Shashidhar Reddy Javaji, Frank Rudzicz, Zining Zhu | 2024-05-29T15:18:59Z | cs.CL | Large Language Models (LLMs) are becoming vital tools that help us solve and
understand complex problems by acting as digital assistants. LLMs can generate
convincing explanations, even when only given the inputs and outputs of these
problems, i.e., in a ``black-box'' approach. However, our research uncovers a
hidden risk tied to this approach, which we call *adversarial helpfulness*.
This happens when an LLM's explanations make a wrong answer look right,
potentially leading people to trust incorrect solutions. In this paper, we show
that this issue affects not just humans, but also LLM evaluators. Digging
deeper, we identify and examine key persuasive strategies employed by LLMs. Our
findings reveal that these models employ strategies such as reframing the
questions, expressing an elevated level of confidence, and cherry-picking
evidence to paint misleading answers in a credible light. To examine if LLMs
are able to navigate complex-structured knowledge when generating adversarially
helpful explanations, we create a special task based on navigating through
graphs. Most LLMs are not able to find alternative paths along simple graphs,
indicating that their misleading explanations aren't produced by only logical
deductions using complex knowledge. These findings shed light on the
limitations of the black-box explanation setting and allow us to provide advice
on the safe usage of LLMs. | cs.CL | LLM-Generated Black-box Explanations Can Be Adversarially Helpful |
2024-05-10T18:57:35Z | http://arxiv.org/pdf/2405.06772v1 | http://arxiv.org/abs/2405.06772v1 | Urjitkumar Patel, Fang-Chun Yeh, Chinmay Gondhalekar | 2024-05-10T18:57:35Z | cs.CR, cs.AI, cs.CL, 68T50, 68T07 (Primary) 03B65, 91F20 (Secondary), I.2.7; I.2.1; I.5.1; I.5.2; I.5.4; H.3.3 | In today's digital landscape, where cyber attacks have become the norm, the
detection of cyber attacks and threats is critically imperative across diverse
domains. Our research presents a new empirical framework for cyber threat
modeling, adept at parsing and categorizing cyber-related information from news
articles, enhancing real-time vigilance for market stakeholders. At the core of
this framework is a fine-tuned BERT model, which we call CANAL - Cyber Activity
News Alerting Language Model, tailored for cyber categorization using a novel
silver labeling approach powered by Random Forest. We benchmark CANAL against
larger, costlier LLMs, including GPT-4, LLaMA, and Zephyr, highlighting their
zero to few-shot learning in cyber news classification. CANAL demonstrates
superior performance by outperforming all other LLM counterparts in both
accuracy and cost-effectiveness. Furthermore, we introduce the Cyber Signal
Discovery module, a strategic component designed to efficiently detect emerging
cyber signals from news articles. Collectively, CANAL and Cyber Signal
Discovery module equip our framework to provide a robust and cost-effective
solution for businesses that require agile responses to cyber intelligence. | cs.CR | CANAL -- Cyber Activity News Alerting Language Model: Empirical Approach vs. Expensive LLM |
2024-05-10T18:28:00Z | http://arxiv.org/pdf/2405.06762v1 | http://arxiv.org/abs/2405.06762v1 | Jinwei Lin | 2024-05-10T18:28:00Z | cs.HC, cs.CL | LaTex coding is one of the main methods of writing an academic paper. When
writing a paper, abundant proper visual or graphic components will represent
more information volume than the textual data. However, most of the
implementation of LaTex graphic items are designed as static items that have
some weaknesses in representing more informative figures or tables with an
interactive reading experience. To address this problem, we propose LIVE, a
novel design methods idea to design interactive LaTex graphic items. To make a
lucid representation of the main idea of LIVE, we designed several novels
representing implementations that are interactive and enough explanation for
the basic level principles. Using LIVE can design more graphic items, which we
call the Gitems, and easily and automatically get the relationship of the
mutual application of a specific range of papers, which will add more vitality
and performance factors into writing of traditional papers especially the
review papers. For vividly representing the functions of LIVE, we use the
papers from NeRF as the example reference papers. The code of the
implementation project is open source. | cs.HC | LIVE: LaTex Interactive Visual Editing |
2024-05-10T18:24:55Z | http://arxiv.org/pdf/2405.06760v1 | http://arxiv.org/abs/2405.06760v1 | Arash Rasti Meymandi, Zahra Hosseini, Sina Davari, Abolfazl Moshiri, Shabnam Rahimi-Golkhandan, Khashayar Namdar, Nikta Feizi, Mohamad Tavakoli-Targhi, Farzad Khalvati | 2024-05-10T18:24:55Z | cs.CL, cs.AI | This study explores the integration of advanced Natural Language Processing
(NLP) and Artificial Intelligence (AI) techniques to analyze and interpret
Persian literature, focusing on the poetry of Forough Farrokhzad. Utilizing
computational methods, we aim to unveil thematic, stylistic, and linguistic
patterns in Persian poetry. Specifically, the study employs AI models including
transformer-based language models for clustering of the poems in an
unsupervised framework. This research underscores the potential of AI in
enhancing our understanding of Persian literary heritage, with Forough
Farrokhzad's work providing a comprehensive case study. This approach not only
contributes to the field of Persian Digital Humanities but also sets a
precedent for future research in Persian literary studies using computational
techniques. | cs.CL | Opportunities for Persian Digital Humanities Research with Artificial Intelligence Language Models; Case Study: Forough Farrokhzad |
2024-05-10T17:59:08Z | http://arxiv.org/pdf/2405.06640v1 | http://arxiv.org/abs/2405.06640v1 | Jean Mercat, Igor Vasiljevic, Sedrick Keh, Kushal Arora, Achal Dave, Adrien Gaidon, Thomas Kollar | 2024-05-10T17:59:08Z | cs.CL | Linear transformers have emerged as a subquadratic-time alternative to
softmax attention and have garnered significant interest due to their
fixed-size recurrent state that lowers inference cost. However, their original
formulation suffers from poor scaling and underperforms compute-matched
transformers. Recent linear models such as RWKV and Mamba have attempted to
address these shortcomings by proposing novel time-mixing and gating
architectures, but pre-training large language models requires significant data
and compute investments. Thus, the search for subquadratic architectures is
limited by the availability of compute and quality pre-training datasets. As a
cost-effective alternative to pre-training linear transformers, we propose
Scalable UPtraining for Recurrent Attention (SUPRA). We present a method to
uptrain existing large pre-trained transformers into Recurrent Neural Networks
(RNNs) with a modest compute budget. This allows us to leverage the strong
pre-training data and performance of existing transformer LLMs, while requiring
5% of the training cost. We find that our linearization technique leads to
competitive performance on standard benchmarks, but we identify persistent
in-context learning and long-context modeling shortfalls for even the largest
linear models. Our code and models can be found at
https://github.com/TRI-ML/linear_open_lm. | cs.CL | Linearizing Large Language Models |
2024-05-10T17:59:04Z | http://arxiv.org/pdf/2405.06639v1 | http://arxiv.org/abs/2405.06639v1 | Seungwook Han, Idan Shenfeld, Akash Srivastava, Yoon Kim, Pulkit Agrawal | 2024-05-10T17:59:04Z | cs.LG, cs.AI, cs.CL | Aligning Large Language Models (LLMs) to cater to different human
preferences, learning new skills, and unlearning harmful behavior is an
important problem. Search-based methods, such as Best-of-N or Monte-Carlo Tree
Search, are performant, but impractical for LLM adaptation due to their high
inference cost. On the other hand, using Reinforcement Learning (RL) for
adaptation is computationally efficient, but performs worse due to the
optimization challenges in co-training the value function and the policy. We
present a new framework for reward optimization, Value Augmented Sampling
(VAS), that can maximize different reward functions using data sampled from
only the initial, frozen LLM. VAS solves for the optimal reward-maximizing
policy without co-training the policy and the value function, making the
optimization stable, outperforming established baselines, such as PPO and DPO,
on standard benchmarks, and achieving comparable results to Best-of-128 with
lower inference cost. Unlike existing RL methods that require changing the
weights of the LLM, VAS does not require access to the weights of the
pre-trained LLM. Thus, it can even adapt LLMs (e.g., ChatGPT), which are
available only as APIs. In addition, our algorithm unlocks the new capability
of composing several rewards and controlling the extent of each one during
deployment time, paving the road ahead for the future of aligned, personalized
LLMs. | cs.LG | Value Augmented Sampling for Language Model Alignment and Personalization |
2024-05-10T17:53:05Z | http://arxiv.org/pdf/2405.06636v2 | http://arxiv.org/abs/2405.06636v2 | Khanh Nguyen, Dimosthenis Karatzas | 2024-05-22T11:01:22Z | cs.CV, cs.AI, cs.CL, cs.LG | An important handicap of document analysis research is that documents tend to
be copyrighted or contain private information, which prohibits their open
publication and the creation of centralised, large-scale document datasets.
Instead, documents are scattered in private data silos, making extensive
training over heterogeneous data a tedious task. In this work, we explore the
use of a federated learning (FL) scheme as a way to train a shared model on
decentralised private document data. We focus on the problem of Document VQA, a
task particularly suited to this approach, as the type of reasoning
capabilities required from the model can be quite different in diverse domains.
Enabling training over heterogeneous document datasets can thus substantially
enrich DocVQA models. We assemble existing DocVQA datasets from diverse domains
to reflect the data heterogeneity in real-world applications. We explore the
self-pretraining technique in this multi-modal setting, where the same data is
used for both pretraining and finetuning, making it relevant for privacy
preservation. We further propose combining self-pretraining with a Federated
DocVQA training method using centralized adaptive optimization that outperforms
the FedAvg baseline. With extensive experiments, we also present a
multi-faceted analysis on training DocVQA models with FL, which provides
insights for future research on this task. We show that our pretraining
strategies can effectively learn and scale up under federated training with
diverse DocVQA datasets and tuning hyperparameters is essential for practical
document tasks under federation. | cs.CV | Federated Document Visual Question Answering: A Pilot Study |
2024-05-10T17:51:35Z | http://arxiv.org/pdf/2405.06634v2 | http://arxiv.org/abs/2405.06634v2 | Evan M. Williams, Kathleen M. Carley | 2024-06-10T15:28:16Z | cs.CV, cs.AI, cs.CL | We evaluate the zero-shot ability of GPT-4 and LLaVa to perform simple Visual
Network Analysis (VNA) tasks on small-scale graphs. We evaluate the Vision
Language Models (VLMs) on 5 tasks related to three foundational network science
concepts: identifying nodes of maximal degree on a rendered graph, identifying
whether signed triads are balanced or unbalanced, and counting components. The
tasks are structured to be easy for a human who understands the underlying
graph theoretic concepts, and can all be solved by counting the appropriate
elements in graphs. We find that while GPT-4 consistently outperforms LLaVa,
both models struggle with every visual network analysis task we propose. We
publicly release the first benchmark for the evaluation of VLMs on foundational
VNA tasks. | cs.CV | Multimodal LLMs Struggle with Basic Visual Network Analysis: a VNA Benchmark |
2024-05-10T17:40:02Z | http://arxiv.org/pdf/2405.06626v1 | http://arxiv.org/abs/2405.06626v1 | Chakshu Moar, Michael Pellauer, Hyoukjun Kwon | 2024-05-10T17:40:02Z | cs.LG, cs.CL | Large language models (LLMs) have emerged and presented their general
problem-solving capabilities with one model. However, the model size has
increased dramatically with billions of parameters to enable such broad
problem-solving capabilities. In addition, due to the dominance of
matrix-matrix and matrix-vector multiplications in LLMs, the compute-to-model
size ratio is significantly lower than that of CNNs. This shift pushes LLMs
from a computation-bound regime to a memory-bound regime. Therefore, optimizing
the memory footprint and traffic is an important optimization direction for
LLMs today.
Model compression methods such as quantization and parameter pruning have
been actively explored for achieving the memory footprint and traffic
optimization. However, the accuracy-efficiency trade-off of rank pruning for
LLMs is not well-understood yet. Therefore, we characterize the
accuracy-efficiency trade-off of a low-rank decomposition method, specifically
Tucker decomposition, on recent language models, including an open-source LLM,
Llama 2.
We formalize the low-rank decomposition design space and show that the
decomposition design space is enormous (e.g., O($2^{37}$) for Llama2-7B). To
navigate such a vast design space, we formulate the design space and perform
thorough case studies of accuracy-efficiency trade-offs using six widely used
LLM benchmarks on BERT and Llama 2 models. Our results show that we can achieve
a 9\% model size reduction with minimal accuracy drops, which range from 4\%p
to 10\%p, depending on the difficulty of the benchmark, without any retraining
to recover accuracy after decomposition. The results show that low-rank
decomposition can be a promising direction for LLM-based applications that
require real-time service in scale (e.g., AI agent assist and real-time coding
assistant), where the latency is as important as the model accuracy. | cs.LG | Characterizing the Accuracy - Efficiency Trade-off of Low-rank Decomposition in Language Models |
2024-05-10T17:11:31Z | http://arxiv.org/pdf/2405.06604v1 | http://arxiv.org/abs/2405.06604v1 | Alexandros Vasileiou, Oliver Eberle | 2024-05-10T17:11:31Z | cs.CL, cs.LG | As Transformers have become state-of-the-art models for natural language
processing (NLP) tasks, the need to understand and explain their predictions is
increasingly apparent. Especially in unsupervised applications, such as
information retrieval tasks, similarity models built on top of foundation model
representations have been widely applied. However, their inner prediction
mechanisms have mostly remained opaque. Recent advances in explainable AI have
made it possible to mitigate these limitations by leveraging improved
explanations for Transformers through layer-wise relevance propagation (LRP).
Using BiLRP, an extension developed for computing second-order explanations in
bilinear similarity models, we investigate which feature interactions drive
similarity in NLP models. We validate the resulting explanations and
demonstrate their utility in three corpus-level use cases, analyzing
grammatical interactions, multilingual semantics, and biomedical text
retrieval. Our findings contribute to a deeper understanding of different
semantic similarity tasks and models, highlighting how novel explainable AI
methods enable in-depth analyses and corpus-level insights. | cs.CL | Explaining Text Similarity in Transformer Models |
2024-05-10T16:06:43Z | http://arxiv.org/pdf/2405.06563v1 | http://arxiv.org/abs/2405.06563v1 | Ilia Kuznetsov, Osama Mohammed Afzal, Koen Dercksen, Nils Dycke, Alexander Goldberg, Tom Hope, Dirk Hovy, Jonathan K. Kummerfeld, Anne Lauscher, Kevin Leyton-Brown, Sheng Lu, Mausam, Margot Mieskes, Aurélie Névéol, Danish Pruthi, Lizhen Qu, Roy Schwartz, Noah A. Smith, Thamar Solorio, Jingyan Wang, Xiaodan Zhu, Anna Rogers, Nihar B. Shah, Iryna Gurevych | 2024-05-10T16:06:43Z | cs.CL | The number of scientific articles produced every year is growing rapidly.
Providing quality control over them is crucial for scientists and, ultimately,
for the public good. In modern science, this process is largely delegated to
peer review -- a distributed procedure in which each submission is evaluated by
several independent experts in the field. Peer review is widely used, yet it is
hard, time-consuming, and prone to error. Since the artifacts involved in peer
review -- manuscripts, reviews, discussions -- are largely text-based, Natural
Language Processing has great potential to improve reviewing. As the emergence
of large language models (LLMs) has enabled NLP assistance for many new tasks,
the discussion on machine-assisted peer review is picking up the pace. Yet,
where exactly is help needed, where can NLP help, and where should it stand
aside? The goal of our paper is to provide a foundation for the future efforts
in NLP for peer-reviewing assistance. We discuss peer review as a general
process, exemplified by reviewing at AI conferences. We detail each step of the
process from manuscript submission to camera-ready revision, and discuss the
associated challenges and opportunities for NLP assistance, illustrated by
existing work. We then turn to the big challenges in NLP for peer review as a
whole, including data acquisition and licensing, operationalization and
experimentation, and ethical issues. To help consolidate community efforts, we
create a companion repository that aggregates key datasets pertaining to peer
review. Finally, we issue a detailed call for action for the scientific
community, NLP and AI researchers, policymakers, and funding bodies to help
bring the research in NLP for peer review forward. We hope that our work will
help set the agenda for research in machine-assisted scientific quality control
in the age of AI, within the NLP community and beyond. | cs.CL | What Can Natural Language Processing Do for Peer Review? |
2024-05-10T15:49:01Z | http://arxiv.org/pdf/2405.06551v1 | http://arxiv.org/abs/2405.06551v1 | Piyush Kumar Garg, Roshni Chakraborty, Sourav Kumar Dandapat | 2024-05-10T15:49:01Z | cs.CL, cs.SI | Online social media platforms, such as Twitter, provide valuable information
during disaster events. Existing tweet disaster summarization approaches
provide a summary of these events to aid government agencies, humanitarian
organizations, etc., to ensure effective disaster response. In the literature,
there are two types of approaches for disaster summarization, namely,
supervised and unsupervised approaches. Although supervised approaches are
typically more effective, they necessitate a sizable number of disaster event
summaries for testing and training. However, there is a lack of good number of
disaster summary datasets for training and evaluation. This motivates us to add
more datasets to make supervised learning approaches more efficient. In this
paper, we present ADSumm, which adds annotated ground-truth summaries for eight
disaster events which consist of both natural and man-made disaster events
belonging to seven different countries. Our experimental analysis shows that
the newly added datasets improve the performance of the supervised
summarization approaches by 8-28% in terms of ROUGE-N F1-score. Moreover, in
newly annotated dataset, we have added a category label for each input tweet
which helps to ensure good coverage from different categories in summary.
