publication_date
stringlengths 20
20
| doc_url
stringlengths 33
34
| id
stringlengths 33
34
| authors
stringlengths 5
20.1k
| update_date
stringlengths 20
20
| category_all
stringlengths 5
138
| abstract
stringlengths 123
3.09k
| category_primary
stringclasses 78
values | title
stringlengths 10
211
|
|---|---|---|---|---|---|---|---|---|
2024-02-06T01:13:53Z | http://arxiv.org/pdf/2402.03620v1 | http://arxiv.org/abs/2402.03620v1 | Pei Zhou, Jay Pujara, Xiang Ren, Xinyun Chen, Heng-Tze Cheng, Quoc V. Le, Ed H. Chi, Denny Zhou, Swaroop Mishra, Huaixiu Steven Zheng | 2024-02-06T01:13:53Z | cs.AI, cs.CL | We introduce SELF-DISCOVER, a general framework for LLMs to self-discover the
task-intrinsic reasoning structures to tackle complex reasoning problems that
are challenging for typical prompting methods. Core to the framework is a
self-discovery process where LLMs select multiple atomic reasoning modules such
as critical thinking and step-by-step thinking, and compose them into an
explicit reasoning structure for LLMs to follow during decoding. SELF-DISCOVER
substantially improves GPT-4 and PaLM 2's performance on challenging reasoning
benchmarks such as BigBench-Hard, grounded agent reasoning, and MATH, by as
much as 32% compared to Chain of Thought (CoT). Furthermore, SELF-DISCOVER
outperforms inference-intensive methods such as CoT-Self-Consistency by more
than 20%, while requiring 10-40x fewer inference compute. Finally, we show that
the self-discovered reasoning structures are universally applicable across
model families: from PaLM 2-L to GPT-4, and from GPT-4 to Llama2, and share
commonalities with human reasoning patterns. | cs.AI | Self-Discover: Large Language Models Self-Compose Reasoning Structures |
2024-02-06T01:07:56Z | http://arxiv.org/pdf/2402.03618v1 | http://arxiv.org/abs/2402.03618v1 | Sreejan Kumar, Raja Marjieh, Byron Zhang, Declan Campbell, Michael Y. Hu, Umang Bhatt, Brenden Lake, Thomas L. Griffiths | 2024-02-06T01:07:56Z | cs.AI, cs.CL, q-bio.NC | Humans extract useful abstractions of the world from noisy sensory data.
Serial reproduction allows us to study how people construe the world through a
paradigm similar to the game of telephone, where one person observes a stimulus
and reproduces it for the next to form a chain of reproductions. Past serial
reproduction experiments typically employ a single sensory modality, but humans
often communicate abstractions of the world to each other through language. To
investigate the effect language on the formation of abstractions, we implement
a novel multimodal serial reproduction framework by asking people who receive a
visual stimulus to reproduce it in a linguistic format, and vice versa. We ran
unimodal and multimodal chains with both humans and GPT-4 and find that adding
language as a modality has a larger effect on human reproductions than GPT-4's.
This suggests human visual and linguistic representations are more dissociable
than those of GPT-4. | cs.AI | Comparing Abstraction in Humans and Large Language Models Using Multimodal Serial Reproduction |
2024-02-06T01:05:14Z | http://arxiv.org/pdf/2402.03616v1 | http://arxiv.org/abs/2402.03616v1 | Yujin Kim, Chin-Chia Hsu | 2024-02-06T01:05:14Z | cs.CL, cs.AI, cs.HC, cs.IR | Large Language Models (LLMs) hold the potential to perform a variety of text
processing tasks and provide textual explanations for proposed actions or
decisions. In the era of hybrid work, LLMs can provide intelligent decision
support for workers who are designing their hybrid work plans. In particular,
they can offer suggestions and explanations to workers balancing numerous
decision factors, thereby enhancing their work experience. In this paper, we
present a decision support model for workspaces in hybrid work environments,
leveraging the reasoning skill of LLMs. We first examine LLM's capability of
making suitable workspace suggestions. We find that its reasoning extends
beyond the guidelines in the prompt and the LLM can manage the trade-off among
the available resources in the workspaces. We conduct an extensive user study
to understand workers' decision process for workspace choices and evaluate the
effectiveness of the system. We observe that a worker's decision could be
influenced by the LLM's suggestions and explanations. The participants in our
study find the system to be convenient, regardless of whether reasons are
provided or not. Our results show that employees can benefit from the
LLM-empowered system for their workspace selection in hybrid workplace. | cs.CL | Leveraging Large Language Models for Hybrid Workplace Decision Support |
2024-02-06T00:53:27Z | http://arxiv.org/pdf/2402.03610v1 | http://arxiv.org/abs/2402.03610v1 | Tomoyuki Kagaya, Thong Jing Yuan, Yuxuan Lou, Jayashree Karlekar, Sugiri Pranata, Akira Kinose, Koki Oguri, Felix Wick, Yang You | 2024-02-06T00:53:27Z | cs.LG, cs.AI, cs.CL | Owing to recent advancements, Large Language Models (LLMs) can now be
deployed as agents for increasingly complex decision-making applications in
areas including robotics, gaming, and API integration. However, reflecting past
experiences in current decision-making processes, an innate human behavior,
continues to pose significant challenges. Addressing this, we propose
Retrieval-Augmented Planning (RAP) framework, designed to dynamically leverage
past experiences corresponding to the current situation and context, thereby
enhancing agents' planning capabilities. RAP distinguishes itself by being
versatile: it excels in both text-only and multimodal environments, making it
suitable for a wide range of tasks. Empirical evaluations demonstrate RAP's
effectiveness, where it achieves SOTA performance in textual scenarios and
notably enhances multimodal LLM agents' performance for embodied tasks. These
results highlight RAP's potential in advancing the functionality and
applicability of LLM agents in complex, real-world applications. | cs.LG | RAP: Retrieval-Augmented Planning with Contextual Memory for Multimodal LLM Agents |
2024-02-06T00:51:27Z | http://arxiv.org/pdf/2402.03607v1 | http://arxiv.org/abs/2402.03607v1 | Trilok Padhi, Ugur Kursuncu, Yaman Kumar, Valerie L. Shalin, Lane Peterson Fronczek | 2024-02-06T00:51:27Z | cs.AI, cs.CL, cs.CV, cs.CY, cs.HC, I.2.7; I.2.10; I.2.4; I.2.1 | The prevalence of smart devices with the ability to capture moments in
multiple modalities has enabled users to experience multimodal information
online. However, large Language (LLMs) and Vision models (LVMs) are still
limited in capturing holistic meaning with cross-modal semantic relationships.
Without explicit, common sense knowledge (e.g., as a knowledge graph), Visual
Language Models (VLMs) only learn implicit representations by capturing
high-level patterns in vast corpora, missing essential contextual cross-modal
cues. In this work, we design a framework to couple explicit commonsense
knowledge in the form of knowledge graphs with large VLMs to improve the
performance of a downstream task, predicting the effectiveness of multi-modal
marketing campaigns. While the marketing application provides a compelling
metric for assessing our methods, our approach enables the early detection of
likely persuasive multi-modal campaigns and the assessment and augmentation of
marketing theory. | cs.AI | Improving Contextual Congruence Across Modalities for Effective Multimodal Marketing using Knowledge-infused Learning |
2024-02-06T00:32:28Z | http://arxiv.org/pdf/2402.05969v2 | http://arxiv.org/abs/2402.05969v2 | Chunsheng Zuo, Michael Guerzhoy | 2024-06-16T22:18:36Z | cs.LG, cs.CL | As we show in this paper, the prediction for output token $n+1$ of
Transformer architectures without one of the mechanisms of positional encodings
and causal attention is invariant to permutations of input tokens $1, 2, ...,
n-1$. Usually, both mechanisms are employed and the symmetry with respect to
the input tokens is broken. Recently, it has been shown that one can train
Transformers without positional encodings. This must be enabled by the causal
attention mechanism. In this paper, we elaborate on the argument that the
causal connection mechanism must be responsible for the fact that Transformers
are able to model input sequences where the order is important. Vertical
"slices" of Transformers are all encouraged to represent the same location $k$
in the input sequence. We hypothesize that residual connections contribute to
this phenomenon, and demonstrate evidence for this. | cs.LG | Breaking Symmetry When Training Transformers |
2024-02-06T00:14:53Z | http://arxiv.org/pdf/2402.03597v1 | http://arxiv.org/abs/2402.03597v1 | Brenda Y. Miao, Christopher YK Williams, Ebenezer Chinedu-Eneh, Travis Zack, Emily Alsentzer, Atul J. Butte, Irene Y. Chen | 2024-02-06T00:14:53Z | cs.CL, cs.IR, cs.LG | Prescription contraceptives play a critical role in supporting women's
reproductive health. With nearly 50 million women in the United States using
contraceptives, understanding the factors that drive contraceptives selection
and switching is of significant interest. However, many factors related to
medication switching are often only captured in unstructured clinical notes and
can be difficult to extract. Here, we evaluate the zero-shot abilities of a
recently developed large language model, GPT-4 (via HIPAA-compliant Microsoft
Azure API), to identify reasons for switching between classes of contraceptives
from the UCSF Information Commons clinical notes dataset. We demonstrate that
GPT-4 can accurately extract reasons for contraceptive switching, outperforming
baseline BERT-based models with microF1 scores of 0.849 and 0.881 for
contraceptive start and stop extraction, respectively. Human evaluation of
GPT-4-extracted reasons for switching showed 91.4% accuracy, with minimal
hallucinations. Using extracted reasons, we identified patient preference,
adverse events, and insurance as key reasons for switching using unsupervised
topic modeling approaches. Notably, we also showed using our approach that
"weight gain/mood change" and "insurance coverage" are disproportionately found
as reasons for contraceptive switching in specific demographic populations. Our
code and supplemental data are available at
https://github.com/BMiao10/contraceptive-switching. | cs.CL | Identifying Reasons for Contraceptive Switching from Real-World Data Using Large Language Models |
2024-02-05T22:48:28Z | http://arxiv.org/pdf/2402.05132v2 | http://arxiv.org/abs/2402.05132v2 | Kaan Kale, Homa Esfahanizadeh, Noel Elias, Oguzhan Baser, Muriel Medard, Sriram Vishwanath | 2024-05-11T20:03:26Z | cs.CL, cs.IT, math.IT | With the exponential growth in data volume and the emergence of
data-intensive applications, particularly in the field of machine learning,
concerns related to resource utilization, privacy, and fairness have become
paramount. This paper focuses on the textual domain of data and addresses
challenges regarding encoding sentences to their optimized representations
through the lens of information-theory. In particular, we use empirical
estimates of mutual information, using the Donsker-Varadhan definition of
Kullback-Leibler divergence. Our approach leverages this estimation to train an
information-theoretic sentence embedding, called TexShape, for (task-based)
data compression or for filtering out sensitive information, enhancing privacy
and fairness. In this study, we employ a benchmark language model for initial
text representation, complemented by neural networks for information-theoretic
compression and mutual information estimations. Our experiments demonstrate
significant advancements in preserving maximal targeted information and minimal
sensitive information over adverse compression ratios, in terms of predictive
accuracy of downstream models that are trained using the compressed data. | cs.CL | TexShape: Information Theoretic Sentence Embedding for Language Models |
2024-02-05T22:35:42Z | http://arxiv.org/pdf/2402.05131v3 | http://arxiv.org/abs/2402.05131v3 | Antonio Jimeno Yepes, Yao You, Jan Milczek, Sebastian Laverde, Renyu Li | 2024-03-16T09:08:26Z | cs.CL | Chunking information is a key step in Retrieval Augmented Generation (RAG).