Additionally, we have added two other features relevance label and key-phrase,
which provide information about the quality of a tweet and explanation about
the inclusion of the tweet into summary, respectively. For ground-truth summary
creation, we provide the annotation procedure adapted in detail, which has not
been described in existing literature. Experimental analysis shows the quality
of ground-truth summary is very good with Coverage, Relevance and Diversity. | cs.CL | ADSumm: Annotated Ground-truth Summary Datasets for Disaster Tweet Summarization |
2024-05-10T15:46:05Z | http://arxiv.org/pdf/2405.06549v1 | http://arxiv.org/abs/2405.06549v1 | Garett Ordway, Vic Patrangenaru | 2024-05-10T15:46:05Z | stat.AP, cs.CL | Communication plays a vital role in human interaction. Studying language is a
worthwhile task and more recently has become quantitative in nature with
developments of fields like quantitative comparative linguistics and
lexicostatistics. With respect to the authors own native languages, the
ancestry of the English language and the Latin alphabet are of the primary
interest. The Indo-European Tree traces many modern languages back to the
Proto-Indo-European root. Swadesh's cognates played a large role in developing
that historical perspective where some of the primary branches are Germanic,
Celtic, Italic, and Balto-Slavic. This paper will use data analysis on open
books where the simplest singular space is the 3-spider - a union T3 of three
rays with their endpoints glued at a point 0 - which can represent these tree
spaces for language clustering. These trees are built using a single linkage
method for clustering based on distances between samples from languages which
use the Latin Script. Taking three languages at a time, the barycenter is
determined. Some initial results have found both non-sticky and sticky sample
means. If the mean exhibits non-sticky properties, then one language may come
from a different ancestor than the other two. If the mean is considered sticky,
then the languages may share a common ancestor or all languages may have
different ancestry. | stat.AP | Sampling the Swadesh List to Identify Similar Languages with Tree Spaces |
2024-05-10T15:40:50Z | http://arxiv.org/pdf/2405.06545v1 | http://arxiv.org/abs/2405.06545v1 | Mengjia Niu, Hao Li, Jie Shi, Hamed Haddadi, Fan Mo | 2024-05-10T15:40:50Z | cs.CL, cs.LG, I.2.7; H.3.3 | Large language models (LLMs) have demonstrated remarkable capabilities across
various domains, although their susceptibility to hallucination poses
significant challenges for their deployment in critical areas such as
healthcare. To address this issue, retrieving relevant facts from knowledge
graphs (KGs) is considered a promising method. Existing KG-augmented approaches
tend to be resource-intensive, requiring multiple rounds of retrieval and
verification for each factoid, which impedes their application in real-world
scenarios.
In this study, we propose Self-Refinement-Enhanced Knowledge Graph Retrieval
(Re-KGR) to augment the factuality of LLMs' responses with less retrieval
efforts in the medical field. Our approach leverages the attribution of
next-token predictive probability distributions across different tokens, and
various model layers to primarily identify tokens with a high potential for
hallucination, reducing verification rounds by refining knowledge triples
associated with these tokens. Moreover, we rectify inaccurate content using
retrieved knowledge in the post-processing stage, which improves the
truthfulness of generated responses. Experimental results on a medical dataset
demonstrate that our approach can enhance the factual capability of LLMs across
various foundational models as evidenced by the highest scores on truthfulness. | cs.CL | Mitigating Hallucinations in Large Language Models via Self-Refinement-Enhanced Knowledge Retrieval |
2024-05-10T15:36:56Z | http://arxiv.org/pdf/2405.06541v1 | http://arxiv.org/abs/2405.06541v1 | Piyush Kumar Garg, Roshni Chakraborty, Sourav Kumar Dandapat | 2024-05-10T15:36:56Z | cs.CL, cs.SI | The abundance of situational information on Twitter poses a challenge for
users to manually discern vital and relevant information during disasters. A
concise and human-interpretable overview of this information helps
decision-makers in implementing efficient and quick disaster response. Existing
abstractive summarization approaches can be categorized as sentence-based or
key-phrase-based approaches. This paper focuses on sentence-based approach,
which is typically implemented as a dual-phase procedure in literature. The
initial phase, known as the extractive phase, involves identifying the most
relevant tweets. The subsequent phase, referred to as the abstractive phase,
entails generating a more human-interpretable summary. In this study, we adopt
the methodology from prior research for the extractive phase. For the
abstractive phase of summarization, most existing approaches employ deep
learning-based frameworks, which can either be pre-trained or require training
from scratch. However, to achieve the appropriate level of performance, it is
imperative to have substantial training data for both methods, which is not
readily available. This work presents an Abstractive Tweet Summarizer (ATSumm)
that effectively addresses the issue of data sparsity by using auxiliary
information. We introduced the Auxiliary Pointer Generator Network (AuxPGN)
model, which utilizes a unique attention mechanism called Key-phrase attention.
This attention mechanism incorporates auxiliary information in the form of
key-phrases and their corresponding importance scores from the input tweets. We
evaluate the proposed approach by comparing it with 10 state-of-the-art
approaches across 13 disaster datasets. The evaluation results indicate that
ATSumm achieves superior performance compared to state-of-the-art approaches,
with improvement of 4-80% in ROUGE-N F1-score. | cs.CL | ATSumm: Auxiliary information enhanced approach for abstractive disaster Tweet Summarization with sparse training data |
2024-05-10T15:10:20Z | http://arxiv.org/pdf/2405.06524v1 | http://arxiv.org/abs/2405.06524v1 | Wenyu Huang, Guancheng Zhou, Mirella Lapata, Pavlos Vougiouklis, Sebastien Montella, Jeff Z. Pan | 2024-05-10T15:10:20Z | cs.CL | Although Large Language Models (LLMs) are effective in performing various NLP
tasks, they still struggle to handle tasks that require extensive, real-world
knowledge, especially when dealing with long-tail facts (facts related to
long-tail entities). This limitation highlights the need to supplement LLMs
with non-parametric knowledge. To address this issue, we analysed the effects
of different types of non-parametric knowledge, including textual passage and
knowledge graphs (KGs). Since LLMs have probably seen the majority of factual
question-answering datasets already, to facilitate our analysis, we proposed a
fully automatic pipeline for creating a benchmark that requires knowledge of
long-tail facts for answering the involved questions. Using this pipeline, we
introduce the LTGen benchmark. We evaluate state-of-the-art LLMs in different
knowledge settings using the proposed benchmark. Our experiments show that LLMs
alone struggle with answering these questions, especially when the long-tail
level is high or rich knowledge is required. Nonetheless, the performance of
the same models improved significantly when they were prompted with
non-parametric knowledge. We observed that, in most cases, prompting LLMs with
KG triples surpasses passage-based prompting using a state-of-the-art
retriever. In addition, while prompting LLMs with both KG triples and documents
does not consistently improve knowledge coverage, it can dramatically reduce
hallucinations in the generated content. | cs.CL | Prompting Large Language Models with Knowledge Graphs for Question Answering Involving Long-tail Facts |
2024-05-10T14:23:43Z | http://arxiv.org/pdf/2405.06499v1 | http://arxiv.org/abs/2405.06499v1 | Haifa Alrdahi, Riza Batista-Navarro | 2024-05-10T14:23:43Z | cs.CL | The chess domain is well-suited for creating an artificial intelligence (AI)
system that mimics real-world challenges, including decision-making. Throughout
the years, minimal attention has been paid to investigating insights derived
from unstructured chess data sources. In this study, we examine the complicated
relationships between multiple referenced moves in a chess-teaching textbook,
and propose a novel method designed to encapsulate chess knowledge derived from
move-action phrases. This study investigates the feasibility of using a
modified sentiment analysis method as a means for evaluating chess moves based
on text. Our proposed Aspect-Based Sentiment Analysis (ABSA) method represents
an advancement in evaluating the sentiment associated with referenced chess
moves. By extracting insights from move-action phrases, our approach aims to
provide a more fine-grained and contextually aware `chess move'-based sentiment
classification. Through empirical experiments and analysis, we evaluate the
performance of our fine-tuned ABSA model, presenting results that confirm the
efficiency of our approach in advancing aspect-based sentiment classification
within the chess domain. This research contributes to the area of game-playing
by machines and shows the practical applicability of leveraging NLP techniques
to understand the context of strategic games. | cs.CL | Aspect-based Sentiment Evaluation of Chess Moves (ASSESS): an NLP-based Method for Evaluating Chess Strategies from Textbooks |
2024-05-10T14:03:37Z | http://arxiv.org/pdf/2405.06483v1 | http://arxiv.org/abs/2405.06483v1 | Ana Ezquerro, David Vilares | 2024-05-10T14:03:37Z | cs.CL | This paper describes our participation in SemEval 2024 Task 3, which focused
on Multimodal Emotion Cause Analysis in Conversations. We developed an early
prototype for an end-to-end system that uses graph-based methods from
dependency parsing to identify causal emotion relations in multi-party
conversations. Our model comprises a neural transformer-based encoder for
contextualizing multimodal conversation data and a graph-based decoder for
generating the adjacency matrix scores of the causal graph. We ranked 7th out
of 15 valid and official submissions for Subtask 1, using textual inputs only.
We also discuss our participation in Subtask 2 during post-evaluation using
multi-modal inputs. | cs.CL | LyS at SemEval-2024 Task 3: An Early Prototype for End-to-End Multimodal Emotion Linking as Graph-Based Parsing |
2024-05-10T13:27:32Z | http://arxiv.org/pdf/2405.06468v1 | http://arxiv.org/abs/2405.06468v1 | Yaoqin Ye, Junjie Zhang, Hongwei Shi | 2024-05-10T13:27:32Z | cs.CV, cs.CL | The task of medical image recognition is notably complicated by the presence
of varied and multiple pathological indications, presenting a unique challenge
in multi-label classification with unseen labels. This complexity underlines
the need for computer-aided diagnosis methods employing multi-label zero-shot
learning. Recent advancements in pre-trained vision-language models (VLMs) have
showcased notable zero-shot classification abilities on medical images.
However, these methods have limitations on leveraging extensive pre-trained
knowledge from broader image datasets, and often depend on manual prompt
construction by expert radiologists. By automating the process of prompt
tuning, prompt learning techniques have emerged as an efficient way to adapt
VLMs to downstream tasks. Yet, existing CoOp-based strategies fall short in
performing class-specific prompts on unseen categories, limiting
generalizability in fine-grained scenarios. To overcome these constraints, we
introduce a novel prompt generation approach inspirited by text generation in
natural language processing (NLP). Our method, named Pseudo-Prompt Generating
(PsPG), capitalizes on the priori knowledge of multi-modal features. Featuring
a RNN-based decoder, PsPG autoregressively generates class-tailored embedding
vectors, i.e., pseudo-prompts. Comparative evaluations on various multi-label
chest radiograph datasets affirm the superiority of our approach against
leading medical vision-language and multi-label prompt learning methods. The
source code is available at https://github.com/fallingnight/PsPG | cs.CV | Pseudo-Prompt Generating in Pre-trained Vision-Language Models for Multi-Label Medical Image Classification |
2024-05-10T13:22:20Z | http://arxiv.org/pdf/2405.06725v3 | http://arxiv.org/abs/2405.06725v3 | Jingkai Li | 2024-05-15T02:46:45Z | q-bio.NC, cs.AI, cs.CL, cs.LG | With the rise of Large Language Models (LLMs), the novel metric "Brainscore"
emerged as a means to evaluate the functional similarity between LLMs and human
brain/neural systems. Our efforts were dedicated to mining the meaning of the
novel score by constructing topological features derived from both human fMRI
data involving 190 subjects, and 39 LLMs plus their untrained counterparts.
Subsequently, we trained 36 Linear Regression Models and conducted thorough
statistical analyses to discern reliable and valid features from our
constructed ones. Our findings reveal distinctive feature combinations
conducive to interpreting existing brainscores across various brain regions of
interest (ROIs) and hemispheres, thereby significantly contributing to
advancing interpretable machine learning (iML) studies. The study is enriched
by our further discussions and analyses concerning existing brainscores. To our
knowledge, this study represents the first attempt to comprehend the novel
metric brainscore within this interdisciplinary domain. | q-bio.NC | On the Shape of Brainscores for Large Language Models (LLMs) |
2024-05-10T13:10:55Z | http://arxiv.org/pdf/2405.06459v3 | http://arxiv.org/abs/2405.06459v3 | Hyejeong Jo, Yiqian Yang, Juhyeok Han, Yiqun Duan, Hui Xiong, Won Hee Lee | 2024-06-14T02:27:00Z | cs.CL, cs.AI | This work critically analyzes existing models for open-vocabulary EEG-to-Text
translation. We identify a crucial limitation: previous studies often employed
implicit teacher-forcing during evaluation, artificially inflating performance
metrics. Additionally, they lacked a critical benchmark - comparing model
performance on pure noise inputs. We propose a methodology to differentiate
between models that truly learn from EEG signals and those that simply memorize
training data. Our analysis reveals that model performance on noise data can be
comparable to that on EEG data. These findings highlight the need for stricter
evaluation practices in EEG-to-Text research, emphasizing transparent reporting
and rigorous benchmarking with noise inputs. This approach will lead to more
reliable assessments of model capabilities and pave the way for robust
EEG-to-Text communication systems. | cs.CL | Are EEG-to-Text Models Working? |
2024-05-10T13:04:21Z | http://arxiv.org/pdf/2405.06454v2 | http://arxiv.org/abs/2405.06454v2 | Mohammad Ghiasvand Mohammadkhani, Niloofar Ranjbar, Saeedeh Momtazi | 2024-05-16T11:37:51Z | cs.CL | Generative approaches have significantly influenced Aspect-Based Sentiment
Analysis (ABSA), garnering considerable attention. However, existing studies
often predict target text components monolithically, neglecting the benefits of
utilizing single elements for tuple prediction. In this paper, we introduce
Element to Tuple Prompting (E2TP), employing a two-step architecture. The
former step focuses on predicting single elements, while the latter step
completes the process by mapping these predicted elements to their
corresponding tuples. E2TP is inspired by human problem-solving, breaking down
tasks into manageable parts, using the first step's output as a guide in the
second step. Within this strategy, three types of paradigms, namely
E2TP($diet$), E2TP($f_1$), and E2TP($f_2$), are designed to facilitate the
training process. Beyond dataset-specific experiments, our paper addresses
cross-domain scenarios, demonstrating the effectiveness and generalizability of
the approach. By conducting a comprehensive analysis on various benchmarks, we
show that E2TP achieves new state-of-the-art results in nearly all cases. | cs.CL | E2TP: Element to Tuple Prompting Improves Aspect Sentiment Tuple Prediction |
2024-05-10T12:14:11Z | http://arxiv.org/pdf/2405.06424v2 | http://arxiv.org/abs/2405.06424v2 | JoonHo Lee, Jae Oh Woo, Juree Seok, Parisa Hassanzadeh, Wooseok Jang, JuYoun Son, Sima Didari, Baruch Gutow, Heng Hao, Hankyu Moon, Wenjun Hu, Yeong-Dae Kwon, Taehee Lee, Seungjai Min | 2024-05-19T17:35:43Z | cs.CL, cs.AI, cs.LG | Assessing response quality to instructions in language models is vital but
challenging due to the complexity of human language across different contexts.