Current research primarily centers on paragraph-level chunking. This approach
treats all texts as equal and neglects the information contained in the
structure of documents. We propose an expanded approach to chunk documents by
moving beyond mere paragraph-level chunking to chunk primary by structural
element components of documents. Dissecting documents into these constituent
elements creates a new way to chunk documents that yields the best chunk size
without tuning. We introduce a novel framework that evaluates how chunking
based on element types annotated by document understanding models contributes
to the overall context and accuracy of the information retrieved. We also
demonstrate how this approach impacts RAG assisted Question & Answer task
performance. Our research includes a comprehensive analysis of various element
types, their role in effective information retrieval, and the impact they have
on the quality of RAG outputs. Findings support that element type based
chunking largely improve RAG results on financial reporting. Through this
research, we are also able to answer how to uncover highly accurate RAG. | cs.CL | Financial Report Chunking for Effective Retrieval Augmented Generation |
2024-02-05T22:22:49Z | http://arxiv.org/pdf/2402.03563v2 | http://arxiv.org/abs/2402.03563v2 | Gustaf Ahdritz, Tian Qin, Nikhil Vyas, Boaz Barak, Benjamin L. Edelman | 2024-02-27T07:37:08Z | cs.LG, cs.AI, cs.CL | We study the feasibility of identifying epistemic uncertainty (reflecting a
lack of knowledge), as opposed to aleatoric uncertainty (reflecting entropy in
the underlying distribution), in the outputs of large language models (LLMs)
over free-form text. In the absence of ground-truth probabilities, we explore a
setting where, in order to (approximately) disentangle a given LLM's
uncertainty, a significantly larger model stands in as a proxy for the ground
truth. We show that small linear probes trained on the embeddings of frozen,
pretrained models accurately predict when larger models will be more confident
at the token level and that probes trained on one text domain generalize to
others. Going further, we propose a fully unsupervised method that achieves
non-trivial accuracy on the same task. Taken together, we interpret these
results as evidence that LLMs naturally contain internal representations of
different types of uncertainty that could potentially be leveraged to devise
more informative indicators of model confidence in diverse practical settings. | cs.LG | Distinguishing the Knowable from the Unknowable with Language Models |
2024-02-05T22:20:19Z | http://arxiv.org/pdf/2402.03561v2 | http://arxiv.org/abs/2402.03561v2 | Jialu Li, Aishwarya Padmakumar, Gaurav Sukhatme, Mohit Bansal | 2024-02-07T18:02:51Z | cs.CV, cs.AI, cs.CL | Outdoor Vision-and-Language Navigation (VLN) requires an agent to navigate
through realistic 3D outdoor environments based on natural language
instructions. The performance of existing VLN methods is limited by
insufficient diversity in navigation environments and limited training data. To
address these issues, we propose VLN-Video, which utilizes the diverse outdoor
environments present in driving videos in multiple cities in the U.S. augmented
with automatically generated navigation instructions and actions to improve
outdoor VLN performance. VLN-Video combines the best of intuitive classical
approaches and modern deep learning techniques, using template infilling to
generate grounded navigation instructions, combined with an image rotation
similarity-based navigation action predictor to obtain VLN style data from
driving videos for pretraining deep learning VLN models. We pre-train the model
on the Touchdown dataset and our video-augmented dataset created from driving
videos with three proxy tasks: Masked Language Modeling, Instruction and
Trajectory Matching, and Next Action Prediction, so as to learn
temporally-aware and visually-aligned instruction representations. The learned
instruction representation is adapted to the state-of-the-art navigator when
fine-tuning on the Touchdown dataset. Empirical results demonstrate that
VLN-Video significantly outperforms previous state-of-the-art models by 2.1% in
task completion rate, achieving a new state-of-the-art on the Touchdown
dataset. | cs.CV | VLN-Video: Utilizing Driving Videos for Outdoor Vision-and-Language Navigation |
2024-02-05T21:05:35Z | http://arxiv.org/pdf/2402.03519v1 | http://arxiv.org/abs/2402.03519v1 | Xiliang Zhu, Chia-Tien Chang, Shayna Gardiner, David Rossouw, Jonas Robertson | 2024-02-05T21:05:35Z | cs.CL, cs.AI | Punctuation restoration is a crucial step after Automatic Speech Recognition
(ASR) systems to enhance transcript readability and facilitate subsequent NLP
tasks. Nevertheless, conventional lexical-based approaches are inadequate for
solving the punctuation restoration task in Spanish, where ambiguity can be
often found between unpunctuated declaratives and questions. In this study, we
propose a novel hybrid acoustic-lexical punctuation restoration system for
Spanish transcription, which consolidates acoustic and lexical signals through
a modular process. Our experiment results show that the proposed system can
effectively improve F1 score of question marks and overall punctuation
restoration on both public and internal Spanish conversational datasets.
Additionally, benchmark comparison against LLMs (Large Language Model)
indicates the superiority of our approach in accuracy, reliability and latency.
Furthermore, we demonstrate that the Word Error Rate (WER) of the ASR module
also benefits from our proposed system. | cs.CL | Resolving Transcription Ambiguity in Spanish: A Hybrid Acoustic-Lexical System for Punctuation Restoration |
2024-02-05T20:51:11Z | http://arxiv.org/pdf/2402.03509v1 | http://arxiv.org/abs/2402.03509v1 | Sanjana Ramprasad, Kundan Krishna, Zachary C Lipton, Byron C Wallace | 2024-02-05T20:51:11Z | cs.CL, cs.AI, cs.LG | Recent work has shown that large language models (LLMs) are capable of
generating summaries zero-shot (i.e., without explicit supervision) that, under
human assessment, are often comparable or even preferred to manually composed
reference summaries. However, this prior work has focussed almost exclusively
on evaluating news article summarization. How do zero-shot summarizers perform
in other (potentially more specialized) domains? In this work we evaluate
zero-shot generated summaries across specialized domains including biomedical
articles, and legal bills (in addition to standard news benchmarks for
reference). We focus especially on the factuality of outputs. We acquire
annotations from domain experts to identify inconsistencies in summaries and
systematically categorize these errors. We analyze whether the prevalence of a
given domain in the pretraining corpus affects extractiveness and faithfulness
of generated summaries of articles in this domain. We release all collected
annotations to facilitate additional research toward measuring and realizing
factually accurate summarization, beyond news articles. The dataset can be
downloaded from https://github.com/sanjanaramprasad/zero_shot_faceval_domains | cs.CL | Evaluating the Factuality of Zero-shot Summarizers Across Varied Domains |
2024-02-05T20:48:57Z | http://arxiv.org/pdf/2402.03507v1 | http://arxiv.org/abs/2402.03507v1 | Mikel Bober-Irizar, Soumya Banerjee | 2024-02-05T20:48:57Z | cs.AI, cs.CL, cs.LG | For half a century, artificial intelligence research has attempted to
reproduce the human qualities of abstraction and reasoning - creating computer
systems that can learn new concepts from a minimal set of examples, in settings
where humans find this easy. While specific neural networks are able to solve
an impressive range of problems, broad generalisation to situations outside
their training data has proved elusive.In this work, we look at several novel
approaches for solving the Abstraction & Reasoning Corpus (ARC), a dataset of
abstract visual reasoning tasks introduced to test algorithms on broad
generalization. Despite three international competitions with $100,000 in
prizes, the best algorithms still fail to solve a majority of ARC tasks and
rely on complex hand-crafted rules, without using machine learning at all. We
revisit whether recent advances in neural networks allow progress on this task.