This complexity often results in ambiguous or inconsistent interpretations,
making accurate assessment difficult. To address this issue, we propose a novel
Uncertainty-aware Reward Model (URM) that introduces a robust uncertainty
estimation for the quality of paired responses based on Bayesian approximation.
Trained with preference datasets, our uncertainty-enabled proxy not only scores
rewards for responses but also evaluates their inherent uncertainty. Empirical
results demonstrate significant benefits of incorporating the proposed proxy
into language model training. Our method boosts the instruction following
capability of language models by refining data curation for training and
improving policy optimization objectives, thereby surpassing existing methods
by a large margin on benchmarks such as Vicuna and MT-bench. These findings
highlight that our proposed approach substantially advances language model
training and paves a new way of harnessing uncertainty within language models. | cs.CL | Improving Instruction Following in Language Models through Proxy-Based Uncertainty Estimation |
2024-05-10T11:53:53Z | http://arxiv.org/pdf/2405.06414v1 | http://arxiv.org/abs/2405.06414v1 | Hunter McNichols, Jaewook Lee, Stephen Fancsali, Steve Ritter, Andrew Lan | 2024-05-10T11:53:53Z | cs.CL | Intelligent Tutoring Systems (ITSs) often contain an automated feedback
component, which provides a predefined feedback message to students when they
detect a predefined error. To such a feedback component, we often resort to
template-based approaches. These approaches require significant effort from
human experts to detect a limited number of possible student errors and provide
corresponding feedback. This limitation is exemplified in open-ended math
questions, where there can be a large number of different incorrect errors. In
our work, we examine the capabilities of large language models (LLMs) to
generate feedback for open-ended math questions, similar to that of an
established ITS that uses a template-based approach. We fine-tune both
open-source and proprietary LLMs on real student responses and corresponding
ITS-provided feedback. We measure the quality of the generated feedback using
text similarity metrics. We find that open-source and proprietary models both
show promise in replicating the feedback they see during training, but do not
generalize well to previously unseen student errors. These results suggest that
despite being able to learn the formatting of feedback, LLMs are not able to
fully understand mathematical errors made by students. | cs.CL | Can Large Language Models Replicate ITS Feedback on Open-Ended Math Questions? |
2024-05-10T11:44:05Z | http://arxiv.org/pdf/2405.06410v1 | http://arxiv.org/abs/2405.06410v1 | Ning Cheng, Zhaohui Yan, Ziming Wang, Zhijie Li, Jiaming Yu, Zilong Zheng, Kewei Tu, Jinan Xu, Wenjuan Han | 2024-05-10T11:44:05Z | cs.CL | Large Language Models (LLMs) play a crucial role in capturing structured
semantics to enhance language understanding, improve interpretability, and
reduce bias. Nevertheless, an ongoing controversy exists over the extent to
which LLMs can grasp structured semantics. To assess this, we propose using
Semantic Role Labeling (SRL) as a fundamental task to explore LLMs' ability to
extract structured semantics. In our assessment, we employ the prompting
approach, which leads to the creation of our few-shot SRL parser, called
PromptSRL. PromptSRL enables LLMs to map natural languages to explicit semantic
structures, which provides an interpretable window into the properties of LLMs.
We find interesting potential: LLMs can indeed capture semantic structures, and
scaling-up doesn't always mirror potential. Additionally, limitations of LLMs
are observed in C-arguments, etc. Lastly, we are surprised to discover that
significant overlap in the errors is made by both LLMs and untrained humans,
accounting for almost 30% of all errors. | cs.CL | Potential and Limitations of LLMs in Capturing Structured Semantics: A Case Study on SRL |
2024-05-10T10:19:14Z | http://arxiv.org/pdf/2405.06373v3 | http://arxiv.org/abs/2405.06373v3 | Li-Chun Lu, Shou-Jen Chen, Tsung-Min Pai, Chan-Hung Yu, Hung-yi Lee, Shao-Hua Sun | 2024-05-18T16:33:46Z | cs.CL, cs.AI | Large language models (LLMs) have shown exceptional proficiency in natural
language processing but often fall short of generating creative and original
responses to open-ended questions. To enhance LLM creativity, our key insight
is to emulate the human process of inducing collective creativity through
engaging discussions with participants from diverse backgrounds and
perspectives. To this end, we propose LLM Discussion, a three-phase discussion
framework that facilitates vigorous and diverging idea exchanges and ensures
convergence to creative answers. Moreover, we adopt a role-playing technique by
assigning distinct roles to LLMs to combat the homogeneity of LLMs. We evaluate
the efficacy of the proposed framework with the Alternative Uses Test,
Similarities Test, Instances Test, and Scientific Creativity Test through both
LLM evaluation and human study. Our proposed framework outperforms single-LLM
approaches and existing multi-LLM frameworks across various creativity metrics. | cs.CL | LLM Discussion: Enhancing the Creativity of Large Language Models via Discussion Framework and Role-Play |
2024-05-10T09:26:12Z | http://arxiv.org/pdf/2405.06346v1 | http://arxiv.org/abs/2405.06346v1 | Rishav Hada, Safiya Husain, Varun Gumma, Harshita Diddee, Aditya Yadavalli, Agrima Seth, Nidhi Kulkarni, Ujwal Gadiraju, Aditya Vashistha, Vivek Seshadri, Kalika Bali | 2024-05-10T09:26:12Z | cs.CL | Existing research in measuring and mitigating gender bias predominantly
centers on English, overlooking the intricate challenges posed by non-English
languages and the Global South. This paper presents the first comprehensive
study delving into the nuanced landscape of gender bias in Hindi, the third
most spoken language globally. Our study employs diverse mining techniques,
computational models, field studies and sheds light on the limitations of
current methodologies. Given the challenges faced with mining gender biased
statements in Hindi using existing methods, we conducted field studies to
bootstrap the collection of such sentences. Through field studies involving
rural and low-income community women, we uncover diverse perceptions of gender
bias, underscoring the necessity for context-specific approaches. This paper
advocates for a community-centric research design, amplifying voices often
marginalized in previous studies. Our findings not only contribute to the
understanding of gender bias in Hindi but also establish a foundation for
further exploration of Indic languages. By exploring the intricacies of this
understudied context, we call for thoughtful engagement with gender bias,
promoting inclusivity and equity in linguistic and cultural contexts beyond the
Global North. | cs.CL | Akal Badi ya Bias: An Exploratory Study of Gender Bias in Hindi Language Technology |
2024-05-10T09:03:27Z | http://arxiv.org/pdf/2405.06331v1 | http://arxiv.org/abs/2405.06331v1 | John Kirchenbauer, Garrett Honke, Gowthami Somepalli, Jonas Geiping, Daphne Ippolito, Katherine Lee, Tom Goldstein, David Andre | 2024-05-10T09:03:27Z | cs.LG, cs.CL | We develop a methodology for analyzing language model task performance at the
individual example level based on training data density estimation. Experiments
with paraphrasing as a controlled intervention on finetuning data demonstrate
that increasing the support in the training distribution for specific test
queries results in a measurable increase in density, which is also a
significant predictor of the performance increase caused by the intervention.
Experiments with pretraining data demonstrate that we can explain a significant
fraction of the variance in model perplexity via density measurements. We
conclude that our framework can provide statistical evidence of the dependence
of a target model's predictions on subsets of its training data, and can more
generally be used to characterize the support (or lack thereof) in the training
data for a given test task. | cs.LG | LMD3: Language Model Data Density Dependence |
2024-05-10T08:48:03Z | http://arxiv.org/pdf/2405.06321v2 | http://arxiv.org/abs/2405.06321v2 | Xin Du, Kumiko Tanaka-Ishii | 2024-05-15T07:46:01Z | cs.CL, cond-mat.stat-mech, cs.AI | The correlation dimension of natural language is measured by applying the
Grassberger-Procaccia algorithm to high-dimensional sequences produced by a
large-scale language model. This method, previously studied only in a Euclidean
space, is reformulated in a statistical manifold via the Fisher-Rao distance.
Language exhibits a multifractal, with global self-similarity and a universal
dimension around 6.5, which is smaller than those of simple discrete random
sequences and larger than that of a Barab\'asi-Albert process. Long memory is
the key to producing self-similarity. Our method is applicable to any
probabilistic model of real-world discrete sequences, and we show an
application to music data. | cs.CL | Correlation Dimension of Natural Language in a Statistical Manifold |
2024-05-10T08:45:23Z | http://arxiv.org/pdf/2405.06319v1 | http://arxiv.org/abs/2405.06319v1 | Hanna-Sophia Widhoelzl, Ece Takmaz | 2024-05-10T08:45:23Z | cs.CV, cs.CL | This study investigates the cognitive plausibility of a pretrained multimodal
model, CLIP, in recognizing emotions evoked by abstract visual art. We employ a
dataset comprising images with associated emotion labels and textual rationales
of these labels provided by human annotators. We perform linguistic analyses of
rationales, zero-shot emotion classification of images and rationales, apply
similarity-based prediction of emotion, and investigate color-emotion
associations. The relatively low, yet above baseline, accuracy in recognizing
emotion for abstract images and rationales suggests that CLIP decodes emotional
complexities in a manner not well aligned with human cognitive processes.
Furthermore, we explore color-emotion interactions in images and rationales.
Expected color-emotion associations, such as red relating to anger, are
identified in images and texts annotated with emotion labels by both humans and
CLIP, with the latter showing even stronger interactions. Our results highlight
the disparity between human processing and machine processing when connecting
image features and emotions. | cs.CV | Decoding Emotions in Abstract Art: Cognitive Plausibility of CLIP in Recognizing Color-Emotion Associations |
2024-05-10T08:31:04Z | http://arxiv.org/pdf/2405.06306v1 | http://arxiv.org/abs/2405.06306v1 | Javier Coronado-Blázquez | 2024-05-10T08:31:04Z | cs.CL, cs.LG | Many videogames suffer "review bombing" -a large volume of unusually low
scores that in many cases do not reflect the real quality of the product- when
rated by users. By taking Metacritic's 50,000+ user score aggregations for PC
games in English language, we use a Natural Language Processing (NLP) approach
to try to understand the main words and concepts appearing in such cases,
reaching a 0.88 accuracy on a validation set when distinguishing between just
bad ratings and review bombings. By uncovering and analyzing the patterns
driving this phenomenon, these results could be used to further mitigate these
situations. | cs.CL | A NLP Approach to "Review Bombing" in Metacritic PC Videogames User Ratings |
2024-05-10T08:09:53Z | http://arxiv.org/pdf/2405.06301v1 | http://arxiv.org/abs/2405.06301v1 | David Lindsay, Sian Lindsay | 2024-05-10T08:09:53Z | cs.LG, cs.AI, cs.CE, cs.CL, cs.CV | The Universal Similarity Metric (USM) has been demonstrated to give
practically useful measures of "similarity" between sequence data. Here we have
used the USM as an alternative distance metric in a K-Nearest Neighbours (K-NN)
learner to allow effective pattern recognition of variable length sequence
data. We compare this USM approach with the commonly used string-to-word vector
approach. Our experiments have used two data sets of divergent domains: (1)
spam email filtering and (2) protein subcellular localization. Our results with
this data reveal that the USM-based K-NN learner (1) gives predictions with
higher classification accuracy than those output by techniques that use the
string-to-word vector approach, and (2) can be used to generate reliable
probability forecasts. | cs.LG | Learning from String Sequences |
2024-05-10T07:52:43Z | http://arxiv.org/pdf/2405.06295v1 | http://arxiv.org/abs/2405.06295v1 | Rochana Chaturvedi, Abari Bhattacharya, Shweta Yadav | 2024-05-10T07:52:43Z | cs.CL, cs.AI, H.4.3; I.2.7; J.3; J.7; K.6.4 | Community Question-Answering (CQA) forums have revolutionized how people seek
information, especially those related to their healthcare needs, placing their
trust in the collective wisdom of the public. However, there can be several
answers in response to a single query, which makes it hard to grasp the key
information related to the specific health concern. Typically, CQA forums
feature a single top-voted answer as a representative summary for each query.
However, a single answer overlooks the alternative solutions and other
information frequently offered in other responses. Our research focuses on
aspect-based summarization of health answers to address this limitation.
Summarization of responses under different aspects such as suggestions,
information, personal experiences, and questions can enhance the usability of
the platforms. We formalize a multi-stage annotation guideline and contribute a
unique dataset comprising aspect-based human-written health answer summaries.
We build an automated multi-faceted answer summarization pipeline with this
dataset based on task-specific fine-tuning of several state-of-the-art models.
The pipeline leverages question similarity to retrieve relevant answer
sentences, subsequently classifying them into the appropriate aspect type.
Following this, we employ several recent abstractive summarization models to
generate aspect-based summaries. Finally, we present a comprehensive human
analysis and find that our summaries rank high in capturing relevant content
and a wide range of solutions. | cs.CL | Aspect-oriented Consumer Health Answer Summarization |
2024-05-10T07:05:02Z | http://arxiv.org/pdf/2405.06275v1 | http://arxiv.org/abs/2405.06275v1 | Nan Zhang, Yanchi Liu, Xujiang Zhao, Wei Cheng, Runxue Bao, Rui Zhang, Prasenjit Mitra, Haifeng Chen | 2024-05-10T07:05:02Z | cs.CL | Large Language Models (LLMs) have exhibited remarkable proficiency across a
wide array of NLP tasks. However, the escalation in model size also engenders
substantial deployment costs. While few efforts have explored model pruning
techniques to reduce the size of LLMs, they mainly center on general or
task-specific weights. This leads to suboptimal performance due to lacking
specificity on the target domain or generality on different tasks when applied
to domain-specific challenges. This work introduces an innovative unstructured
dual-pruning methodology, D-Pruner, for domain-specific compression on LLM. It
extracts a compressed, domain-specific, and task-agnostic LLM by identifying
LLM weights that are pivotal for general capabilities, like linguistic
capability and multi-task solving, and domain-specific knowledge. More
specifically, we first assess general weight importance by quantifying the
error incurred upon their removal with the help of an open-domain calibration
dataset. Then, we utilize this general weight importance to refine the training
loss, so that it preserves generality when fitting into a specific domain.
Moreover, by efficiently approximating weight importance with the refined
training loss on a domain-specific calibration dataset, we obtain a pruned
model emphasizing generality and specificity. Our comprehensive experiments
across various tasks in healthcare and legal domains show the effectiveness of
D-Pruner in domain-specific compression. Our code is available at
https://github.com/psunlpgroup/D-Pruner. | cs.CL | Pruning as a Domain-specific LLM Extractor |
2024-05-10T06:52:44Z | http://arxiv.org/pdf/2405.06270v3 | http://arxiv.org/abs/2405.06270v3 | Fatemeh Nazary, Yashar Deldjoo, Tommaso Di Noia, Eugenio di Sciascio | 2024-06-03T16:23:28Z | cs.LG, cs.AI, cs.CL | The integration of Large Language Models (LLMs) into healthcare diagnostics
offers a promising avenue for clinical decision-making. This study outlines the
development of a novel method for zero-shot/few-shot in-context learning (ICL)
by integrating medical domain knowledge using a multi-layered structured
prompt. We also explore the efficacy of two communication styles between the
user and LLMs: the Numerical Conversational (NC) style, which processes data
incrementally, and the Natural Language Single-Turn (NL-ST) style, which
employs long narrative prompts.
Our study systematically evaluates the diagnostic accuracy and risk factors,
including gender bias and false negative rates, using a dataset of 920 patient
records in various few-shot scenarios. Results indicate that traditional
clinical machine learning (ML) models generally outperform LLMs in zero-shot
and few-shot settings. However, the performance gap narrows significantly when
employing few-shot examples alongside effective explainable AI (XAI) methods as
sources of domain knowledge. Moreover, with sufficient time and an increased
number of examples, the conversational style (NC) nearly matches the
performance of ML models. Most notably, LLMs demonstrate comparable or superior
cost-sensitive accuracy relative to ML models.
This research confirms that, with appropriate domain knowledge and tailored
communication strategies, LLMs can significantly enhance diagnostic processes.