First, we adapt the DreamCoder neurosymbolic reasoning solver to ARC.
DreamCoder automatically writes programs in a bespoke domain-specific language
to perform reasoning, using a neural network to mimic human intuition. We
present the Perceptual Abstraction and Reasoning Language (PeARL) language,
which allows DreamCoder to solve ARC tasks, and propose a new recognition model
that allows us to significantly improve on the previous best implementation.We
also propose a new encoding and augmentation scheme that allows large language
models (LLMs) to solve ARC tasks, and find that the largest models can solve
some ARC tasks. LLMs are able to solve a different group of problems to
state-of-the-art solvers, and provide an interesting way to complement other
approaches. We perform an ensemble analysis, combining models to achieve better
results than any system alone. Finally, we publish the arckit Python library to
make future research on ARC easier. | cs.AI | Neural networks for abstraction and reasoning: Towards broad generalization in machines |
2024-02-05T20:34:32Z | http://arxiv.org/pdf/2402.03501v1 | http://arxiv.org/abs/2402.03501v1 | Felipe Rodrigues Perche-Mahlow, André Felipe-Zanella, William Alberto Cruz-Castañeda, Marcellus Amadeus | 2024-02-05T20:34:32Z | cs.CV, cs.AI, cs.CL | In recent years, groundbreaking advancements in Generative Artificial
Intelligence (GenAI) have triggered a transformative paradigm shift,
significantly influencing various domains. In this work, we specifically
explore an integrated approach, leveraging advanced techniques in GenAI and
computer vision emphasizing image manipulation. The methodology unfolds through
several stages, including depth estimation, the creation of inpaint masks based
on depth information, the generation and replacement of backgrounds utilizing
Stable Diffusion in conjunction with Latent Consistency Models (LCMs), and the
subsequent replacement of clothes and application of aesthetic changes through
an inpainting pipeline. Experiments conducted in this study underscore the
methodology's efficacy, highlighting its potential to produce visually
captivating content. The convergence of these advanced techniques allows users
to input photographs of individuals and manipulate them to modify clothing and
background based on specific prompts without manually input inpainting masks,
effectively placing the subjects within the vast landscape of creative
imagination. | cs.CV | An Inpainting-Infused Pipeline for Attire and Background Replacement |
2024-02-05T19:56:56Z | http://arxiv.org/pdf/2402.03485v2 | http://arxiv.org/abs/2402.03485v2 | Gianluigi Lopardo, Frederic Precioso, Damien Garreau | 2024-06-17T13:18:30Z | stat.ML, cs.CL, cs.LG | Attention-based architectures, in particular transformers, are at the heart
of a technological revolution. Interestingly, in addition to helping obtain
state-of-the-art results on a wide range of applications, the attention
mechanism intrinsically provides meaningful insights on the internal behavior
of the model. Can these insights be used as explanations? Debate rages on. In
this paper, we mathematically study a simple attention-based architecture and
pinpoint the differences between post-hoc and attention-based explanations. We
show that they provide quite different results, and that, despite their
limitations, post-hoc methods are capable of capturing more useful insights
than merely examining the attention weights. | stat.ML | Attention Meets Post-hoc Interpretability: A Mathematical Perspective |
2024-02-05T19:56:27Z | http://arxiv.org/pdf/2402.03484v1 | http://arxiv.org/abs/2402.03484v1 | Ashley Shin, Qiao Jin, James Anibal, Zhiyong Lu | 2024-02-05T19:56:27Z | cs.IR, cs.CL | Searching for a related article based on a reference article is an integral
part of scientific research. PubMed, like many academic search engines, has a
"similar articles" feature that recommends articles relevant to the current
article viewed by a user. Explaining recommended items can be of great utility
to users, particularly in the literature search process. With more than a
million biomedical papers being published each year, explaining the recommended
similar articles would facilitate researchers and clinicians in searching for
related articles. Nonetheless, the majority of current literature
recommendation systems lack explanations for their suggestions. We employ a
post hoc approach to explaining recommendations by identifying relevant tokens
in the titles of similar articles. Our major contribution is building PubCLogs
by repurposing 5.6 million pairs of coclicked articles from PubMed's user query
logs. Using our PubCLogs dataset, we train the Highlight Similar Article Title
(HSAT), a transformer-based model designed to select the most relevant parts of
the title of a similar article, based on the title and abstract of a seed
article. HSAT demonstrates strong performance in our empirical evaluations,
achieving an F1 score of 91.72 percent on the PubCLogs test set, considerably
outperforming several baselines including BM25 (70.62), MPNet (67.11), MedCPT
(62.22), GPT-3.5 (46.00), and GPT-4 (64.89). Additional evaluations on a
separate, manually annotated test set further verifies HSAT's performance.
Moreover, participants of our user study indicate a preference for HSAT, due to
its superior balance between conciseness and comprehensiveness. Our study
suggests that repurposing user query logs of academic search engines can be a
promising way to train state-of-the-art models for explaining literature
recommendation. | cs.IR | Harnessing PubMed User Query Logs for Post Hoc Explanations of Recommended Similar Articles |
2024-02-05T19:55:06Z | http://arxiv.org/pdf/2402.03483v1 | http://arxiv.org/abs/2402.03483v1 | Zeeshan Patel, Karim El-Refai, Jonathan Pei, Tianle Li | 2024-02-05T19:55:06Z | cs.CL, cs.AI | Automated long-form story generation typically employs long-context large
language models (LLMs) for one-shot creation, which can produce cohesive but
not necessarily engaging content. We introduce Storytelling With Action
Guidance (SWAG), a novel approach to storytelling with LLMs. Our approach
reduces story writing to a search problem through a two-model feedback loop:
one LLM generates story content, and another auxiliary LLM is used to choose
the next best "action" to steer the story's future direction. Our results show
that SWAG can substantially outperform previous end-to-end story generation
techniques when evaluated by GPT-4 and through human evaluation, and our SWAG
pipeline using only open-source models surpasses GPT-3.5-Turbo. | cs.CL | SWAG: Storytelling With Action Guidance |
2024-02-05T19:49:13Z | http://arxiv.org/pdf/2402.07927v1 | http://arxiv.org/abs/2402.07927v1 | Pranab Sahoo, Ayush Kumar Singh, Sriparna Saha, Vinija Jain, Samrat Mondal, Aman Chadha | 2024-02-05T19:49:13Z | cs.AI, cs.CL, cs.HC | Prompt engineering has emerged as an indispensable technique for extending
the capabilities of large language models (LLMs) and vision-language models
(VLMs). This approach leverages task-specific instructions, known as prompts,
to enhance model efficacy without modifying the core model parameters. Rather
than updating the model parameters, prompts allow seamless integration of
pre-trained models into downstream tasks by eliciting desired model behaviors
solely based on the given prompt. Prompts can be natural language instructions
that provide context to guide the model or learned vector representations that
activate relevant knowledge. This burgeoning field has enabled success across
various applications, from question-answering to commonsense reasoning.
However, there remains a lack of systematic organization and understanding of
the diverse prompt engineering methods and techniques. This survey paper
addresses the gap by providing a structured overview of recent advancements in
prompt engineering, categorized by application area. For each prompting
approach, we provide a summary detailing the prompting methodology, its
applications, the models involved, and the datasets utilized. We also delve
into the strengths and limitations of each approach and include a taxonomy
diagram and table summarizing datasets, models, and critical points of each
prompting technique. This systematic analysis enables a better understanding of
this rapidly developing field and facilitates future research by illuminating
open challenges and opportunities for prompt engineering. | cs.AI | A Systematic Survey of Prompt Engineering in Large Language Models: Techniques and Applications |
2024-02-05T19:39:07Z | http://arxiv.org/pdf/2402.03477v1 | http://arxiv.org/abs/2402.03477v1 | Norah Alshahrani, Saied Alshahrani, Esma Wali, Jeanna Matthews | 2024-02-05T19:39:07Z | cs.CL | Text classification systems have been proven vulnerable to adversarial text
examples, modified versions of the original text examples that are often
unnoticed by human eyes, yet can force text classification models to alter
their classification. Often, research works quantifying the impact of
adversarial text attacks have been applied only to models trained in English.