The findings highlight the importance of optimizing the number of training
examples and communication styles to improve accuracy and reduce biases in LLM
applications. | cs.LG | XAI4LLM. Let Machine Learning Models and LLMs Collaborate for Enhanced In-Context Learning in Healthcare |
2024-05-10T06:14:07Z | http://arxiv.org/pdf/2405.06258v2 | http://arxiv.org/abs/2405.06258v2 | Jiarui Liu, Wenkai Li, Zhijing Jin, Mona Diab | 2024-06-19T03:36:01Z | cs.CL | In an era of model and data proliferation in machine learning/AI especially
marked by the rapid advancement of open-sourced technologies, there arises a
critical need for standardized consistent documentation. Our work addresses the
information incompleteness in current human-generated model and data cards. We
propose an automated generation approach using Large Language Models (LLMs).
Our key contributions include the establishment of CardBench, a comprehensive
dataset aggregated from over 4.8k model cards and 1.4k data cards, coupled with
the development of the CardGen pipeline comprising a two-step retrieval
process. Our approach exhibits enhanced completeness, objectivity, and
faithfulness in generated model and data cards, a significant step in
responsible AI documentation practices ensuring better accountability and
traceability. | cs.CL | Automatic Generation of Model and Data Cards: A Step Towards Responsible AI |
2024-05-10T04:22:54Z | http://arxiv.org/pdf/2405.06239v1 | http://arxiv.org/abs/2405.06239v1 | Faisal Qarah | 2024-05-10T04:22:54Z | cs.CL, cs.AI | In this paper, we introduce SaudiBERT, a monodialect Arabic language model
pretrained exclusively on Saudi dialectal text. To demonstrate the model's
effectiveness, we compared SaudiBERT with six different multidialect Arabic
language models across 11 evaluation datasets, which are divided into two
groups: sentiment analysis and text classification. SaudiBERT achieved average
F1-scores of 86.15\% and 87.86\% in these groups respectively, significantly
outperforming all other comparative models. Additionally, we present two novel
Saudi dialectal corpora: the Saudi Tweets Mega Corpus (STMC), which contains
over 141 million tweets in Saudi dialect, and the Saudi Forums Corpus (SFC),
which includes 15.2 GB of text collected from five Saudi online forums. Both
corpora are used in pretraining the proposed model, and they are the largest
Saudi dialectal corpora ever reported in the literature. The results confirm
the effectiveness of SaudiBERT in understanding and analyzing Arabic text
expressed in Saudi dialect, achieving state-of-the-art results in most tasks
and surpassing other language models included in the study. SaudiBERT model is
publicly available on \url{https://huggingface.co/faisalq/SaudiBERT}. | cs.CL | SaudiBERT: A Large Language Model Pretrained on Saudi Dialect Corpora |
2024-05-10T03:16:07Z | http://arxiv.org/pdf/2405.06221v1 | http://arxiv.org/abs/2405.06221v1 | Xiaocong Du, Haipeng Zhang | 2024-05-10T03:16:07Z | cs.CL, cs.CY | Achieving gender equality is a pivotal factor in realizing the UN's Global
Goals for Sustainable Development. Gender bias studies work towards this and
rely on name-based gender inference tools to assign individual gender labels
when gender information is unavailable. However, these tools often inaccurately
predict gender for Chinese Pinyin names, leading to potential bias in such
studies. With the growing participation of Chinese in international activities,
this situation is becoming more severe. Specifically, current tools focus on
pronunciation (Pinyin) information, neglecting the fact that the latent
connections between Pinyin and Chinese characters (Hanzi) behind convey
critical information. As a first effort, we formulate the Pinyin name-gender
guessing problem and design a Multi-Task Learning Network assisted by Knowledge
Distillation that enables the Pinyin embeddings in the model to possess
semantic features of Chinese characters and to learn gender information from
Chinese character names. Our open-sourced method surpasses commercial
name-gender guessing tools by 9.70\% to 20.08\% relatively, and also
outperforms the state-of-the-art algorithms. | cs.CL | For the Misgendered Chinese in Gender Bias Research: Multi-Task Learning with Knowledge Distillation for Pinyin Name-Gender Prediction |
2024-05-10T03:14:26Z | http://arxiv.org/pdf/2405.06719v1 | http://arxiv.org/abs/2405.06719v1 | Xiannan Huang | 2024-05-10T03:14:26Z | cs.CL, cs.AI | Traffic prediction is pivotal for rational transportation supply scheduling
and allocation. Existing researches into short-term traffic prediction,
however, face challenges in adequately addressing exceptional circumstances and
integrating non-numerical contextual information like weather into models.
While, Large language models offer a promising solution due to their inherent
world knowledge. However, directly using them for traffic prediction presents
drawbacks such as high cost, lack of determinism, and limited mathematical
capability. To mitigate these issues, this study proposes a novel approach.
Instead of directly employing large models for prediction, it utilizes them to
process textual information and obtain embeddings. These embeddings are then
combined with historical traffic data and inputted into traditional
spatiotemporal forecasting models. The study investigates two types of special
scenarios: regional-level and node-level. For regional-level scenarios, textual
information is represented as a node connected to the entire network. For
node-level scenarios, embeddings from the large model represent additional
nodes connected only to corresponding nodes. This approach shows a significant
improvement in prediction accuracy according to our experiment of New York Bike
dataset. | cs.CL | Enhancing Traffic Prediction with Textual Data Using Large Language Models |
2024-05-10T03:06:24Z | http://arxiv.org/pdf/2405.06219v2 | http://arxiv.org/abs/2405.06219v2 | Haojie Duanmu, Zhihang Yuan, Xiuhong Li, Jiangfei Duan, Xingcheng Zhang, Dahua Lin | 2024-05-13T14:39:11Z | cs.LG, cs.CL | Large language models (LLMs) can now handle longer sequences of tokens,
enabling complex tasks like book understanding and generating lengthy novels.
However, the key-value (KV) cache required for LLMs consumes substantial memory
as context length increasing, becoming the bottleneck for deployment. In this
paper, we present a strategy called SKVQ, which stands for sliding-window KV
cache quantization, to address the issue of extremely low bitwidth KV cache
quantization. To achieve this, SKVQ rearranges the channels of the KV cache in
order to improve the similarity of channels in quantization groups, and applies
clipped dynamic quantization at the group level. Additionally, SKVQ ensures
that the most recent window tokens in the KV cache are preserved with high
precision. This helps maintain the accuracy of a small but important portion of
the KV cache.SKVQ achieves high compression ratios while maintaining accuracy.
Our evaluation on LLMs demonstrates that SKVQ surpasses previous quantization
approaches, allowing for quantization of the KV cache to 2-bit keys and 1.5-bit
values with minimal loss of accuracy. With SKVQ, it is possible to process
context lengths of up to 1M on an 80GB memory GPU for a 7b model and up to 7
times faster decoding. | cs.LG | SKVQ: Sliding-window Key and Value Cache Quantization for Large Language Models |
2024-05-10T02:48:45Z | http://arxiv.org/pdf/2405.06211v3 | http://arxiv.org/abs/2405.06211v3 | Wenqi Fan, Yujuan Ding, Liangbo Ning, Shijie Wang, Hengyun Li, Dawei Yin, Tat-Seng Chua, Qing Li | 2024-06-17T08:56:38Z | cs.CL, cs.AI, cs.IR | As one of the most advanced techniques in AI, Retrieval-Augmented Generation
(RAG) can offer reliable and up-to-date external knowledge, providing huge
convenience for numerous tasks. Particularly in the era of AI-Generated Content
(AIGC), the powerful capacity of retrieval in providing additional knowledge
enables RAG to assist existing generative AI in producing high-quality outputs.
Recently, Large Language Models (LLMs) have demonstrated revolutionary
abilities in language understanding and generation, while still facing inherent
limitations, such as hallucinations and out-of-date internal knowledge. Given
the powerful abilities of RAG in providing the latest and helpful auxiliary
information, Retrieval-Augmented Large Language Models (RA-LLMs) have emerged
to harness external and authoritative knowledge bases, rather than solely
relying on the model's internal knowledge, to augment the generation quality of
LLMs. In this survey, we comprehensively review existing research studies in
RA-LLMs, covering three primary technical perspectives: architectures, training
strategies, and applications. As the preliminary knowledge, we briefly
introduce the foundations and recent advances of LLMs. Then, to illustrate the
practical significance of RAG for LLMs, we systematically review mainstream
relevant work by their architectures, training strategies, and application
areas, detailing specifically the challenges of each and the corresponding
capabilities of RA-LLMs. Finally, to deliver deeper insights, we discuss
current limitations and several promising directions for future research.
Updated information about this survey can be found at
https://advanced-recommender-systems.github.io/RAG-Meets-LLMs/ | cs.CL | A Survey on RAG Meeting LLMs: Towards Retrieval-Augmented Large Language Models |
2024-05-10T02:40:49Z | http://arxiv.org/pdf/2405.06204v1 | http://arxiv.org/abs/2405.06204v1 | Bowen Xing, Ivor W. Tsang | 2024-05-10T02:40:49Z | cs.CL, cs.AI | State-of-the-art model for zero-shot cross-lingual spoken language
understanding performs cross-lingual unsupervised contrastive learning to
achieve the label-agnostic semantic alignment between each utterance and its
code-switched data. However, it ignores the precious intent/slot labels, whose
label information is promising to help capture the label-aware semantics
structure and then leverage supervised contrastive learning to improve both
source and target languages' semantics. In this paper, we propose Hybrid and
Cooperative Contrastive Learning to address this problem. Apart from
cross-lingual unsupervised contrastive learning, we design a holistic approach
that exploits source language supervised contrastive learning, cross-lingual
supervised contrastive learning and multilingual supervised contrastive
learning to perform label-aware semantics alignments in a comprehensive manner.
Each kind of supervised contrastive learning mechanism includes both
single-task and joint-task scenarios. In our model, one contrastive learning
mechanism's input is enhanced by others. Thus the total four contrastive
learning mechanisms are cooperative to learn more consistent and discriminative
representations in the virtuous cycle during the training process. Experiments
show that our model obtains consistent improvements over 9 languages, achieving
new state-of-the-art performance. | cs.CL | HC$^2$L: Hybrid and Cooperative Contrastive Learning for Cross-lingual Spoken Language Understanding |
2024-05-10T02:23:56Z | http://arxiv.org/pdf/2405.06196v2 | http://arxiv.org/abs/2405.06196v2 | Manish Dhakal, Rabin Adhikari, Safal Thapaliya, Bishesh Khanal | 2024-06-27T14:19:56Z | cs.CV, cs.AI, cs.CL, cs.LG | Foundation Vision-Language Models (VLMs) trained using large-scale
open-domain images and text pairs have recently been adapted to develop
Vision-Language Segmentation Models (VLSMs) that allow providing text prompts
during inference to guide image segmentation. If robust and powerful VLSMs can
be built for medical images, it could aid medical professionals in many
clinical tasks where they must spend substantial time delineating the target
structure of interest. VLSMs for medical images resort to fine-tuning base VLM
or VLSM pretrained on open-domain natural image datasets due to fewer annotated
medical image datasets; this fine-tuning is resource-consuming and expensive as
it usually requires updating all or a significant fraction of the pretrained
parameters. Recently, lightweight blocks called adapters have been proposed in
VLMs that keep the pretrained model frozen and only train adapters during
fine-tuning, substantially reducing the computing resources required. We
introduce a novel adapter, VLSM-Adapter, that can fine-tune pretrained
vision-language segmentation models using transformer encoders. Our experiments
in widely used CLIP-based segmentation models show that with only 3 million
trainable parameters, the VLSM-Adapter outperforms state-of-the-art and is
comparable to the upper bound end-to-end fine-tuning. The source code is
available at: https://github.com/naamiinepal/vlsm-adapter. | cs.CV | VLSM-Adapter: Finetuning Vision-Language Segmentation Efficiently with Lightweight Blocks |
2024-05-10T00:16:58Z | http://arxiv.org/pdf/2405.06150v1 | http://arxiv.org/abs/2405.06150v1 | Dena Mujtaba, Nihar R. Mahapatra, Megan Arney, J. Scott Yaruss, Hope Gerlach-Houck, Caryn Herring, Jia Bin | 2024-05-10T00:16:58Z | cs.CL, cs.CY, eess.AS | Automatic speech recognition (ASR) systems, increasingly prevalent in
education, healthcare, employment, and mobile technology, face significant
challenges in inclusivity, particularly for the 80 million-strong global
community of people who stutter. These systems often fail to accurately
interpret speech patterns deviating from typical fluency, leading to critical
usability issues and misinterpretations. This study evaluates six leading ASRs,
analyzing their performance on both a real-world dataset of speech samples from
individuals who stutter and a synthetic dataset derived from the widely-used
LibriSpeech benchmark. The synthetic dataset, uniquely designed to incorporate
various stuttering events, enables an in-depth analysis of each ASR's handling
of disfluent speech. Our comprehensive assessment includes metrics such as word
error rate (WER), character error rate (CER), and semantic accuracy of the
transcripts. The results reveal a consistent and statistically significant
accuracy bias across all ASRs against disfluent speech, manifesting in
significant syntactical and semantic inaccuracies in transcriptions. These
findings highlight a critical gap in current ASR technologies, underscoring the
need for effective bias mitigation strategies. Addressing this bias is
imperative not only to improve the technology's usability for people who
stutter but also to ensure their equitable and inclusive participation in the
rapidly evolving digital landscape. | cs.CL | Lost in Transcription: Identifying and Quantifying the Accuracy Biases of Automatic Speech Recognition Systems Against Disfluent Speech |
2024-05-09T23:43:57Z | http://arxiv.org/pdf/2405.06145v1 | http://arxiv.org/abs/2405.06145v1 | Yao Ge, Sudeshna Das, Karen O'Connor, Mohammed Ali Al-Garadi, Graciela Gonzalez-Hernandez, Abeed Sarker | 2024-05-09T23:43:57Z | cs.CL, cs.AI, cs.LG | Substance use disorders (SUDs) are a growing concern globally, necessitating
enhanced understanding of the problem and its trends through data-driven
research. Social media are unique and important sources of information about
SUDs, particularly since the data in such sources are often generated by people
with lived experiences. In this paper, we introduce Reddit-Impacts, a
challenging Named Entity Recognition (NER) dataset curated from subreddits
dedicated to discussions on prescription and illicit opioids, as well as
medications for opioid use disorder. The dataset specifically concentrates on
the lesser-studied, yet critically important, aspects of substance use--its
clinical and social impacts. We collected data from chosen subreddits using the
publicly available Application Programming Interface for Reddit. We manually
annotated text spans representing clinical and social impacts reported by
people who also reported personal nonmedical use of substances including but
not limited to opioids, stimulants and benzodiazepines. Our objective is to
create a resource that can enable the development of systems that can
automatically detect clinical and social impacts of substance use from
text-based social media data. The successful development of such systems may
enable us to better understand how nonmedical use of substances affects
individual health and societal dynamics, aiding the development of effective
public health strategies. In addition to creating the annotated data set, we
applied several machine learning models to establish baseline performances.
Specifically, we experimented with transformer models like BERT, and RoBERTa,
one few-shot learning model DANN by leveraging the full training dataset, and
GPT-3.5 by using one-shot learning, for automatic NER of clinical and social
impacts. The dataset has been made available through the 2024 SMM4H shared
tasks. | cs.CL | Reddit-Impacts: A Named Entity Recognition Dataset for Analyzing Clinical and Social Effects of Substance Use Derived from Social Media |
2024-05-09T22:59:23Z | http://arxiv.org/pdf/2405.06134v1 | http://arxiv.org/abs/2405.06134v1 | Vyas Raina, Rao Ma, Charles McGhee, Kate Knill, Mark Gales | 2024-05-09T22:59:23Z | cs.CL, cs.SD, eess.AS | Recent developments in large speech foundation models like Whisper have led
to their widespread use in many automatic speech recognition (ASR)
applications. These systems incorporate `special tokens' in their vocabulary,
such as $\texttt{<endoftext>}$, to guide their language generation process.