In this paper, we introduce the first word-level study of adversarial attacks
in Arabic. Specifically, we use a synonym (word-level) attack using a Masked
Language Modeling (MLM) task with a BERT model in a black-box setting to assess
the robustness of the state-of-the-art text classification models to
adversarial attacks in Arabic. To evaluate the grammatical and semantic
similarities of the newly produced adversarial examples using our synonym
BERT-based attack, we invite four human evaluators to assess and compare the
produced adversarial examples with their original examples. We also study the
transferability of these newly produced Arabic adversarial examples to various
models and investigate the effectiveness of defense mechanisms against these
adversarial examples on the BERT models. We find that fine-tuned BERT models
were more susceptible to our synonym attacks than the other Deep Neural
Networks (DNN) models like WordCNN and WordLSTM we trained. We also find that
fine-tuned BERT models were more susceptible to transferred attacks. We,
lastly, find that fine-tuned BERT models successfully regain at least 2% in
accuracy after applying adversarial training as an initial defense mechanism. | cs.CL | Arabic Synonym BERT-based Adversarial Examples for Text Classification |
2024-02-05T19:00:02Z | http://arxiv.org/pdf/2402.03435v1 | http://arxiv.org/abs/2402.03435v1 | Sergi Blanco-Cuaresma | 2024-02-05T19:00:02Z | cs.CL, cs.AI, cs.CY | This study explores the use of Large Language Models (LLMs) to analyze text
comments from Reddit users, aiming to achieve two primary objectives: firstly,
to pinpoint critical excerpts that support a predefined psychological
assessment of suicidal risk; and secondly, to summarize the material to
substantiate the preassigned suicidal risk level. The work is circumscribed to
the use of "open-source" LLMs that can be run locally, thereby enhancing data
privacy. Furthermore, it prioritizes models with low computational
requirements, making it accessible to both individuals and institutions
operating on limited computing budgets. The implemented strategy only relies on
a carefully crafted prompt and a grammar to guide the LLM's text completion.
Despite its simplicity, the evaluation metrics show outstanding results, making
it a valuable privacy-focused and cost-effective approach. This work is part of
the Computational Linguistics and Clinical Psychology (CLPsych) 2024 shared
task. | cs.CL | Psychological Assessments with Large Language Models: A Privacy-Focused and Cost-Effective Approach |
2024-02-05T18:58:19Z | http://arxiv.org/pdf/2402.03303v1 | http://arxiv.org/abs/2402.03303v1 | Edward Kim | 2024-02-05T18:58:19Z | cs.CL, cs.AI, cs.LG | Given the impressive capabilities of recent Large Language Models (LLMs), we
investigate and benchmark the most popular proprietary and different sized open
source models on the task of explicit instruction following in conflicting
situations, e.g. overrides. These include the ability of the model to override
the knowledge within the weights of the model, the ability to override (or
moderate) extracted knowledge in the prompt, and lastly the ability to perform
a full jailbreak. Experimentation performed suggest several key findings to
improve instruction following - larger models perform the best in following
instructions that override internal and contextual instructions, and are
obedient, even to a fault. When scaling to longer contexts via rope scaling, a
significant buffer needs to be maintained from the edge of the perplexity cliff
in order to maintain instruction following capabilities. Finally, we observe
improving instruction following, and subsequently instruction
overrides/jailbreaks, is fundamentally at odds with the ability of a language
model to follow given safety filters or guidelines. Thus, we postulate the most
effective approach for safe, trustworthy AI should be dealt external to the LLM
itself. | cs.CL | Nevermind: Instruction Override and Moderation in Large Language Models |
2024-02-05T18:55:32Z | http://arxiv.org/pdf/2402.03300v3 | http://arxiv.org/abs/2402.03300v3 | Zhihong Shao, Peiyi Wang, Qihao Zhu, Runxin Xu, Junxiao Song, Xiao Bi, Haowei Zhang, Mingchuan Zhang, Y. K. Li, Y. Wu, Daya Guo | 2024-04-27T15:25:53Z | cs.CL, cs.AI, cs.LG | Mathematical reasoning poses a significant challenge for language models due
to its complex and structured nature. In this paper, we introduce DeepSeekMath
7B, which continues pre-training DeepSeek-Coder-Base-v1.5 7B with 120B
math-related tokens sourced from Common Crawl, together with natural language
and code data. DeepSeekMath 7B has achieved an impressive score of 51.7% on the
competition-level MATH benchmark without relying on external toolkits and
voting techniques, approaching the performance level of Gemini-Ultra and GPT-4.
Self-consistency over 64 samples from DeepSeekMath 7B achieves 60.9% on MATH.
The mathematical reasoning capability of DeepSeekMath is attributed to two key
factors: First, we harness the significant potential of publicly available web
data through a meticulously engineered data selection pipeline. Second, we
introduce Group Relative Policy Optimization (GRPO), a variant of Proximal
Policy Optimization (PPO), that enhances mathematical reasoning abilities while
concurrently optimizing the memory usage of PPO. | cs.CL | DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models |
2024-02-05T18:54:43Z | http://arxiv.org/pdf/2402.03299v4 | http://arxiv.org/abs/2402.03299v4 | Haibo Jin, Ruoxi Chen, Andy Zhou, Yang Zhang, Haohan Wang | 2024-05-30T21:14:26Z | cs.LG, cs.CL, cs.CV | The discovery of "jailbreaks" to bypass safety filters of Large Language
Models (LLMs) and harmful responses have encouraged the community to implement
safety measures. One major safety measure is to proactively test the LLMs with
jailbreaks prior to the release. Therefore, such testing will require a method
that can generate jailbreaks massively and efficiently. In this paper, we
follow a novel yet intuitive strategy to generate jailbreaks in the style of
the human generation. We propose a role-playing system that assigns four
different roles to the user LLMs to collaborate on new jailbreaks. Furthermore,
we collect existing jailbreaks and split them into different independent
characteristics using clustering frequency and semantic patterns sentence by
sentence. We organize these characteristics into a knowledge graph, making them
more accessible and easier to retrieve. Our system of different roles will
leverage this knowledge graph to generate new jailbreaks, which have proved
effective in inducing LLMs to generate unethical or guideline-violating
responses. In addition, we also pioneer a setting in our system that will
automatically follow the government-issued guidelines to generate jailbreaks to
test whether LLMs follow the guidelines accordingly. We refer to our system as
GUARD (Guideline Upholding through Adaptive Role-play Diagnostics). We have
empirically validated the effectiveness of GUARD on three cutting-edge
open-sourced LLMs (Vicuna-13B, LongChat-7B, and Llama-2-7B), as well as a
widely-utilized commercial LLM (ChatGPT). Moreover, our work extends to the
realm of vision language models (MiniGPT-v2 and Gemini Vision Pro), showcasing
GUARD's versatility and contributing valuable insights for the development of
safer, more reliable LLM-based applications across diverse modalities. | cs.LG | GUARD: Role-playing to Generate Natural-language Jailbreakings to Test Guideline Adherence of Large Language Models |
2024-02-05T18:39:47Z | http://arxiv.org/pdf/2402.03284v1 | http://arxiv.org/abs/2402.03284v1 | Anthony Sicilia, Hyunwoo Kim, Khyathi Raghavi Chandu, Malihe Alikhani, Jack Hessel | 2024-02-05T18:39:47Z | cs.CL, cs.AI, cs.LG | Effective interlocutors account for the uncertain goals, beliefs, and
emotions of others. But even the best human conversationalist cannot perfectly
anticipate the trajectory of a dialogue. How well can language models represent
inherent uncertainty in conversations? We propose FortUne Dial, an expansion of
the long-standing "conversation forecasting" task: instead of just accuracy,
evaluation is conducted with uncertainty-aware metrics, effectively enabling
abstention on individual instances. We study two ways in which language models
potentially represent outcome uncertainty (internally, using scores and
directly, using tokens) and propose fine-tuning strategies to improve
calibration of both representations. Experiments on eight difficult negotiation
corpora demonstrate that our proposed fine-tuning strategies (a traditional
supervision strategy and an off-policy reinforcement learning strategy) can
calibrate smaller open-source models to compete with pre-trained models 10x
their size. | cs.CL | Deal, or no deal (or who knows)? Forecasting Uncertainty in Conversations using Large Language Models |
2024-02-05T18:28:44Z | http://arxiv.org/pdf/2402.03271v2 | http://arxiv.org/abs/2402.03271v2 | Zhiyuan Hu, Chumin Liu, Xidong Feng, Yilun Zhao, See-Kiong Ng, Anh Tuan Luu, Junxian He, Pang Wei Koh, Bryan Hooi | 2024-05-30T14:03:35Z | cs.CL, cs.AI, cs.LG | In the face of uncertainty, the ability to *seek information* is of
fundamental importance. In many practical applications, such as medical
diagnosis and troubleshooting, the information needed to solve the task is not
initially given and has to be actively sought by asking follow-up questions
(for example, a doctor asking a patient for more details about their symptoms).
In this work, we introduce Uncertainty of Thoughts (UoT), an algorithm to
augment large language models with the ability to actively seek information by
asking effective questions. UoT combines 1) an *uncertainty-aware simulation
approach* which enables the model to simulate possible future scenarios and how
likely they are to occur, 2) *uncertainty-based rewards* motivated by
information gain which incentivizes the model to seek information, and 3) a
*reward propagation scheme* to select the optimal question to ask in a way that
maximizes the expected reward. In experiments on medical diagnosis,
troubleshooting, and the `20 Questions` game, UoT achieves an average
performance improvement of 38.1% in the rate of successful task completion
across multiple LLMs compared with direct prompting and also improves
efficiency (i.e., the number of questions needed to complete the task). Our
code has been released [here](https://github.com/zhiyuanhubj/UoT) | cs.CL | Uncertainty of Thoughts: Uncertainty-Aware Planning Enhances Information Seeking in Large Language Models |
2024-02-05T18:27:27Z | http://arxiv.org/pdf/2402.03269v1 | http://arxiv.org/abs/2402.03269v1 | Masato Hagiwara, Marius Miron, Jen-Yu Liu | 2024-02-05T18:27:27Z | cs.SD, cs.CL, cs.LG, eess.AS | Traditionally, bioacoustics has relied on spectrograms and continuous,
per-frame audio representations for the analysis of animal sounds, also serving
as input to machine learning models. Meanwhile, the International Phonetic
Alphabet (IPA) system has provided an interpretable, language-independent
method for transcribing human speech sounds. In this paper, we introduce ISPA
(Inter-Species Phonetic Alphabet), a precise, concise, and interpretable system
designed for transcribing animal sounds into text. We compare acoustics-based
and feature-based methods for transcribing and classifying animal sounds,
demonstrating their comparable performance with baseline methods utilizing
continuous, dense audio representations. By representing animal sounds with
text, we effectively treat them as a "foreign language," and we show that
established human language ML paradigms and models, such as language models,
can be successfully applied to improve performance. | cs.SD | ISPA: Inter-Species Phonetic Alphabet for Transcribing Animal Sounds |
2024-02-05T18:25:51Z | http://arxiv.org/pdf/2402.03268v3 | http://arxiv.org/abs/2402.03268v3 | Xinyi Wang, Alfonso Amayuelas, Kexun Zhang, Liangming Pan, Wenhu Chen, William Yang Wang | 2024-06-20T18:46:06Z | cs.LG, cs.AI, cs.CL | Pre-trained language models (LMs) are able to perform complex reasoning
without explicit fine-tuning. To understand how pre-training with a next-token
prediction objective contributes to the emergence of such reasoning capability,
we propose that we can view an LM as deriving new conclusions by aggregating
indirect reasoning paths seen at pre-training time. We found this perspective
effective in two important cases of reasoning: logic reasoning with knowledge
graphs (KGs) and chain-of-thought (CoT) reasoning. More specifically, we
formalize the reasoning paths as random walk paths on the knowledge/reasoning
graphs. Analyses of learned LM distributions suggest that a weighted sum of
relevant random walk path probabilities is a reasonable way to explain how LMs
reason. Experiments and analysis on multiple KG and CoT datasets reveal the
effect of training on random walk paths and suggest that augmenting unlabeled
random walk reasoning paths can improve real-world multi-step reasoning
performance. code: https://github.com/WANGXinyiLinda/LM_random_walk | cs.LG | Understanding Reasoning Ability of Language Models From the Perspective of Reasoning Paths Aggregation |
2024-02-05T17:59:00Z | http://arxiv.org/pdf/2402.03244v2 | http://arxiv.org/abs/2402.03244v2 | Kolby Nottingham, Bodhisattwa Prasad Majumder, Bhavana Dalvi Mishra, Sameer Singh, Peter Clark, Roy Fox | 2024-06-22T18:58:09Z | cs.LG, cs.CL | Large language models (LLMs) have recently been used for sequential decision
making in interactive environments. However, leveraging environment reward
signals for continual LLM actor improvement is not straightforward. We propose
Skill Set Optimization (SSO) for improving LLM actor performance through
constructing and refining sets of transferable skills. SSO constructs skills by
extracting common subtrajectories with high rewards and generating subgoals and
instructions to represent each skill. These skills are provided to the LLM
actor in-context to reinforce behaviors with high rewards. Then, SSO further
refines the skill set by pruning skills that do not continue to result in high
rewards. We evaluate our method in the classic videogame NetHack and the text
environment ScienceWorld to demonstrate SSO's ability to optimize a set of
skills and perform in-context policy improvement. SSO outperforms baselines by
40% in our custom NetHack task and outperforms the previous state-of-the-art in
ScienceWorld by 35%. | cs.LG | Skill Set Optimization: Reinforcing Language Model Behavior via Transferable Skills |
2024-02-05T17:57:26Z | http://arxiv.org/pdf/2402.03242v1 | http://arxiv.org/abs/2402.03242v1 | Antoine Magron, Anna Dai, Mike Zhang, Syrielle Montariol, Antoine Bosselut | 2024-02-05T17:57:26Z | cs.CL | Recent approaches in skill matching, employing synthetic training data for
classification or similarity model training, have shown promising results,
reducing the need for time-consuming and expensive annotations. However,
previous synthetic datasets have limitations, such as featuring only one skill
per sentence and generally comprising short sentences. In this paper, we
introduce JobSkape, a framework to generate synthetic data that tackles these
limitations, specifically designed to enhance skill-to-taxonomy matching.
Within this framework, we create SkillSkape, a comprehensive open-source
synthetic dataset of job postings tailored for skill-matching tasks. We
introduce several offline metrics that show that our dataset resembles
real-world data. Additionally, we present a multi-step pipeline for skill
extraction and matching tasks using large language models (LLMs), benchmarking
against known supervised methodologies. We outline that the downstream
evaluation results on real-world data can beat baselines, underscoring its
efficacy and adaptability. | cs.CL | JOBSKAPE: A Framework for Generating Synthetic Job Postings to Enhance Skill Matching |
2024-02-05T17:36:19Z | http://arxiv.org/pdf/2402.03223v4 | http://arxiv.org/abs/2402.03223v4 | Patrick Bareiß, Roman Klinger, Jeremy Barnes | 2024-03-07T17:18:18Z | cs.CL | Emotion classification in text is a challenging task due to the processes
involved when interpreting a textual description of a potential emotion
stimulus. In addition, the set of emotion categories is highly domain-specific.
For instance, literature analysis might require the use of aesthetic emotions
(e.g., finding something beautiful), and social media analysis could benefit
from fine-grained sets (e.g., separating anger from annoyance) than only those
that represent basic categories as they have been proposed by Paul Ekman
(anger, disgust, fear, joy, surprise, sadness). This renders the task an
interesting field for zero-shot classifications, in which the label set is not
known at model development time. Unfortunately, most resources for emotion
analysis are English, and therefore, most studies on emotion analysis have been
performed in English, including those that involve prompting language models
for text labels. This leaves us with a research gap that we address in this
paper: In which language should we prompt for emotion labels on non-English
texts? This is particularly of interest when we have access to a multilingual
large language model, because we could request labels with English prompts even
for non-English data. Our experiments with natural language inference-based
language models show that it is consistently better to use English prompts even
if the data is in a different language. | cs.CL | English Prompts are Better for NLI-based Zero-Shot Emotion Classification than Target-Language Prompts |
2024-02-05T17:33:22Z | http://arxiv.org/pdf/2402.03221v1 | http://arxiv.org/abs/2402.03221v1 | Huy Nghiem, Umang Gupta, Fred Morstatter | 2024-02-05T17:33:22Z | cs.CL | The propagation of offensive content through social media channels has
garnered attention of the research community. Multiple works have proposed
various semantically related yet subtle distinct categories of offensive
speech. In this work, we explore meta-earning approaches to leverage the
diversity of offensive speech corpora to enhance their reliable and efficient
detection. We propose a joint embedding architecture that incorporates the
input's label and definition for classification via Prototypical Network. Our
model achieves at least 75% of the maximal F1-score while using less than 10%
of the available training data across 4 datasets. Our experimental findings
also provide a case study of training strategies valuable to combat resource
scarcity. | cs.CL | "Define Your Terms" : Enhancing Efficient Offensive Speech Classification with Definition |
2024-02-05T17:26:49Z | http://arxiv.org/pdf/2402.03216v4 | http://arxiv.org/abs/2402.03216v4 | Jianlv Chen, Shitao Xiao, Peitian Zhang, Kun Luo, Defu Lian, Zheng Liu | 2024-06-28T09:55:49Z | cs.CL, cs.AI, cs.LG | In this paper, we present a new embedding model, called M3-Embedding, which
is distinguished for its versatility in Multi-Linguality, Multi-Functionality,
and Multi-Granularity. It can support more than 100 working languages, leading
to new state-of-the-art performances on multi-lingual and cross-lingual
retrieval tasks. It can simultaneously perform the three common retrieval
functionalities of embedding model: dense retrieval, multi-vector retrieval,
and sparse retrieval, which provides a unified model foundation for real-world
IR applications. It is able to process inputs of different granularities,
spanning from short sentences to long documents of up to 8192 tokens. The
effective training of M3-Embedding involves the following technical
contributions. We propose a novel self-knowledge distillation approach, where
the relevance scores from different retrieval functionalities can be integrated
as the teacher signal to enhance the training quality. We also optimize the
batching strategy, enabling a large batch size and high training throughput to
ensure the discriminativeness of embeddings. To the best of our knowledge,
M3-Embedding is the first embedding model which realizes such a strong
versatility. The model and code will be publicly available at
https://github.com/FlagOpen/FlagEmbedding. | cs.CL | BGE M3-Embedding: Multi-Lingual, Multi-Functionality, Multi-Granularity Text Embeddings Through Self-Knowledge Distillation |
2024-02-05T16:57:24Z | http://arxiv.org/pdf/2402.03191v3 | http://arxiv.org/abs/2402.03191v3 | Timothee Mickus, Stig-Arne Grönroos, Joseph Attieh | 2024-05-27T10:21:08Z | cs.LG, cs.CL | Whether embedding spaces use all their dimensions equally, i.e., whether they
are isotropic, has been a recent subject of discussion. Evidence has been
accrued both for and against enforcing isotropy in embedding spaces. In the
present paper, we stress that isotropy imposes requirements on the embedding
space that are not compatible with the presence of clusters -- which also
negatively impacts linear classification objectives. We demonstrate this fact
both mathematically and empirically and use it to shed light on previous
results from the literature. | cs.LG | Isotropy, Clusters, and Classifiers |
2024-02-05T16:56:11Z | http://arxiv.org/pdf/2402.03190v4 | http://arxiv.org/abs/2402.03190v4 | Xiang Chen, Chenxi Wang, Yida Xue, Ningyu Zhang, Xiaoyan Yang, Qiang Li, Yue Shen, Lei Liang, Jinjie Gu, Huajun Chen | 2024-05-27T11:52:56Z | cs.CL, cs.AI, cs.IR, cs.LG, cs.MM | Despite significant strides in multimodal tasks, Multimodal Large Language
Models (MLLMs) are plagued by the critical issue of hallucination. The reliable
detection of such hallucinations in MLLMs has, therefore, become a vital aspect
of model evaluation and the safeguarding of practical application deployment.