However, we demonstrate that these tokens can be exploited by adversarial
attacks to manipulate the model's behavior. We propose a simple yet effective
method to learn a universal acoustic realization of Whisper's
$\texttt{<endoftext>}$ token, which, when prepended to any speech signal,
encourages the model to ignore the speech and only transcribe the special
token, effectively `muting' the model. Our experiments demonstrate that the
same, universal 0.64-second adversarial audio segment can successfully mute a
target Whisper ASR model for over 97\% of speech samples. Moreover, we find
that this universal adversarial audio segment often transfers to new datasets
and tasks. Overall this work demonstrates the vulnerability of Whisper models
to `muting' adversarial attacks, where such attacks can pose both risks and
potential benefits in real-world settings: for example the attack can be used
to bypass speech moderation systems, or conversely the attack can also be used
to protect private speech data. | cs.CL | Muting Whisper: A Universal Acoustic Adversarial Attack on Speech Foundation Models |
2024-05-09T22:21:40Z | http://arxiv.org/pdf/2405.06129v1 | http://arxiv.org/abs/2405.06129v1 | Saydeh N. Karabatis, Vandana P. Janeja | 2024-05-09T22:21:40Z | cs.IR, cs.CL | Climate change and political unrest in certain regions of the world are
imposing extreme hardship on many communities and are forcing millions of
vulnerable populations to abandon their homelands and seek refuge in safer
lands. As international laws are not fully set to deal with the migration
crisis, people are relying on networks of exploiting smugglers to escape the
devastation in order to live in stability. During the smuggling journey,
migrants can become victims of human trafficking if they fail to pay the
smuggler and may be forced into coerced labor. Government agencies and
anti-trafficking organizations try to identify the trafficking routes based on
stories of survivors in order to gain knowledge and help prevent such crimes.
In this paper, we propose a system called Narrative to Trajectory (N2T+), which
extracts trajectories of trafficking routes. N2T+ uses Data Science and Natural
Language Processing techniques to analyze trafficking narratives, automatically
extract relevant location names, disambiguate possible name ambiguities, and
plot the trafficking route on a map. In a comparative evaluation we show that
the proposed multi-dimensional approach offers significantly higher geolocation
detection than other state of the art techniques. | cs.IR | Narrative to Trajectory (N2T+): Extracting Routes of Life or Death from Human Trafficking Text Corpora |
2024-05-09T21:15:49Z | http://arxiv.org/pdf/2405.06105v1 | http://arxiv.org/abs/2405.06105v1 | Yutong Hu, Quzhe Huang, Mingxu Tao, Chen Zhang, Yansong Feng | 2024-05-09T21:15:49Z | cs.CL | Recent studies have shown that Large Language Models (LLMs) have the
potential to process extremely long text. Many works only evaluate LLMs'
long-text processing ability on the language modeling task, with perplexity
(PPL) as the evaluation metric. However, in our study, we find that there is no
correlation between PPL and LLMs' long-text understanding ability. Besides, PPL
may only reflect the model's ability to model local information instead of
catching long-range dependency. Therefore, only using PPL to prove the model
could process long text is inappropriate. The local focus feature of PPL could
also explain some existing phenomena, such as the great extrapolation ability
of the position method ALiBi. When evaluating a model's ability in long text,
we might pay more attention to PPL's limitation and avoid overly relying on it. | cs.CL | Can Perplexity Reflect Large Language Model's Ability in Long Text Understanding? |
2024-05-09T20:45:58Z | http://arxiv.org/pdf/2405.06093v1 | http://arxiv.org/abs/2405.06093v1 | Bhawesh Kumar, Jonathan Amar, Eric Yang, Nan Li, Yugang Jia | 2024-05-09T20:45:58Z | cs.LG, cs.CL | Large Language Models (LLMs) have demonstrated their efficacy across a broad
spectrum of tasks in healthcare applications. However, often LLMs need to be
fine-tuned on task-specific expert annotated data to achieve optimal
performance, which can be expensive and time consuming. In this study, we
fine-tune PaLM-2 with parameter efficient fine-tuning (PEFT) using noisy labels
obtained from gemini-pro 1.0 for the detection of Schedule-of-Event (SoE)
tables, which specify care plan in clinical trial protocols. We introduce a
filtering mechanism to select high-confidence labels for this table
classification task, thereby reducing the noise in the auto-generated labels.
We show that fine-tuned PaLM-2 with those labels achieves performance that
exceeds the gemini-pro 1.0 and other LLMs. Furthermore, its performance is
close to a PaLM-2 fine-tuned on labels obtained from non-expert annotators. Our
results show that leveraging LLM-generated labels through powerful models like
gemini-pro can potentially serve as a viable strategy for improving LLM
performance through fine-tuning in specialized tasks, particularly in domains
where expert annotations are scarce, expensive, or time-consuming to obtain. | cs.LG | Selective Fine-tuning on LLM-labeled Data May Reduce Reliance on Human Annotation: A Case Study Using Schedule-of-Event Table Detection |
2024-05-09T19:32:49Z | http://arxiv.org/pdf/2405.06067v2 | http://arxiv.org/abs/2405.06067v2 | Zifan He, Zongyue Qin, Neha Prakriya, Yizhou Sun, Jason Cong | 2024-05-14T06:09:52Z | cs.CL, cs.LG | Transformer-based large language models (LLM) have been widely used in
language processing applications. However, most of them restrict the context
window that permits the model to attend to every token in the inputs. Previous
works in recurrent models can memorize past tokens to enable unlimited context
and maintain effectiveness. However, they have "flat" memory architectures,
which have limitations in selecting and filtering information. Since humans are
good at learning and self-adjustment, we speculate that imitating brain memory
hierarchy is beneficial for model memorization. We propose the Hierarchical
Memory Transformer (HMT), a novel framework that enables and improves models'
long-context processing ability by imitating human memorization behavior.
Leveraging memory-augmented segment-level recurrence, we organize the memory
hierarchy by preserving tokens from early input token segments, passing memory
embeddings along the sequence, and recalling relevant information from history.
Evaluating general language modeling (Wikitext-103, PG-19) and
question-answering tasks (PubMedQA), we show that HMT steadily improves the
long-context processing ability of context-constrained and long-context models.
With an additional 0.5% - 2% of parameters, HMT can easily plug in and augment
future LLMs to handle long context effectively. Our code is open-sourced on
Github: https://github.com/OswaldHe/HMT-pytorch. | cs.CL | HMT: Hierarchical Memory Transformer for Long Context Language Processing |
2024-05-09T19:17:47Z | http://arxiv.org/pdf/2405.06064v1 | http://arxiv.org/abs/2405.06064v1 | Alexandra Zytek, Sara Pidò, Kalyan Veeramachaneni | 2024-05-09T19:17:47Z | cs.AI, cs.CL, cs.HC, cs.LG | In response to the demand for Explainable Artificial Intelligence (XAI), we
investigate the use of Large Language Models (LLMs) to transform ML
explanations into natural, human-readable narratives. Rather than directly
explaining ML models using LLMs, we focus on refining explanations computed
using existing XAI algorithms. We outline several research directions,
including defining evaluation metrics, prompt design, comparing LLM models,
exploring further training methods, and integrating external data. Initial
experiments and user study suggest that LLMs offer a promising way to enhance
the interpretability and usability of XAI. | cs.AI | LLMs for XAI: Future Directions for Explaining Explanations |
2024-05-09T19:02:56Z | http://arxiv.org/pdf/2405.06059v1 | http://arxiv.org/abs/2405.06059v1 | Christopher Z. Cui, Xiangyu Peng, Mark O. Riedl | 2024-05-09T19:02:56Z | cs.CL, cs.AI | Open-ended worlds are those in which there are no pre-specified goals or
environmental reward signal. As a consequence, an agent must know how to
perform a multitude of tasks. However, when a new task is presented to an
agent, we expect it to be able to reuse some of what it knows from previous
tasks to rapidly learn that new task. We introduce a novel technique whereby
policies for different a priori known tasks are combined into a
Mixture-of-Experts model with an attention mechanism across a mix of frozen and
unfrozen experts. The model learns when to attend to frozen task-specific
experts when appropriate and learns new experts to handle novel situations. We
work in an open-ended text-based environment in which the agent is tasked with
behaving like different types of character roles and must rapidly learn
behaviors associated with new character role types. We show that our agent both
obtains more rewards in the zero-shot setting, and discovers these rewards with
greater sample efficiency in the few-shot learning settings. | cs.CL | A Mixture-of-Experts Approach to Few-Shot Task Transfer in Open-Ended Text Worlds |
2024-05-09T19:02:53Z | http://arxiv.org/pdf/2405.06058v1 | http://arxiv.org/abs/2405.06058v1 | Aadesh Salecha, Molly E. Ireland, Shashanka Subrahmanya, João Sedoc, Lyle H. Ungar, Johannes C. Eichstaedt | 2024-05-09T19:02:53Z | cs.AI, cs.CL, cs.CY, cs.HC | As Large Language Models (LLMs) become widely used to model and simulate
human behavior, understanding their biases becomes critical. We developed an
experimental framework using Big Five personality surveys and uncovered a
previously undetected social desirability bias in a wide range of LLMs. By
systematically varying the number of questions LLMs were exposed to, we
demonstrate their ability to infer when they are being evaluated. When
personality evaluation is inferred, LLMs skew their scores towards the
desirable ends of trait dimensions (i.e., increased extraversion, decreased
neuroticism, etc). This bias exists in all tested models, including GPT-4/3.5,
Claude 3, Llama 3, and PaLM-2. Bias levels appear to increase in more recent
models, with GPT-4's survey responses changing by 1.20 (human) standard
deviations and Llama 3's by 0.98 standard deviations-very large effects. This
bias is robust to randomization of question order and paraphrasing.
Reverse-coding all the questions decreases bias levels but does not eliminate
them, suggesting that this effect cannot be attributed to acquiescence bias.
Our findings reveal an emergent social desirability bias and suggest
constraints on profiling LLMs with psychometric tests and on using LLMs as
proxies for human participants. | cs.AI | Large Language Models Show Human-like Social Desirability Biases in Survey Responses |
2024-05-09T17:59:32Z | http://arxiv.org/pdf/2405.05966v1 | http://arxiv.org/abs/2405.05966v1 | Juri Opitz, Shira Wein, Nathan Schneider | 2024-05-09T17:59:32Z | cs.CL, cs.AI | Large Language Models (LLMs) have become capable of generating highly fluent
text in certain languages, without modules specially designed to capture
grammar or semantic coherence. What does this mean for the future of linguistic
expertise in NLP? We highlight several aspects in which NLP (still) relies on
linguistics, or where linguistic thinking can illuminate new directions. We
argue our case around the acronym $RELIES$ that encapsulates six major facets
where linguistics contributes to NLP: $R$esources, $E$valuation, $L$ow-resource
settings, $I$nterpretability, $E$xplanation, and the $S$tudy of language. This
list is not exhaustive, nor is linguistics the main point of reference for
every effort under these themes; but at a macro level, these facets highlight
the enduring importance of studying machine systems vis-a-vis systems of human
language. | cs.CL | Natural Language Processing RELIES on Linguistics |
2024-05-09T17:53:28Z | http://arxiv.org/pdf/2405.05957v1 | http://arxiv.org/abs/2405.05957v1 | Dan Qiao, Yi Su, Pinzheng Wang, Jing Ye, Wenjing Xie, Yuechi Zhou, Yuyang Ding, Zecheng Tang, Jikai Wang, Yixin Ji, Yue Wang, Pei Guo, Zechen Sun, Zikang Zhang, Juntao Li, Pingfu Chao, Wenliang Chen, Guohong Fu, Guodong Zhou, Qiaoming Zhu, Min Zhang | 2024-05-09T17:53:28Z | cs.CL | Large Language Models (LLMs) have played an important role in many fields due
to their powerful capabilities.However, their massive number of parameters
leads to high deployment requirements and incurs significant inference costs,
which impedes their practical applications. Training smaller models is an
effective way to address this problem. Therefore, we introduce OpenBA-V2, a
3.4B model derived from multi-stage compression and continual pre-training from
the original 15B OpenBA model. OpenBA-V2 utilizes more data, more flexible
training objectives, and techniques such as layer pruning, neural pruning, and
vocabulary pruning to achieve a compression rate of 77.3\% with minimal
performance loss. OpenBA-V2 demonstrates competitive performance compared to
other open-source models of similar size, achieving results close to or on par
with the 15B OpenBA model in downstream tasks such as common sense reasoning
and Named Entity Recognition (NER). OpenBA-V2 illustrates that LLMs can be
compressed into smaller ones with minimal performance loss by employing
advanced training objectives and data strategies, which may help deploy LLMs in
resource-limited scenarios. | cs.CL | OpenBA-V2: Reaching 77.3% High Compression Ratio with Fast Multi-Stage Pruning |
2024-05-09T17:49:04Z | http://arxiv.org/pdf/2405.05955v3 | http://arxiv.org/abs/2405.05955v3 | Junzhi Chen, Juhao Liang, Benyou Wang | 2024-06-24T02:44:57Z | cs.CL | The emergence of large language models (LLMs) has opened up unprecedented
possibilities for automating complex tasks that are often comparable to human
performance. Despite their capabilities, LLMs still encounter difficulties in
completing tasks that require high levels of accuracy and complexity due to
their inherent limitations in handling multifaceted problems single-handedly.
This paper introduces `Smurfs', a cutting-edge multi-agent framework designed
to revolutionize the application of LLMs. By seamlessly transforming a
conventional LLM into a synergistic multi-agent ensemble, Smurfs can enhance
the model's ability to solve complex tasks at no additional cost. This is
achieved through innovative prompting strategies that allocate distinct roles
within the model, thereby facilitating collaboration among specialized agents
and forming an intelligent multi-agent system. Our empirical investigation on
both open-ended task of StableToolBench and closed-ended task on HotpotQA
showcases Smurfs' superior capability in intricate tool utilization scenarios.
Notably, Smurfs outmatches all the baseline methods in both experiments,
setting new state-of-the-art performance. Furthermore, through comprehensive
ablation studies, we dissect the contribution of the core components of the
multi-agent framework to its overall efficacy. This not only verifies the
effectiveness of the framework, but also sets a route for future exploration of
multi-agent LLM systems. | cs.CL | Smurfs: Leveraging Multiple Proficiency Agents with Context-Efficiency for Tool Planning |
2024-05-09T17:25:31Z | http://arxiv.org/pdf/2405.05938v2 | http://arxiv.org/abs/2405.05938v2 | Chaitanya Malaviya, Priyanka Agrawal, Kuzman Ganchev, Pranesh Srinivasan, Fantine Huot, Jonathan Berant, Mark Yatskar, Dipanjan Das, Mirella Lapata, Chris Alberti | 2024-05-28T17:53:48Z | cs.CL | Experts in various fields routinely perform methodical writing tasks to plan,
organize, and report their work. From a clinician writing a differential
diagnosis for a patient, to a teacher writing a lesson plan for students, these
tasks are pervasive, requiring to methodically generate structured long-form
output for a given input. We develop a typology of methodical tasks structured
in the form of a task objective, procedure, input, and output, and introduce
DoLoMiTes, a novel benchmark with specifications for 519 such tasks elicited
from hundreds of experts from across 25 fields. Our benchmark further contains
specific instantiations of methodical tasks with concrete input and output
examples (1,857 in total) which we obtain by collecting expert revisions of up
to 10 model-generated examples of each task. We use these examples to evaluate
contemporary language models highlighting that automating methodical tasks is a
challenging long-form generation problem, as it requires performing complex
inferences, while drawing upon the given context as well as domain knowledge. | cs.CL | DOLOMITES: Domain-Specific Long-Form Methodical Tasks |
2024-05-09T17:00:22Z | http://arxiv.org/pdf/2405.05904v2 | http://arxiv.org/abs/2405.05904v2 | Zorik Gekhman, Gal Yona, Roee Aharoni, Matan Eyal, Amir Feder, Roi Reichart, Jonathan Herzig | 2024-05-13T07:29:58Z | cs.CL | When large language models are aligned via supervised fine-tuning, they may
encounter new factual information that was not acquired through pre-training.