Prior research in this domain has been constrained by a narrow focus on
singular tasks, an inadequate range of hallucination categories addressed, and
a lack of detailed granularity. In response to these challenges, our work
expands the investigative horizons of hallucination detection. We present a
novel meta-evaluation benchmark, MHaluBench, meticulously crafted to facilitate
the evaluation of advancements in hallucination detection methods.
Additionally, we unveil a novel unified multimodal hallucination detection
framework, UNIHD, which leverages a suite of auxiliary tools to validate the
occurrence of hallucinations robustly. We demonstrate the effectiveness of
UNIHD through meticulous evaluation and comprehensive analysis. We also provide
strategic insights on the application of specific tools for addressing various
categories of hallucinations. | cs.CL | Unified Hallucination Detection for Multimodal Large Language Models |
2024-02-05T16:47:17Z | http://arxiv.org/pdf/2402.05130v2 | http://arxiv.org/abs/2402.05130v2 | Yan Zhao, Zhongyun Li, Yushan Pan, Jiaxing Wang, Yihong Wang | 2024-02-09T02:45:51Z | cs.CL, cs.AI | Generative Artificial Intelligence (AI), because of its emergent abilities,
has empowered various fields, one typical of which is large language models
(LLMs). One of the typical application fields of Generative AI is large
language models (LLMs), and the natural language understanding capability of
LLM is dramatically improved when compared with conventional AI-based methods.
The natural language understanding capability has always been a barrier to the
intent recognition performance of the Knowledge-Based-Question-and-Answer
(KBQA) system, which arises from linguistic diversity and the newly appeared
intent. Conventional AI-based methods for intent recognition can be divided
into semantic parsing-based and model-based approaches. However, both of the
methods suffer from limited resources in intent recognition. To address this
issue, we propose a novel KBQA system based on a Large Language Model(LLM) and
BERT (LB-KBQA). With the help of generative AI, our proposed method could
detect newly appeared intent and acquire new knowledge. In experiments on
financial domain question answering, our model has demonstrated superior
effectiveness. | cs.CL | LB-KBQA: Large-language-model and BERT based Knowledge-Based Question and Answering System |
2024-02-05T16:46:16Z | http://arxiv.org/pdf/2402.03181v4 | http://arxiv.org/abs/2402.03181v4 | Mintong Kang, Nezihe Merve Gürel, Ning Yu, Dawn Song, Bo Li | 2024-06-04T04:51:08Z | cs.AI, cs.CL, cs.IR | Despite the impressive capabilities of large language models (LLMs) across
diverse applications, they still suffer from trustworthiness issues, such as
hallucinations and misalignments. Retrieval-augmented language models (RAG)
have been proposed to enhance the credibility of generations by grounding
external knowledge, but the theoretical understandings of their generation
risks remains unexplored. In this paper, we answer: 1) whether RAG can indeed
lead to low generation risks, 2) how to provide provable guarantees on the
generation risks of RAG and vanilla LLMs, and 3) what sufficient conditions
enable RAG models to reduce generation risks. We propose C-RAG, the first
framework to certify generation risks for RAG models. Specifically, we provide
conformal risk analysis for RAG models and certify an upper confidence bound of
generation risks, which we refer to as conformal generation risk. We also
provide theoretical guarantees on conformal generation risks for general
bounded risk functions under test distribution shifts. We prove that RAG
achieves a lower conformal generation risk than that of a single LLM when the
quality of the retrieval model and transformer is non-trivial. Our intensive
empirical results demonstrate the soundness and tightness of our conformal
generation risk guarantees across four widely-used NLP datasets on four
state-of-the-art retrieval models. | cs.AI | C-RAG: Certified Generation Risks for Retrieval-Augmented Language Models |
2024-02-05T16:44:17Z | http://arxiv.org/pdf/2402.03177v1 | http://arxiv.org/abs/2402.03177v1 | Zaid Alyafeai, Khalid Almubarak, Ahmed Ashraf, Deema Alnuhait, Saied Alshahrani, Gubran A. Q. Abdulrahman, Gamil Ahmed, Qais Gawah, Zead Saleh, Mustafa Ghaleb, Yousef Ali, Maged S. Al-Shaibani | 2024-02-05T16:44:17Z | cs.CL, cs.LG | Instruction tuning has emerged as a prominent methodology for teaching Large
Language Models (LLMs) to follow instructions. However, current instruction
datasets predominantly cater to English or are derived from English-dominated
LLMs, resulting in inherent biases toward Western culture. This bias
significantly impacts the linguistic structures of non-English languages such
as Arabic, which has a distinct grammar reflective of the diverse cultures
across the Arab region. This paper addresses this limitation by introducing
CIDAR: https://hf.co/datasets/arbml/CIDAR, the first open Arabic
instruction-tuning dataset culturally-aligned by human reviewers. CIDAR
contains 10,000 instruction and output pairs that represent the Arab region. We
discuss the cultural relevance of CIDAR via the analysis and comparison to
other models fine-tuned on other datasets. Our experiments show that CIDAR can
help enrich research efforts in aligning LLMs with the Arabic culture. All the
code is available at https://github.com/ARBML/CIDAR. | cs.CL | CIDAR: Culturally Relevant Instruction Dataset For Arabic |
2024-02-05T16:41:02Z | http://arxiv.org/pdf/2402.03173v2 | http://arxiv.org/abs/2402.03173v2 | Zichen Zhu, Yang Xu, Lu Chen, Jingkai Yang, Yichuan Ma, Yiming Sun, Hailin Wen, Jiaqi Liu, Jinyu Cai, Yingzi Ma, Situo Zhang, Zihan Zhao, Liangtai Sun, Kai Yu | 2024-02-20T07:55:52Z | cs.CL, cs.AI, cs.CV | Rapid progress in multimodal large language models (MLLMs) highlights the
need to introduce challenging yet realistic benchmarks to the academic
community, while existing benchmarks primarily focus on understanding simple
natural images and short context. In this paper, we present MULTI as a
cutting-edge benchmark for evaluating MLLMs on understanding complex tables and
images, and reasoning with long context. MULTI provides multimodal inputs and
requires responses that are either precise or open-ended, reflecting real-life
examination styles. MULTI includes over 18,000 questions and challenges MLLMs
with a variety of tasks, ranging from formula derivation to image detail
analysis and cross-modality reasoning. We also introduce MULTI-Elite, a
500-question selected hard subset, and MULTI-Extend, with more than 4,500
external knowledge context pieces. Our evaluation indicates significant
potential for MLLM advancement, with GPT-4V achieving a 63.7% accuracy rate on
MULTI, in contrast to other MLLMs scoring between 28.5% and 55.3%. MULTI serves
not only as a robust evaluation platform but also paves the way for the
development of expert-level AI. | cs.CL | MULTI: Multimodal Understanding Leaderboard with Text and Images |
2024-02-05T16:40:23Z | http://arxiv.org/pdf/2402.03172v1 | http://arxiv.org/abs/2402.03172v1 | Gonçalo Gomes, Isabel Coutinho, Bruno Martins | 2024-02-05T16:40:23Z | cs.CL, cs.AI | Although the International Classification of Diseases (ICD) has been adopted
worldwide, manually assigning ICD codes to clinical text is time-consuming,
error-prone, and expensive, motivating the development of automated approaches.
This paper describes a novel approach for automated ICD coding, combining
several ideas from previous related work. We specifically employ a strong
Transformer-based model as a text encoder and, to handle lengthy clinical
narratives, we explored either (a) adapting the base encoder model into a
Longformer, or (b) dividing the text into chunks and processing each chunk
independently. The representations produced by the encoder are combined with a
label embedding mechanism that explores diverse ICD code synonyms. Experiments
with different splits of the MIMIC-III dataset show that the proposed approach
outperforms the current state-of-the-art models in ICD coding, with the label
embeddings significantly contributing to the good performance. Our approach
also leads to properly calibrated classification results, which can effectively
inform downstream tasks such as quantification. | cs.CL | Accurate and Well-Calibrated ICD Code Assignment Through Attention Over Diverse Label Embeddings |
2024-02-05T16:39:15Z | http://arxiv.org/pdf/2402.03171v1 | http://arxiv.org/abs/2402.03171v1 | Fatima Zahra Qachfar, Rakesh M. Verma | 2024-02-05T16:39:15Z | cs.CL, cs.CR, cs.LG | We examine the impact of homograph attacks on the Sentiment Analysis (SA)
task of different Arabic dialects from the Maghreb North-African countries.
Homograph attacks result in a 65.3% decrease in transformer classification from
an F1-score of 0.95 to 0.33 when data is written in "Arabizi". The goal of this
study is to highlight LLMs weaknesses' and to prioritize ethical and
responsible Machine Learning. | cs.CL | Homograph Attacks on Maghreb Sentiment Analyzers |
2024-02-05T16:31:03Z | http://arxiv.org/pdf/2402.03163v1 | http://arxiv.org/abs/2402.03163v1 | Adrian-Gabriel Chifu, Sébastien Fournier | 2024-02-05T16:31:03Z | cs.CL, cs.IR | One of the challenges of natural language understanding is to deal with the
subjectivity of sentences, which may express opinions and emotions that add
layers of complexity and nuance. Sentiment analysis is a field that aims to
extract and analyze these subjective elements from text, and it can be applied
at different levels of granularity, such as document, paragraph, sentence, or
aspect. Aspect-based sentiment analysis is a well-studied topic with many
available data sets and models. However, there is no clear definition of what
makes a sentence difficult for aspect-based sentiment analysis. In this paper,
we explore this question by conducting an experiment with three data sets:
"Laptops", "Restaurants", and "MTSC" (Multi-Target-dependent Sentiment
Classification), and a merged version of these three datasets. We study the
impact of domain diversity and syntactic diversity on difficulty. We use a
combination of classifiers to identify the most difficult sentences and analyze
their characteristics. We employ two ways of defining sentence difficulty. The
first one is binary and labels a sentence as difficult if the classifiers fail
to correctly predict the sentiment polarity. The second one is a six-level
scale based on how many of the top five best-performing classifiers can
correctly predict the sentiment polarity. We also define 9 linguistic features
that, combined, aim at estimating the difficulty at sentence level. | cs.CL | Linguistic features for sentence difficulty prediction in ABSA |
2024-02-05T16:30:49Z | http://arxiv.org/pdf/2402.03161v3 | http://arxiv.org/abs/2402.03161v3 | Yang Jin, Zhicheng Sun, Kun Xu, Kun Xu, Liwei Chen, Hao Jiang, Quzhe Huang, Chengru Song, Yuliang Liu, Di Zhang, Yang Song, Kun Gai, Yadong Mu | 2024-06-03T08:09:09Z | cs.CV, cs.CL | In light of recent advances in multimodal Large Language Models (LLMs), there
is increasing attention to scaling them from image-text data to more
informative real-world videos. Compared to static images, video poses unique
challenges for effective large-scale pre-training due to the modeling of its
spatiotemporal dynamics. In this paper, we address such limitations in
video-language pre-training with an efficient video decomposition that
represents each video as keyframes and temporal motions. These are then adapted
to an LLM using well-designed tokenizers that discretize visual and temporal
information as a few tokens, thus enabling unified generative pre-training of
videos, images, and text. At inference, the generated tokens from the LLM are
carefully recovered to the original continuous pixel space to create various
video content. Our proposed framework is both capable of comprehending and
generating image and video content, as demonstrated by its competitive
performance across 13 multimodal benchmarks in image and video understanding
and generation. Our code and models are available at
https://video-lavit.github.io. | cs.CV | Video-LaVIT: Unified Video-Language Pre-training with Decoupled Visual-Motional Tokenization |
2024-02-05T16:11:43Z | http://arxiv.org/pdf/2402.03142v2 | http://arxiv.org/abs/2402.03142v2 | Michele Mastromattei, Fabio Massimo Zanzotto | 2024-06-09T10:32:03Z | cs.LG, cs.CL | Neural network pruning has become increasingly crucial due to the complexity
of these models and their widespread use in various fields. Existing pruning
algorithms often suffer from limitations such as architecture specificity,
excessive complexity and reliance on demanding calculations, rendering them
impractical for real-world applications. This paper introduces KEN: a
straightforward, universal and unstructured pruning algorithm based on Kernel
Density Estimation (KDE). KEN aims to construct optimized transformers by
selectively preserving the most significant parameters while restoring others
to their pre-training state. This strategy preserves model performance while
enabling storage of only the optimized subnetwork, leading to substantial
memory savings. Extensive evaluations across seven different LLMs demonstrate
that KEN achieves equal or better performance than their original unpruned
versions, with a minimum parameter reduction of 25%. Furthermore, in-depth
comparisons with established pruning and PEFT algorithms confirm KEN
effectiveness. We further introduce KEN$_{viz}$, an explainable tool that
visualizes the optimized model composition achieved by KEN from different
points of view. | cs.LG | Less is KEN: a Universal and Simple Non-Parametric Pruning Algorithm for Large Language Models |
2024-02-05T16:05:32Z | http://arxiv.org/pdf/2402.03137v1 | http://arxiv.org/abs/2402.03137v1 | Kushal Tatariya, Heather Lent, Johannes Bjerva, Miryam de Lhoneux | 2024-02-05T16:05:32Z | cs.CL, cs.LG | Emotion classification is a challenging task in NLP due to the inherent
idiosyncratic and subjective nature of linguistic expression, especially with
code-mixed data. Pre-trained language models (PLMs) have achieved high
performance for many tasks and languages, but it remains to be seen whether
these models learn and are robust to the differences in emotional expression
across languages. Sociolinguistic studies have shown that Hinglish speakers
switch to Hindi when expressing negative emotions and to English when
expressing positive emotions. To understand if language models can learn these
associations, we study the effect of language on emotion prediction across 3
PLMs on a Hinglish emotion classification dataset. Using LIME and token level
language ID, we find that models do learn these associations between language
choice and emotional expression. Moreover, having code-mixed data present in
the pre-training can augment that learning when task-specific data is scarce.
We also conclude from the misclassifications that the models may overgeneralise
this heuristic to other infrequent examples where this sociolinguistic
phenomenon does not apply. | cs.CL | Sociolinguistically Informed Interpretability: A Case Study on Hinglish Emotion Classification |
2024-02-05T15:57:32Z | http://arxiv.org/pdf/2402.03131v1 | http://arxiv.org/abs/2402.03131v1 | Duong Minh Le, Yang Chen, Alan Ritter, Wei Xu | 2024-02-05T15:57:32Z | cs.CL, cs.LG | Zero-shot cross-lingual transfer utilizing multilingual LLMs has become a
popular learning paradigm for low-resource languages with no labeled training
data. However, for NLP tasks that involve fine-grained predictions on words and
phrases, the performance of zero-shot cross-lingual transfer learning lags far
behind supervised fine-tuning methods. Therefore, it is common to exploit
translation and label projection to further improve the performance by (1)
translating training data that is available in a high-resource language (e.g.,
English) together with the gold labels into low-resource languages, and/or (2)
translating test data in low-resource languages to a high-source language to
run inference on, then projecting the predicted span-level labels back onto the
original test data. However, state-of-the-art marker-based label projection
methods suffer from translation quality degradation due to the extra label
markers injected in the input to the translation model. In this work, we
explore a new direction that leverages constrained decoding for label
projection to overcome the aforementioned issues. Our new method not only can
preserve the quality of translated texts but also has the versatility of being
applicable to both translating training and translating test data strategies.
This versatility is crucial as our experiments reveal that translating test
data can lead to a considerable boost in performance compared to translating
only training data. We evaluate on two cross-lingual transfer tasks, namely
Named Entity Recognition and Event Argument Extraction, spanning 20 languages.
The results demonstrate that our approach outperforms the state-of-the-art
marker-based method by a large margin and also shows better performance than
other label projection methods that rely on external word alignment. | cs.CL | Constrained Decoding for Cross-lingual Label Projection |
2024-02-05T15:43:50Z | http://arxiv.org/pdf/2402.05129v1 | http://arxiv.org/abs/2402.05129v1 | Petter Törnberg | 2024-02-05T15:43:50Z | cs.CL | Large Language Models (LLMs) have ushered in a new era of text annotation, as
their ease-of-use, high accuracy, and relatively low costs have meant that
their use has exploded in recent months. However, the rapid growth of the field
has meant that LLM-based annotation has become something of an academic Wild
West: the lack of established practices and standards has led to concerns about
the quality and validity of research. Researchers have warned that the
ostensible simplicity of LLMs can be misleading, as they are prone to bias,
misunderstandings, and unreliable results. Recognizing the transformative
potential of LLMs, this paper proposes a comprehensive set of standards and
best practices for their reliable, reproducible, and ethical use. These
guidelines span critical areas such as model selection, prompt engineering,
structured prompting, prompt stability analysis, rigorous model validation, and
the consideration of ethical and legal implications. The paper emphasizes the
need for a structured, directed, and formalized approach to using LLMs, aiming
to ensure the integrity and robustness of text annotation practices, and
advocates for a nuanced and critical engagement with LLMs in social scientific
research. | cs.CL | Best Practices for Text Annotation with Large Language Models |
2024-02-05T15:28:43Z | http://arxiv.org/pdf/2402.03099v1 | http://arxiv.org/abs/2402.03099v1 | Elad Levi, Eli Brosh, Matan Friedmann | 2024-02-05T15:28:43Z | cs.CL, cs.AI, cs.LG | Prompt engineering is a challenging and important task due to the high
sensitivity of Large Language Models (LLMs) to the given prompt and the
inherent ambiguity of a textual task instruction. Automatic prompt engineering
is essential to achieve optimized performance from LLMs. Recent studies have
demonstrated the capabilities of LLMs to automatically conduct prompt
engineering by employing a meta-prompt that incorporates the outcomes of the
last trials and proposes an improved prompt. However, this requires a
high-quality benchmark to compare different prompts, which is difficult and
expensive to acquire in many real-world use cases. In this work, we introduce a
new method for automatic prompt engineering, using a calibration process that
iteratively refines the prompt to the user intent. During the optimization
process, the system jointly generates synthetic data of boundary use cases and
optimizes the prompt according to the generated dataset. We demonstrate the
effectiveness of our method with respect to strong proprietary models on
real-world tasks such as moderation and generation. Our method outperforms
state-of-the-art methods with a limited number of annotated samples.