It is often conjectured that this can teach the model the behavior of
hallucinating factually incorrect responses, as the model is trained to
generate facts that are not grounded in its pre-existing knowledge. In this
work, we study the impact of such exposure to new knowledge on the capability
of the fine-tuned model to utilize its pre-existing knowledge. To this end, we
design a controlled setup, focused on closed-book QA, where we vary the
proportion of the fine-tuning examples that introduce new knowledge. We
demonstrate that large language models struggle to acquire new factual
knowledge through fine-tuning, as fine-tuning examples that introduce new
knowledge are learned significantly slower than those consistent with the
model's knowledge. However, we also find that as the examples with new
knowledge are eventually learned, they linearly increase the model's tendency
to hallucinate. Taken together, our results highlight the risk in introducing
new factual knowledge through fine-tuning, and support the view that large
language models mostly acquire factual knowledge through pre-training, whereas
fine-tuning teaches them to use it more efficiently. | cs.CL | Does Fine-Tuning LLMs on New Knowledge Encourage Hallucinations? |
2024-05-09T16:45:27Z | http://arxiv.org/pdf/2405.05894v2 | http://arxiv.org/abs/2405.05894v2 | Adian Liusie, Vatsal Raina, Yassir Fathullah, Mark Gales | 2024-06-09T17:56:11Z | cs.CL | LLM-as-a-judge approaches are a practical and effective way of assessing a
range of text tasks, aligning with human judgements especially when applied in
a comparative assessment fashion. However, when using pairwise comparisons to
rank a set of candidates the computational costs scale quadratically with the
number of candidates, which can have practical limitations. This paper
introduces a Product of Expert (PoE) framework for efficient LLM Comparative
Assessment. Here individual comparisons are considered experts that provide
information on a pair's score difference. The PoE framework combines the
information from these experts to yield an expression that can be maximized
with respect to the underlying set of candidates, and is highly flexible where
any form of expert can be assumed. When Gaussian experts are used one can
derive simple closed-form solutions for the optimal candidate ranking, as well
as expressions for selecting which comparisons should be made to maximize the
probability of this ranking. Our approach enables efficient comparative
assessment, where by using only a small subset of the possible comparisons, one
can generate score predictions that correlate as well to human judgements as
the predictions when all comparisons are used. We evaluate the approach on
multiple NLG tasks and demonstrate that our framework can yield considerable
computational savings when performing pairwise comparative assessment. When N
is large, with as few as 2% of comparisons the PoE solution can achieve similar
performance to when all comparisons are used. | cs.CL | Efficient LLM Comparative Assessment: a Product of Experts Framework for Pairwise Comparisons |
2024-05-09T16:42:29Z | http://arxiv.org/pdf/2405.06715v1 | http://arxiv.org/abs/2405.06715v1 | Pronita Mehrotra, Aishni Parab, Sumit Gulwani | 2024-05-09T16:42:29Z | cs.CL, cs.AI | This paper explores the enhancement of creativity in Large Language Models
(LLMs) like vGPT-4 through associative thinking, a cognitive process where
creative ideas emerge from linking seemingly unrelated concepts. Associative
thinking strategies have been found to effectively help humans boost
creativity. However, whether the same strategies can help LLMs become more
creative remains under-explored. In this work, we investigate whether prompting
LLMs to connect disparate concepts can augment their creative outputs. Focusing
on three domains -- Product Design, Storytelling, and Marketing -- we introduce
creativity tasks designed to assess vGPT-4's ability to generate original and
useful content. By challenging the models to form novel associations, we
evaluate the potential of associative thinking to enhance the creative
capabilities of LLMs. Our findings show that leveraging associative thinking
techniques can significantly improve the originality of vGPT-4's responses. | cs.CL | Enhancing Creativity in Large Language Models through Associative Thinking Strategies |
2024-05-09T16:36:56Z | http://arxiv.org/pdf/2405.06714v2 | http://arxiv.org/abs/2405.06714v2 | David Heineman, Reba Koenen, Sashank Varma | 2024-05-14T02:10:05Z | cs.CL, cs.AI | Category fluency is a widely studied cognitive phenomenon, yet two
conflicting accounts have been proposed as the underlying retrieval mechanism
-- an optimal foraging process deliberately searching through memory (Hills et
al., 2012) and a random walk sampling from a semantic network (Abbott et al.,
2015). Evidence for both accounts has centered around predicting human patch
switches, where both existing models of category fluency produce paradoxically
identical results. We begin by peeling back the assumptions made by existing
models, namely that each named example only depends on the previous example, by
(i) adding an additional bias to model the category transition probability
directly and (ii) relying on a large language model to predict based on the
entire existing sequence. Then, we present evidence towards resolving the
disagreement between each account of foraging by reformulating models as
sequence generators. To evaluate, we compare generated category fluency runs to
a bank of human-written sequences by proposing a metric based on n-gram
overlap. We find category switch predictors do not necessarily produce
human-like sequences, in fact the additional biases used by the Hills et al.
(2012) model are required to improve generation quality, which are later
improved by our category modification. Even generating exclusively with an LLM
requires an additional global cue to trigger the patch switching behavior
during production. Further tests on only the search process on top of the
semantic network highlight the importance of deterministic search to replicate
human behavior. | cs.CL | Towards a Path Dependent Account of Category Fluency |
2024-05-09T15:48:07Z | http://arxiv.org/pdf/2405.05860v1 | http://arxiv.org/abs/2405.05860v1 | Eve Fleisig, Su Lin Blodgett, Dan Klein, Zeerak Talat | 2024-05-09T15:48:07Z | cs.LG, cs.CL, cs.CY | Longstanding data labeling practices in machine learning involve collecting
and aggregating labels from multiple annotators. But what should we do when
annotators disagree? Though annotator disagreement has long been seen as a
problem to minimize, new perspectivist approaches challenge this assumption by
treating disagreement as a valuable source of information. In this position
paper, we examine practices and assumptions surrounding the causes of
disagreement--some challenged by perspectivist approaches, and some that remain
to be addressed--as well as practical and normative challenges for work
operating under these assumptions. We conclude with recommendations for the
data labeling pipeline and avenues for future research engaging with
subjectivity and disagreement. | cs.LG | The Perspectivist Paradigm Shift: Assumptions and Challenges of Capturing Human Labels |
2024-05-09T15:39:54Z | http://arxiv.org/pdf/2405.05852v1 | http://arxiv.org/abs/2405.05852v1 | Gunshi Gupta, Karmesh Yadav, Yarin Gal, Dhruv Batra, Zsolt Kira, Cong Lu, Tim G. J. Rudner | 2024-05-09T15:39:54Z | cs.CV, cs.AI, cs.CL, cs.LG, cs.RO, stat.ML | Embodied AI agents require a fine-grained understanding of the physical world
mediated through visual and language inputs. Such capabilities are difficult to
learn solely from task-specific data. This has led to the emergence of
pre-trained vision-language models as a tool for transferring representations
learned from internet-scale data to downstream tasks and new domains. However,
commonly used contrastively trained representations such as in CLIP have been
shown to fail at enabling embodied agents to gain a sufficiently fine-grained
scene understanding -- a capability vital for control. To address this
shortcoming, we consider representations from pre-trained text-to-image
diffusion models, which are explicitly optimized to generate images from text
prompts and as such, contain text-conditioned representations that reflect
highly fine-grained visuo-spatial information. Using pre-trained text-to-image
diffusion models, we construct Stable Control Representations which allow
learning downstream control policies that generalize to complex, open-ended
environments. We show that policies learned using Stable Control
Representations are competitive with state-of-the-art representation learning
approaches across a broad range of simulated control settings, encompassing
challenging manipulation and navigation tasks. Most notably, we show that
Stable Control Representations enable learning policies that exhibit
state-of-the-art performance on OVMM, a difficult open-vocabulary navigation
benchmark. | cs.CV | Pre-trained Text-to-Image Diffusion Models Are Versatile Representation Learners for Control |
2024-05-09T15:39:19Z | http://arxiv.org/pdf/2405.06713v2 | http://arxiv.org/abs/2405.06713v2 | Zhenhui Jiang, Jiaxin Li, Yang Liu | 2024-05-21T05:37:58Z | cs.CL, cs.AI | The strategic significance of Large Language Models (LLMs) in economic
expansion, innovation, societal development, and national security has been
increasingly recognized since the advent of ChatGPT. This study provides a
comprehensive comparative evaluation of American and Chinese LLMs in both
English and Chinese contexts. We proposed a comprehensive evaluation framework
that encompasses natural language proficiency, disciplinary expertise, and
safety and responsibility, and systematically assessed 16 prominent models from
the US and China under various operational tasks and scenarios. Our key
findings show that GPT 4-Turbo is at the forefront in English contexts, whereas
Ernie-Bot 4 stands out in Chinese contexts. The study also highlights
disparities in LLM performance across languages and tasks, stressing the
necessity for linguistically and culturally nuanced model development. The
complementary strengths of American and Chinese LLMs point to the value of
Sino-US collaboration in advancing LLM technology. The research presents the
current LLM competition landscape and offers valuable insights for policymakers
and businesses regarding strategic LLM investments and development. Future work
will expand on this framework to include emerging LLM multimodal capabilities
and business application assessments. | cs.CL | Unveiling the Competitive Dynamics: A Comparative Evaluation of American and Chinese LLMs |
2024-05-09T15:12:24Z | http://arxiv.org/pdf/2405.06712v1 | http://arxiv.org/abs/2405.06712v1 | Gaurav Kumar Gupta, Aditi Singh, Sijo Valayakkad Manikandan, Abul Ehtesham | 2024-05-09T15:12:24Z | cs.CL, cs.AI | The recent swift development of LLMs like GPT-4, Gemini, and GPT-3.5 offers a
transformative opportunity in medicine and healthcare, especially in digital
diagnostics. This study evaluates each model diagnostic abilities by
interpreting a user symptoms and determining diagnoses that fit well with
common illnesses, and it demonstrates how each of these models could
significantly increase diagnostic accuracy and efficiency. Through a series of
diagnostic prompts based on symptoms from medical databases, GPT-4 demonstrates
higher diagnostic accuracy from its deep and complete history of training on
medical data. Meanwhile, Gemini performs with high precision as a critical tool
in disease triage, demonstrating its potential to be a reliable model when
physicians are trying to make high-risk diagnoses. GPT-3.5, though slightly
less advanced, is a good tool for medical diagnostics. This study highlights
the need to study LLMs for healthcare and clinical practices with more care and
attention, ensuring that any system utilizing LLMs promotes patient privacy and
complies with health information privacy laws such as HIPAA compliance, as well
as the social consequences that affect the varied individuals in complex
healthcare contexts. This study marks the start of a larger future effort to
study the various ways in which assigning ethical concerns to LLMs task of
learning from human biases could unearth new ways to apply AI in complex
medical settings. | cs.CL | Digital Diagnostics: The Potential Of Large Language Models In Recognizing Symptoms Of Common Illnesses |
2024-05-09T13:54:22Z | http://arxiv.org/pdf/2405.05777v1 | http://arxiv.org/abs/2405.05777v1 | Ronny Paul, Himanshu Buckchash, Shantipriya Parida, Dilip K. Prasad | 2024-05-09T13:54:22Z | cs.CL, cs.AI | S\'ami, an indigenous language group comprising multiple languages, faces
digital marginalization due to the limited availability of data and
sophisticated language models designed for its linguistic intricacies. This
work focuses on increasing technological participation for the S\'ami language.
We draw the attention of the ML community towards the language modeling problem
of Ultra Low Resource (ULR) languages. ULR languages are those for which the
amount of available textual resources is very low, and the speaker count for
them is also very low. ULRLs are also not supported by mainstream Large
Language Models (LLMs) like ChatGPT, due to which gathering artificial training
data for them becomes even more challenging. Mainstream AI foundational model
development has given less attention to this category of languages. Generally,
these languages have very few speakers, making it hard to find them. However,
it is important to develop foundational models for these ULR languages to
promote inclusion and the tangible abilities and impact of LLMs. To this end,
we have compiled the available S\'ami language resources from the web to create
a clean dataset for training language models. In order to study the behavior of
modern LLM models with ULR languages (S\'ami), we have experimented with
different kinds of LLMs, mainly at the order of $\sim$ seven billion
parameters. We have also explored the effect of multilingual LLM training for
ULRLs. We found that the decoder-only models under a sequential multilingual
training scenario perform better than joint multilingual training, whereas
multilingual training with high semantic overlap, in general, performs better
than training from scratch.This is the first study on the S\'ami language for
adapting non-statistical language models that use the latest developments in
the field of natural language processing (NLP). | cs.CL | Towards a More Inclusive AI: Progress and Perspectives in Large Language Model Training for the Sámi Language |
2024-05-09T13:54:15Z | http://arxiv.org/pdf/2405.05776v1 | http://arxiv.org/abs/2405.05776v1 | Polina Tsvilodub, Paul Marty, Sonia Ramotowska, Jacopo Romoli, Michael Franke | 2024-05-09T13:54:15Z | cs.CL | Human communication is based on a variety of inferences that we draw from
sentences, often going beyond what is literally said. While there is wide
agreement on the basic distinction between entailment, implicature, and
presupposition, the status of many inferences remains controversial. In this
paper, we focus on three inferences of plain and embedded disjunctions, and
compare them with regular scalar implicatures. We investigate this comparison
from the novel perspective of the predictions of state-of-the-art large
language models, using the same experimental paradigms as recent studies
investigating the same inferences with humans. The results of our best
performing models mostly align with those of humans, both in the large
differences we find between those inferences and implicatures, as well as in
fine-grained distinctions among different aspects of those inferences. | cs.CL | Experimental Pragmatics with Machines: Testing LLM Predictions for the Inferences of Plain and Embedded Disjunctions |
2024-05-09T13:32:26Z | http://arxiv.org/pdf/2405.05760v2 | http://arxiv.org/abs/2405.05760v2 | Zhizhen Zhang, Ning Wang, Haojie Li, Zhihui Wang | 2024-06-23T10:05:18Z | cs.CV, cs.CL | Semantic location prediction aims to derive meaningful location insights from
multimodal social media posts, offering a more contextual understanding of
daily activities than using GPS coordinates. This task faces significant
challenges due to the noise and modality heterogeneity in "text-image" posts.
Existing methods are generally constrained by inadequate feature
representations and modal interaction, struggling to effectively reduce noise
and modality heterogeneity. To address these challenges, we propose a
Similarity-Guided Multimodal Fusion Transformer (SG-MFT) for predicting the
semantic locations of users from their multimodal posts. First, we incorporate
high-quality text and image representations by utilizing a pre-trained large
vision-language model. Then, we devise a Similarity-Guided Interaction Module
(SIM) to alleviate modality heterogeneity and noise interference by
incorporating both coarse-grained and fine-grained similarity guidance for
improving modality interactions. Specifically, we propose a novel
similarity-aware feature interpolation attention mechanism at the
coarse-grained level, leveraging modality-wise similarity to mitigate
heterogeneity and reduce noise within each modality. At the fine-grained level,
we utilize a similarity-aware feed-forward block and element-wise similarity to
further address the issue of modality heterogeneity. Finally, building upon
pre-processed features with minimal noise and modal interference, we devise a
Similarity-aware Fusion Module (SFM) to fuse two modalities with a
cross-attention mechanism. Comprehensive experimental results clearly
demonstrate the superior performance of our proposed method. | cs.CV | Similarity Guided Multimodal Fusion Transformer for Semantic Location Prediction in Social Media |
2024-05-09T13:27:22Z | http://arxiv.org/pdf/2405.05758v1 | http://arxiv.org/abs/2405.05758v1 | Han Meng, Yitian Yang, Yunan Li, Jungup Lee, Yi-Chieh Lee | 2024-05-09T13:27:22Z | cs.HC, cs.CL, cs.CY | Qualitative analysis is a challenging, yet crucial aspect of advancing
research in the field of Human-Computer Interaction (HCI). Recent studies show
that large language models (LLMs) can perform qualitative coding within
existing schemes, but their potential for collaborative human-LLM discovery and
new insight generation in qualitative analysis is still underexplored. To
bridge this gap and advance qualitative analysis by harnessing the power of
LLMs, we propose CHALET, a novel methodology that leverages the human-LLM
collaboration paradigm to facilitate conceptualization and empower qualitative
research. The CHALET approach involves LLM-supported data collection,
performing both human and LLM deductive coding to identify disagreements, and
performing collaborative inductive coding on these disagreement cases to derive
new conceptual insights. We validated the effectiveness of CHALET through its
application to the attribution model of mental-illness stigma, uncovering
implicit stigmatization themes on cognitive, emotional and behavioral
dimensions. We discuss the implications for future research, methodology, and
the transdisciplinary opportunities CHALET presents for the HCI community and
beyond. | cs.HC | Exploring the Potential of Human-LLM Synergy in Advancing Qualitative Analysis: A Case Study on Mental-Illness Stigma |
2024-05-09T12:58:22Z | http://arxiv.org/pdf/2405.05741v1 | http://arxiv.org/abs/2405.05741v1 | Jinyang Wu, Feihu Che, Xinxin Zheng, Shuai Zhang, Ruihan Jin, Shuai Nie, Pengpeng Shao, Jianhua Tao | 2024-05-09T12:58:22Z | cs.CL, cs.AI | Large language models (LLMs) like ChatGPT have shown significant advancements
across diverse natural language understanding (NLU) tasks, including
intelligent dialogue and autonomous agents. Yet, lacking widely acknowledged
testing mechanisms, answering `whether LLMs are stochastic parrots or genuinely
comprehend the world' remains unclear, fostering numerous studies and sparking
heated debates. Prevailing research mainly focuses on surface-level NLU,
neglecting fine-grained explorations. However, such explorations are crucial
for understanding their unique comprehension mechanisms, aligning with human
cognition, and finally enhancing LLMs' general NLU capacities. To address this
gap, our study delves into LLMs' nuanced semantic comprehension capabilities,
particularly regarding common words with uncommon meanings. The idea stems from
foundational principles of human communication within psychology, which
underscore accurate shared understandings of word semantics. Specifically, this
paper presents the innovative construction of a Lexical Semantic Comprehension
(LeSC) dataset with novel evaluation metrics, the first benchmark encompassing
both fine-grained and cross-lingual dimensions. Introducing models of both
open-source and closed-source, varied scales and architectures, our extensive
empirical experiments demonstrate the inferior performance of existing models
in this basic lexical-meaning understanding task. Notably, even the
state-of-the-art LLMs GPT-4 and GPT-3.5 lag behind 16-year-old humans by 3.9%
and 22.3%, respectively. Additionally, multiple advanced prompting techniques
and retrieval-augmented generation are also introduced to help alleviate this
trouble, yet limitations persist. By highlighting the above critical
shortcomings, this research motivates further investigation and offers novel
insights for developing more intelligent LLMs. | cs.CL | Can large language models understand uncommon meanings of common words? |
2024-05-09T12:39:50Z | http://arxiv.org/pdf/2405.06710v1 | http://arxiv.org/abs/2405.06710v1 | Cem Bozsahin | 2024-05-09T12:39:50Z | cs.CL, cs.AI, cs.CY | The article is an attempt to contribute to explorations of a common origin
for language and planned-collaborative action. It gives `semantics of change'
the central stage in the synthesis, from its history and recordkeeping to its
development, its syntax, delivery and reception, including substratal aspects.
It is suggested that to arrive at a common core, linguistic semantics must be
understood as studying through syntax mobile agent's representing, tracking and
coping with change and no change. Semantics of actions can be conceived the
same way, but through plans instead of syntax. The key point is the following:
Sequencing itself, of words and action sequences, brings in more structural
interpretation to the sequence than which is immediately evident from the
sequents themselves. Mobile sequencers can be understood as subjects
structuring reporting, understanding and keeping track of change and no change.
The idea invites rethinking of the notion of category, both in language and in
planning.
Understanding understanding change by mobile agents is suggested to be about
human extended practice, not extended-human practice. That's why linguistics is
as important as computer science in the synthesis. It must rely on
representational history of acts, thoughts and expressions, personal and
public, crosscutting overtness and covertness of these phenomena. It has
implication for anthropology in the extended practice, which is covered
briefly. | cs.CL | Mobile Sequencers |
2024-05-09T12:35:33Z | http://arxiv.org/pdf/2405.05723v1 | http://arxiv.org/abs/2405.05723v1 | Pablo Rosillo-Rodes, Maxi San Miguel, David Sanchez | 2024-05-09T12:35:33Z | cs.CL, cs.AI, cs.IR | Flamenco, recognized by UNESCO as part of the Intangible Cultural Heritage of
Humanity, is a profound expression of cultural identity rooted in Andalusia,
Spain. However, there is a lack of quantitative studies that help identify
characteristic patterns in this long-lived music tradition. In this work, we
present a computational analysis of Flamenco lyrics, employing natural language
processing and machine learning to categorize over 2000 lyrics into their
respective Flamenco genres, termed as $\textit{palos}$. Using a Multinomial
Naive Bayes classifier, we find that lexical variation across styles enables to
accurately identify distinct $\textit{palos}$. More importantly, from an
automatic method of word usage, we obtain the semantic fields that characterize
each style. Further, applying a metric that quantifies the inter-genre distance
we perform a network analysis that sheds light on the relationship between
Flamenco styles. Remarkably, our results suggest historical connections and
$\textit{palo}$ evolutions. Overall, our work illuminates the intricate
relationships and cultural significance embedded within Flamenco lyrics,
complementing previous qualitative discussions with quantitative analyses and
sparking new discussions on the origin and development of traditional music
genres. | cs.CL | Computational lexical analysis of Flamenco genres |
2024-05-09T12:03:38Z | http://arxiv.org/pdf/2405.05705v1 | http://arxiv.org/abs/2405.05705v1 | Sandrine Chausson, Björn Ross | 2024-05-09T12:03:38Z | cs.CL | Many tasks related to Computational Social Science and Web Content Analysis
involve classifying pieces of text based on the claims they contain.
State-of-the-art approaches usually involve fine-tuning models on large
annotated datasets, which are costly to produce. In light of this, we propose
and release a qualitative and versatile few-shot learning methodology as a
common paradigm for any claim-based textual classification task. This
methodology involves defining the classes as arbitrarily sophisticated
taxonomies of claims, and using Natural Language Inference models to obtain the
textual entailment between these and a corpus of interest. The performance of
these models is then boosted by annotating a minimal sample of data points,
dynamically sampled using the well-established statistical heuristic of
Probabilistic Bisection. We illustrate this methodology in the context of three
tasks: climate change contrarianism detection, topic/stance classification and
depression-relates symptoms detection. This approach rivals traditional
pre-train/fine-tune approaches while drastically reducing the need for data
annotation. | cs.CL | Detecting Statements in Text: A Domain-Agnostic Few-Shot Solution |
2024-05-09T11:49:05Z | http://arxiv.org/pdf/2405.06001v1 | http://arxiv.org/abs/2405.06001v1 | Ruihao Gong, Yang Yong, Shiqiao Gu, Yushi Huang, Yunchen Zhang, Xianglong Liu, Dacheng Tao | 2024-05-09T11:49:05Z | cs.LG, cs.AI, cs.CL | Recent advancements in large language models (LLMs) are propelling us toward
artificial general intelligence, thanks to their remarkable emergent abilities
and reasoning capabilities. However, the substantial computational and memory
requirements of LLMs limit their widespread adoption. Quan- tization, a key
compression technique, offers a viable solution to mitigate these demands by
compressing and accelerating LLMs, albeit with poten- tial risks to model
accuracy. Numerous studies have aimed to minimize the accuracy loss associated
with quantization. However, the quantization configurations in these studies
vary and may not be optimized for hard- ware compatibility. In this paper, we
focus on identifying the most effective practices for quantizing LLMs, with the
goal of balancing performance with computational efficiency. For a fair
analysis, we develop a quantization toolkit LLMC, and design four crucial
principles considering the inference efficiency, quantized accuracy,
calibration cost, and modularization. By benchmarking on various models and
datasets with over 500 experiments, three takeaways corresponding to
calibration data, quantization algorithm, and quantization schemes are derived.
Finally, a best practice of LLM PTQ pipeline is constructed. All the benchmark
results and the toolkit can be found at https://github.com/ModelTC/llmc. | cs.LG | LLM-QBench: A Benchmark Towards the Best Practice for Post-training Quantization of Large Language Models |
2024-05-09T11:38:23Z | http://arxiv.org/pdf/2405.05688v1 | http://arxiv.org/abs/2405.05688v1 | Dipankar Srirag, Aditya Joshi | 2024-05-09T11:38:23Z | cs.CL | With an evergrowing number of LLMs reporting superlative performance for
English, their ability to perform equitably for different dialects of English
(i.e., dialect robustness) needs to be ascertained. Specifically, we use
English language (US English or Indian English) conversations between humans
who play the word-guessing game of `taboo'. We formulate two evaluative tasks:
target word prediction (TWP) (i.e.predict the masked target word in a
conversation) and target word selection (TWS) (i.e., select the most likely
masked target word in a conversation, from among a set of candidate words).
Extending MD3, an existing dialectic dataset of taboo-playing conversations, we
introduce M-MD3, a target-word-masked version of MD3 with the USEng and IndEng
subsets. We add two subsets: AITrans (where dialectic information is removed
from IndEng) and AIGen (where LLMs are prompted to generate conversations). Our
evaluation uses pre-trained and fine-tuned versions of two closed-source
(GPT-4/3.5) and two open-source LLMs (Mistral and Gemma). LLMs perform
significantly better for US English than Indian English for both TWP and TWS,
for all settings. While GPT-based models perform the best, the comparatively
smaller models work more equitably for short conversations (<8 turns). Our
results on AIGen and AITrans (the best and worst-performing subset)
respectively show that LLMs may learn a dialect of their own based on the
composition of the training data, and that dialect robustness is indeed a
challenging task. Our evaluation methodology exhibits a novel way to examine
attributes of language models using pre-existing dialogue datasets. | cs.CL | Evaluating Dialect Robustness of Language Models via Conversation Understanding |
2024-05-09T11:29:25Z | http://arxiv.org/pdf/2405.06709v1 | http://arxiv.org/abs/2405.06709v1 | Dimitris Asimopoulos, Ilias Siniosoglou, Vasileios Argyriou, Sotirios K. Goudos, Konstantinos E. Psannis, Nikoleta Karditsioti, Theocharis Saoulidis, Panagiotis Sarigiannidis | 2024-05-09T11:29:25Z | cs.CL, cs.AI | In the digital era, with escalating privacy concerns, it's imperative to
devise robust strategies that protect private data while maintaining the
intrinsic value of textual information. This research embarks on a
comprehensive examination of text anonymisation methods, focusing on
Conditional Random Fields (CRF), Long Short-Term Memory (LSTM), Embeddings from
Language Models (ELMo), and the transformative capabilities of the Transformers
architecture. Each model presents unique strengths since LSTM is modeling
long-term dependencies, CRF captures dependencies among word sequences, ELMo
delivers contextual word representations using deep bidirectional language
models and Transformers introduce self-attention mechanisms that provide
enhanced scalability. Our study is positioned as a comparative analysis of
these models, emphasising their synergistic potential in addressing text
anonymisation challenges. Preliminary results indicate that CRF, LSTM, and ELMo
individually outperform traditional methods. The inclusion of Transformers,
when compared alongside with the other models, offers a broader perspective on
achieving optimal text anonymisation in contemporary settings. | cs.CL | Evaluating the Efficacy of AI Techniques in Textual Anonymization: A Comparative Study |
2024-05-09T11:10:29Z | http://arxiv.org/pdf/2405.05678v1 | http://arxiv.org/abs/2405.05678v1 | Yoonsu Kim, Kihoon Son, Seoyoung Kim, Juho Kim | 2024-05-09T11:10:29Z | cs.HC, cs.CL | AI intent alignment, ensuring that AI produces outcomes as intended by users,
is a critical challenge in human-AI interaction. The emergence of generative
AI, including LLMs, has intensified the significance of this problem, as
interactions increasingly involve users specifying desired results for AI
systems. In order to support better AI intent alignment, we aim to explore
human strategies for intent specification in human-human communication. By
studying and comparing human-human and human-LLM communication, we identify key
strategies that can be applied to the design of AI systems that are more
effective at understanding and aligning with user intent. This study aims to
advance toward a human-centered AI system by bringing together human
communication strategies for the design of AI systems. | cs.HC | Beyond Prompts: Learning from Human Communication for Enhanced AI Intent Alignment |
2024-05-09T10:04:05Z | http://arxiv.org/pdf/2405.06708v5 | http://arxiv.org/abs/2405.06708v5 | Suyuan Zhao, Jiahuan Zhang, Yushuai Wu, Yizhen Luo, Zaiqing Nie | 2024-06-11T07:31:13Z | q-bio.GN, cs.AI, cs.CL | Cell identity encompasses various semantic aspects of a cell, including cell
type, pathway information, disease information, and more, which are essential
for biologists to gain insights into its biological characteristics.
Understanding cell identity from the transcriptomic data, such as annotating
cell types, has become an important task in bioinformatics. As these semantic
aspects are determined by human experts, it is impossible for AI models to
effectively carry out cell identity understanding tasks without the supervision
signals provided by single-cell and label pairs. The single-cell pre-trained
language models (PLMs) currently used for this task are trained only on a
single modality, transcriptomics data, lack an understanding of cell identity
knowledge. As a result, they have to be fine-tuned for downstream tasks and
struggle when lacking labeled data with the desired semantic labels. To address
this issue, we propose an innovative solution by constructing a unified
representation of single-cell data and natural language during the pre-training
phase, allowing the model to directly incorporate insights related to cell
identity. More specifically, we introduce $\textbf{LangCell}$, the first
$\textbf{Lang}$uage-$\textbf{Cell}$ pre-training framework. LangCell utilizes
texts enriched with cell identity information to gain a profound comprehension
of cross-modal knowledge. Results from experiments conducted on different
benchmarks show that LangCell is the only single-cell PLM that can work
effectively in zero-shot cell identity understanding scenarios, and also
significantly outperforms existing models in few-shot and fine-tuning cell
identity understanding scenarios. | q-bio.GN | LangCell: Language-Cell Pre-training for Cell Identity Understanding |
2024-05-09T08:46:17Z | http://arxiv.org/pdf/2405.06707v1 | http://arxiv.org/abs/2405.06707v1 | Yitian Li, Jidong Tian, Hao He, Yaohui Jin | 2024-05-09T08:46:17Z | cs.CL, cs.AI | Combining different forms of prompts with pre-trained large language models
has yielded remarkable results on reasoning tasks (e.g. Chain-of-Thought
prompting). However, along with testing on more complex reasoning, these
methods also expose problems such as invalid reasoning and fictional reasoning
paths. In this paper, we develop \textit{Hypothesis Testing Prompting}, which
adds conclusion assumptions, backward reasoning, and fact verification during
intermediate reasoning steps. \textit{Hypothesis Testing prompting} involves
multiple assumptions and reverses validation of conclusions leading to its
unique correct answer. Experiments on two challenging deductive reasoning
datasets ProofWriter and RuleTaker show that hypothesis testing prompting not
only significantly improves the effect, but also generates a more reasonable
and standardized reasoning process. | cs.CL | Hypothesis Testing Prompting Improves Deductive Reasoning in Large Language Models |
2024-05-09T08:32:55Z | http://arxiv.org/pdf/2405.05618v1 | http://arxiv.org/abs/2405.05618v1 | Ashlesha Akella, Abhijit Manatkar, Brij Chavda, Hima Patel | 2024-05-09T08:32:55Z | cs.LG, cs.CL | Efficient processing of tabular data is important in various industries,
especially when working with datasets containing a large number of columns.
Large language models (LLMs) have demonstrated their ability on several tasks
through carefully crafted prompts. However, creating effective prompts for
tabular datasets is challenging due to the structured nature of the data and
the need to manage numerous columns. This paper presents an innovative
auto-prompt generation system suitable for multiple LLMs, with minimal
training. It proposes two novel methods; 1) A Reinforcement Learning-based
algorithm for identifying and sequencing task-relevant columns 2) Cell-level
similarity-based approach for enhancing few-shot example selection. Our
approach has been extensively tested across 66 datasets, demonstrating improved
performance in three downstream tasks: data imputation, error detection, and
entity matching using two distinct LLMs; Google flan-t5-xxl and Mixtral 8x7B. | cs.LG | An Automatic Prompt Generation System for Tabular Data Tasks |
2024-05-09T08:28:12Z | http://arxiv.org/pdf/2405.05616v1 | http://arxiv.org/abs/2405.05616v1 | Ruiting Dai, Yuqiao Tan, Lisi Mo, Shuang Liang, Guohao Huo, Jiayi Luo, Yao Cheng | 2024-05-09T08:28:12Z | cs.CL, cs.AI | Commonsense question answering has demonstrated considerable potential across
various applications like assistants and social robots. Although fully
fine-tuned pre-trained Language Models(LM) have achieved remarkable performance
in commonsense reasoning, their tendency to excessively prioritize textual
information hampers the precise transfer of structural knowledge and undermines
interpretability. Some studies have explored combining LMs with Knowledge
Graphs(KGs) by coarsely fusing the two modalities to perform Graph Neural
Network(GNN)-based reasoning that lacks a profound interaction between
heterogeneous modalities. In this paper, we propose a novel Graph-based
Structure-Aware Prompt Learning Model for commonsense reasoning, named G-SAP,
aiming to maintain a balance between heterogeneous knowledge and enhance the
cross-modal interaction within the LM+GNNs model. In particular, an evidence
graph is constructed by integrating multiple knowledge sources, i.e.
ConceptNet, Wikipedia, and Cambridge Dictionary to boost the performance.
Afterward, a structure-aware frozen PLM is employed to fully incorporate the
structured and textual information from the evidence graph, where the
generation of prompts is driven by graph entities and relations. Finally, a
heterogeneous message-passing reasoning module is used to facilitate deep
interaction of knowledge between the LM and graph-based networks. Empirical
validation, conducted through extensive experiments on three benchmark
datasets, demonstrates the notable performance of the proposed model. The
results reveal a significant advancement over the existing models, especially,
with 6.12% improvement over the SoTA LM+GNNs model on the OpenbookQA dataset. | cs.CL | G-SAP: Graph-based Structure-Aware Prompt Learning over Heterogeneous Knowledge for Commonsense Reasoning |
2024-05-09T08:23:20Z | http://arxiv.org/pdf/2405.05615v1 | http://arxiv.org/abs/2405.05615v1 | Shibo Jie, Yehui Tang, Ning Ding, Zhi-Hong Deng, Kai Han, Yunhe Wang | 2024-05-09T08:23:20Z | cs.CV, cs.CL, cs.LG | Current solutions for efficiently constructing large vision-language (VL)
models follow a two-step paradigm: projecting the output of pre-trained vision
encoders to the input space of pre-trained language models as visual prompts;
and then transferring the models to downstream VL tasks via end-to-end
parameter-efficient fine-tuning (PEFT). However, this paradigm still exhibits
inefficiency since it significantly increases the input length of the language
models. In this paper, in contrast to integrating visual prompts into inputs,
we regard visual prompts as additional knowledge that facilitates language
models in addressing tasks associated with visual information. Motivated by the
finding that Feed-Forward Network (FFN) of language models acts as "key-value
memory", we introduce a novel approach termed memory-space visual prompting
(MemVP), wherein visual prompts are concatenated with the weights of FFN for
visual knowledge injection. Experimental results across various VL tasks and
language models reveal that MemVP significantly reduces the training time and
inference latency of the finetuned VL models and surpasses the performance of
previous PEFT methods. Code: https://github.com/JieShibo/MemVP | cs.CV | Memory-Space Visual Prompting for Efficient Vision-Language Fine-Tuning |
2024-05-09T08:15:21Z | http://arxiv.org/pdf/2405.05610v1 | http://arxiv.org/abs/2405.05610v1 | Xikang Yang, Xuehai Tang, Songlin Hu, Jizhong Han | 2024-05-09T08:15:21Z | cs.CL, cs.CR, cs.LG | Large language models (LLMs) have achieved remarkable performance in various
natural language processing tasks, especially in dialogue systems. However, LLM
may also pose security and moral threats, especially in multi round
conversations where large models are more easily guided by contextual content,
resulting in harmful or biased responses. In this paper, we present a novel
method to attack LLMs in multi-turn dialogues, called CoA (Chain of Attack).
CoA is a semantic-driven contextual multi-turn attack method that adaptively
adjusts the attack policy through contextual feedback and semantic relevance
during multi-turn of dialogue with a large model, resulting in the model
producing unreasonable or harmful content. We evaluate CoA on different LLMs
and datasets, and show that it can effectively expose the vulnerabilities of
LLMs, and outperform existing attack methods. Our work provides a new
perspective and tool for attacking and defending LLMs, and contributes to the
security and ethical assessment of dialogue systems. | cs.CL | Chain of Attack: a Semantic-Driven Contextual Multi-Turn attacker for LLM |
2024-05-09T07:47:25Z | http://arxiv.org/pdf/2405.06706v1 | http://arxiv.org/abs/2405.06706v1 | Shuo Zhang, Biao Yang, Zhang Li, Zhiyin Ma, Yuliang Liu, Xiang Bai | 2024-05-09T07:47:25Z | cs.CL, cs.AI | While large multi-modal models (LMM) have shown notable progress in
multi-modal tasks, their capabilities in tasks involving dense textual content
remains to be fully explored. Dense text, which carries important information,
is often found in documents, tables, and product descriptions. Understanding
dense text enables us to obtain more accurate information, assisting in making
better decisions. To further explore the capabilities of LMM in complex text
tasks, we propose the DT-VQA dataset, with 170k question-answer pairs. In this
paper, we conduct a comprehensive evaluation of GPT4V, Gemini, and various
open-source LMMs on our dataset, revealing their strengths and weaknesses.
Furthermore, we evaluate the effectiveness of two strategies for LMM: prompt
engineering and downstream fine-tuning. We find that even with automatically
labeled training datasets, significant improvements in model performance can be
achieved. We hope that this research will promote the study of LMM in dense
text tasks. Code will be released at
https://github.com/Yuliang-Liu/MultimodalOCR. | cs.CL | Exploring the Capabilities of Large Multimodal Models on Dense Text |
2024-05-09T07:39:19Z | http://arxiv.org/pdf/2405.05600v1 | http://arxiv.org/abs/2405.05600v1 | Zahra Abbasiantaeb, Chuan Meng, Leif Azzopardi, Mohammad Aliannejadi | 2024-05-09T07:39:19Z | cs.IR, cs.CL | Incomplete relevance judgments limit the re-usability of test collections.
When new systems are compared against previous systems used to build the pool
of judged documents, they often do so at a disadvantage due to the ``holes'' in
test collection (i.e., pockets of un-assessed documents returned by the new
system). In this paper, we take initial steps towards extending existing test
collections by employing Large Language Models (LLM) to fill the holes by
leveraging and grounding the method using existing human judgments. We explore
this problem in the context of Conversational Search using TREC iKAT, where
information needs are highly dynamic and the responses (and, the results
retrieved) are much more varied (leaving bigger holes). While previous work has
shown that automatic judgments from LLMs result in highly correlated rankings,
we find substantially lower correlates when human plus automatic judgments are
used (regardless of LLM, one/two/few shot, or fine-tuned). We further find
that, depending on the LLM employed, new runs will be highly favored (or
penalized), and this effect is magnified proportionally to the size of the
holes. Instead, one should generate the LLM annotations on the whole document
pool to achieve more consistent rankings with human-generated labels. Future
work is required to prompt engineering and fine-tuning LLMs to reflect and
represent the human annotations, in order to ground and align the models, such
that they are more fit for purpose. | cs.IR | Can We Use Large Language Models to Fill Relevance Judgment Holes? |
2024-05-09T07:37:34Z | http://arxiv.org/pdf/2405.06705v1 | http://arxiv.org/abs/2405.06705v1 | Zhuoxuan Jiang, Haoyuan Peng, Shanshan Feng, Fan Li, Dongsheng Li | 2024-05-09T07:37:34Z | cs.CL, cs.AI | Self-correction is emerging as a promising approach to mitigate the issue of
hallucination in Large Language Models (LLMs). To facilitate effective
self-correction, recent research has proposed mistake detection as its initial
step. However, current literature suggests that LLMs often struggle with
reliably identifying reasoning mistakes when using simplistic prompting
strategies. To address this challenge, we introduce a unique prompting
strategy, termed the Pedagogical Chain-of-Thought (PedCoT), which is
specifically designed to guide the identification of reasoning mistakes,
particularly mathematical reasoning mistakes. PedCoT consists of pedagogical
principles for prompts (PPP) design, two-stage interaction process (TIP) and
grounded PedCoT prompts, all inspired by the educational theory of the Bloom
Cognitive Model (BCM). We evaluate our approach on two public datasets
featuring math problems of varying difficulty levels. The experiments
demonstrate that our zero-shot prompting strategy significantly outperforms
strong baselines. The proposed method can achieve the goal of reliable
mathematical mistake identification and provide a foundation for automatic math
answer grading. The results underscore the significance of educational theory,
serving as domain knowledge, in guiding prompting strategy design for
addressing challenging tasks with LLMs effectively. | cs.CL | LLMs can Find Mathematical Reasoning Mistakes by Pedagogical Chain-of-Thought |
2024-05-09T07:15:19Z | http://arxiv.org/pdf/2405.05583v1 | http://arxiv.org/abs/2405.05583v1 | Yuxia Wang, Minghan Wang, Hasan Iqbal, Georgi Georgiev, Jiahui Geng, Preslav Nakov | 2024-05-09T07:15:19Z | cs.CL | The increased use of large language models (LLMs) across a variety of
real-world applications calls for mechanisms to verify the factual accuracy of
their outputs. Difficulties lie in assessing the factuality of free-form
responses in open domains. Also, different papers use disparate evaluation
benchmarks and measurements, which renders them hard to compare and hampers
future progress. To mitigate these issues, we propose OpenFactCheck, a unified
factuality evaluation framework for LLMs. OpenFactCheck consists of three
modules: (i) CUSTCHECKER allows users to easily customize an automatic
fact-checker and verify the factual correctness of documents and claims, (ii)
LLMEVAL, a unified evaluation framework assesses LLM's factuality ability from
various perspectives fairly, and (iii) CHECKEREVAL is an extensible solution
for gauging the reliability of automatic fact-checkers' verification results
using human-annotated datasets. OpenFactCheck is publicly released at
https://github.com/yuxiaw/OpenFactCheck. | cs.CL | OpenFactCheck: A Unified Framework for Factuality Evaluation of LLMs |
2024-05-09T07:12:45Z | http://arxiv.org/pdf/2405.05581v1 | http://arxiv.org/abs/2405.05581v1 | Yoonjoo Lee, Kihoon Son, Tae Soo Kim, Jisu Kim, John Joon Young Chung, Eytan Adar, Juho Kim | 2024-05-09T07:12:45Z | cs.HC, cs.AI, cs.CL | As Large Language Models (LLMs) are nondeterministic, the same input can
generate different outputs, some of which may be incorrect or hallucinated. If
run again, the LLM may correct itself and produce the correct answer.
Unfortunately, most LLM-powered systems resort to single results which, correct
or not, users accept. Having the LLM produce multiple outputs may help identify
disagreements or alternatives. However, it is not obvious how the user will
interpret conflicts or inconsistencies. To this end, we investigate how users
perceive the AI model and comprehend the generated information when they
receive multiple, potentially inconsistent, outputs. Through a preliminary
study, we identified five types of output inconsistencies. Based on these
categories, we conducted a study (N=252) in which participants were given one
or more LLM-generated passages to an information-seeking question. We found
that inconsistency within multiple LLM-generated outputs lowered the
participants' perceived AI capacity, while also increasing their comprehension
of the given information. Specifically, we observed that this positive effect
of inconsistencies was most significant for participants who read two passages,
compared to those who read three. Based on these findings, we present design
implications that, instead of regarding LLM output inconsistencies as a
drawback, we can reveal the potential inconsistencies to transparently indicate
the limitations of these models and promote critical LLM usage. | cs.HC | One vs. Many: Comprehending Accurate Information from Multiple Erroneous and Inconsistent AI Generations |
2024-05-09T06:40:39Z | http://arxiv.org/pdf/2405.05572v1 | http://arxiv.org/abs/2405.05572v1 | Prashant Kodali, Anmol Goel, Likhith Asapu, Vamshi Krishna Bonagiri, Anirudh Govil, Monojit Choudhury, Manish Shrivastava, Ponnurangam Kumaraguru | 2024-05-09T06:40:39Z | cs.CL, cs.AI | Current computational approaches for analysing or generating code-mixed
sentences do not explicitly model "naturalness" or "acceptability" of
code-mixed sentences, but rely on training corpora to reflect distribution of
acceptable code-mixed sentences. Modelling human judgement for the
acceptability of code-mixed text can help in distinguishing natural code-mixed
text and enable quality-controlled generation of code-mixed text. To this end,
we construct Cline - a dataset containing human acceptability judgements for
English-Hindi (en-hi) code-mixed text. Cline is the largest of its kind with
16,642 sentences, consisting of samples sourced from two sources: synthetically
generated code-mixed text and samples collected from online social media. Our
analysis establishes that popular code-mixing metrics such as CMI, Number of
Switch Points, Burstines, which are used to filter/curate/compare code-mixed
corpora have low correlation with human acceptability judgements, underlining
the necessity of our dataset. Experiments using Cline demonstrate that simple
Multilayer Perceptron (MLP) models trained solely on code-mixing metrics are
outperformed by fine-tuned pre-trained Multilingual Large Language Models
(MLLMs). Specifically, XLM-Roberta and Bernice outperform IndicBERT across
different configurations in challenging data settings. Comparison with
ChatGPT's zero and fewshot capabilities shows that MLLMs fine-tuned on larger
data outperform ChatGPT, providing scope for improvement in code-mixed tasks.
Zero-shot transfer from English-Hindi to English-Telugu acceptability judgments
using our model checkpoints proves superior to random baselines, enabling
application to other code-mixed language pairs and providing further avenues of
research. We publicly release our human-annotated dataset, trained checkpoints,
code-mix corpus, and code for data generation and model training. | cs.CL | From Human Judgements to Predictive Models: Unravelling Acceptability in Code-Mixed Sentences |
2024-05-09T02:44:42Z | http://arxiv.org/pdf/2405.05513v1 | http://arxiv.org/abs/2405.05513v1 | Yicheng Yang, Xinyu Wang, Haoming Yu, Zhiyuan Li | 2024-05-09T02:44:42Z | cs.CL, cs.DM | The increase in academic dishonesty cases among college students has raised
concern, particularly due to the shift towards online learning caused by the
pandemic. We aim to develop and implement a method capable of generating
tailored questions for each student. The use of Automatic Question Generation
(AQG) is a possible solution. Previous studies have investigated AQG frameworks
in education, which include validity, user-defined difficulty, and personalized
problem generation. Our new AQG approach produces logical equivalence problems
for Discrete Mathematics, which is a core course for year-one computer science
students. This approach utilizes a syntactic grammar and a semantic attribute
system through top-down parsing and syntax tree transformations. Our
experiments show that the difficulty level of questions generated by our AQG
approach is similar to the questions presented to students in the textbook [1].
These results confirm the practicality of our AQG approach for automated
question generation in education, with the potential to significantly enhance
learning experiences. | cs.CL | Automatic question generation for propositional logical equivalences |
2024-05-09T02:35:32Z | http://arxiv.org/pdf/2405.05990v2 | http://arxiv.org/abs/2405.05990v2 | Yang Bai, Ge Pei, Jindong Gu, Yong Yang, Xingjun Ma | 2024-05-20T14:40:03Z | cs.CR, cs.AI, cs.CL, cs.LG | Large language models (LLMs) have achieved remarkable performance on a wide
range of tasks. However, recent studies have shown that LLMs can memorize
training data and simple repeated tokens can trick the model to leak the data.
In this paper, we take a step further and show that certain special characters
or their combinations with English letters are stronger memory triggers,
leading to more severe data leakage. The intuition is that, since LLMs are
trained with massive data that contains a substantial amount of special
characters (e.g. structural symbols {, } of JSON files, and @, # in emails and
online posts), the model may memorize the co-occurrence between these special
characters and the raw texts. This motivates us to propose a simple but
effective Special Characters Attack (SCA) to induce training data leakage. Our
experiments verify the high effectiveness of SCA against state-of-the-art LLMs:
they can leak diverse training data, such as code corpus, web pages, and
personally identifiable information, and sometimes generate non-stop outputs as
a byproduct. We further show that the composition of the training data corpus
can be revealed by inspecting the leaked data -- one crucial piece of
information for pre-training high-performance LLMs. Our work can help
understand the sensitivity of LLMs to special characters and identify potential
areas for improvement. | cs.CR | Special Characters Attack: Toward Scalable Training Data Extraction From Large Language Models |
Subsets and Splits
No saved queries yet
Save your SQL queries to embed, download, and access them later. Queries will appear here once saved.