Furthermore, we validate the advantages of each one of the system's key
components. Our system is built in a modular way, facilitating easy adaptation
to other tasks. The code is available
$\href{https://github.com/Eladlev/AutoPrompt}{here}$. | cs.CL | Intent-based Prompt Calibration: Enhancing prompt optimization with synthetic boundary cases |
2024-02-05T15:08:19Z | http://arxiv.org/pdf/2402.03407v1 | http://arxiv.org/abs/2402.03407v1 | Álvaro Martín-Cortinas, Daniel Sáez-Trigueros, Iván Vallés-Pérez, Biel Tura-Vecino, Piotr Biliński, Mateusz Lajszczak, Grzegorz Beringer, Roberto Barra-Chicote, Jaime Lorenzo-Trueba | 2024-02-05T15:08:19Z | eess.AS, cs.CL, cs.LG | Large Language Models (LLMs) are one of the most promising technologies for
the next era of speech generation systems, due to their scalability and
in-context learning capabilities. Nevertheless, they suffer from multiple
stability issues at inference time, such as hallucinations, content skipping or
speech repetitions. In this work, we introduce a new self-supervised Voice
Conversion (VC) architecture which can be used to learn to encode transitory
features, such as content, separately from stationary ones, such as speaker ID
or recording conditions, creating speaker-disentangled representations. Using
speaker-disentangled codes to train LLMs for text-to-speech (TTS) allows the
LLM to generate the content and the style of the speech only from the text,
similarly to humans, while the speaker identity is provided by the decoder of
the VC model. Results show that LLMs trained over speaker-disentangled
self-supervised representations provide an improvement of 4.7pp in speaker
similarity over SOTA entangled representations, and a word error rate (WER)
5.4pp lower. Furthermore, they achieve higher naturalness than human recordings
of the LibriTTS test-other dataset. Finally, we show that using explicit
reference embedding negatively impacts intelligibility (stability), with WER
increasing by 14pp compared to the model that only uses text to infer the
style. | eess.AS | Enhancing the Stability of LLM-based Speech Generation Systems through Self-Supervised Representations |
2024-02-05T14:59:29Z | http://arxiv.org/pdf/2402.03067v1 | http://arxiv.org/abs/2402.03067v1 | Darija Medvecki, Bojana Bašaragin, Adela Ljajić, Nikola Milošević | 2024-02-05T14:59:29Z | cs.CL, cs.AI | This paper presents the results of the first application of BERTopic, a
state-of-the-art topic modeling technique, to short text written in a
morphologi-cally rich language. We applied BERTopic with three multilingual
embed-ding models on two levels of text preprocessing (partial and full) to
evalu-ate its performance on partially preprocessed short text in Serbian. We
also compared it to LDA and NMF on fully preprocessed text. The experiments
were conducted on a dataset of tweets expressing hesitancy toward COVID-19
vaccination. Our results show that with adequate parameter setting, BERTopic
can yield informative topics even when applied to partially pre-processed short
text. When the same parameters are applied in both prepro-cessing scenarios,
the performance drop on partially preprocessed text is minimal. Compared to LDA
and NMF, judging by the keywords, BERTopic offers more informative topics and
gives novel insights when the number of topics is not limited. The findings of
this paper can be significant for re-searchers working with other
morphologically rich low-resource languages and short text. | cs.CL | Multilingual transformer and BERTopic for short text topic modeling: The case of Serbian |
2024-02-05T14:36:51Z | http://arxiv.org/pdf/2402.03053v1 | http://arxiv.org/abs/2402.03053v1 | Husein Zolkepli, Aisyah Razak, Kamarul Adha, Ariff Nazhan | 2024-02-05T14:36:51Z | cs.CL, cs.LG | In this work, we present a comprehensive exploration of finetuning Malaysian
language models, specifically Llama2 and Mistral, on embedding tasks involving
negative and positive pairs. We release two distinct models tailored for
Semantic Similarity and Retrieval-Augmented Generation (RAG).
For Semantic Similarity, our 600 million parameter Llama2 model outperforms
OpenAI text-embedding-ada-002 across all recall@k metrics for b.cari.com.my,
c.cari.com.my, Malay news, and Malaysian Twitter test sets.
In the realm of RAG models, our approach proves competitive with OpenAI
text-embedding-ada-002 in the Malaysian context. Notably, our 2 billion
parameter Llama2 model achieves superior Recall@5, Recall@10 for the "Melayu"
keyword research papers dataset and excels in Recall@3, Recall@5, and Recall@10
for the lom.agc.gov.my dataset.
These findings underscore the effectiveness of our finetuning strategy and
highlight the performance gains in both Semantic Similarity and RAG tasks.
All models released at
https://huggingface.co/collections/mesolitica/malaysian-embedding-6523612bfe5881ad35f81b99 | cs.CL | Multi-Lingual Malaysian Embedding: Leveraging Large Language Models for Semantic Representations |
2024-02-05T14:34:14Z | http://arxiv.org/pdf/2402.03050v1 | http://arxiv.org/abs/2402.03050v1 | Rupak Raj Ghimire, Bal Krishna Bal, Prakash Poudyal | 2024-02-05T14:34:14Z | cs.SD, cs.CL, eess.AS | In this paper, we examine the research conducted in the field of Nepali
Automatic Speech Recognition (ASR). The primary objective of this survey is to
conduct a comprehensive review of the works on Nepali Automatic Speech
Recognition Systems completed to date, explore the different datasets used,
examine the technology utilized, and take account of the obstacles encountered
in implementing the Nepali ASR system. In tandem with the global trends of
ever-increasing research on speech recognition based research, the number of
Nepalese ASR-related projects are also growing. Nevertheless, the investigation
of language and acoustic models of the Nepali language has not received
adequate attention compared to languages that possess ample resources. In this
context, we provide a framework as well as directions for future
investigations. | cs.SD | A Comprehensive Study of the Current State-of-the-Art in Nepali Automatic Speech Recognition Systems |
2024-02-05T14:33:56Z | http://arxiv.org/pdf/2402.03049v4 | http://arxiv.org/abs/2402.03049v4 | Yixin Ou, Ningyu Zhang, Honghao Gui, Ziwen Xu, Shuofei Qiao, Yida Xue, Runnan Fang, Kangwei Liu, Lei Li, Zhen Bi, Guozhou Zheng, Huajun Chen | 2024-06-24T02:10:23Z | cs.CL, cs.AI, cs.HC, cs.IR, cs.LG | In recent years, instruction tuning has gained increasing attention and
emerged as a crucial technique to enhance the capabilities of Large Language
Models (LLMs). To construct high-quality instruction datasets, many instruction
processing approaches have been proposed, aiming to achieve a delicate balance
between data quantity and data quality. Nevertheless, due to inconsistencies
that persist among various instruction processing methods, there is no standard
open-source instruction processing implementation framework available for the
community, which hinders practitioners from further developing and advancing.
To facilitate instruction processing research and development, we present
EasyInstruct, an easy-to-use instruction processing framework for LLMs, which
modularizes instruction generation, selection, and prompting, while also
considering their combination and interaction. EasyInstruct is publicly
released and actively maintained at https://github.com/zjunlp/EasyInstruct,
along with an online demo app and a demo video for quick-start, calling for
broader research centered on instruction data and synthetic data. | cs.CL | EasyInstruct: An Easy-to-use Instruction Processing Framework for Large Language Models |
2024-02-05T14:29:54Z | http://arxiv.org/pdf/2402.03043v1 | http://arxiv.org/abs/2402.03043v1 | Mohammad N. S. Jahromi, Satya. M. Muddamsetty, Asta Sofie Stage Jarlner, Anna Murphy Høgenhaug, Thomas Gammeltoft-Hansen, Thomas B. Moeslund | 2024-02-05T14:29:54Z | cs.CL, cs.LG | Explainable AI (XAI) aids in deciphering 'black-box' models. While several
methods have been proposed and evaluated primarily in the image domain, the
exploration of explainability in the text domain remains a growing research
area. In this paper, we delve into the applicability of XAI methods for the
text domain. In this context, the 'Similarity Difference and Uniqueness' (SIDU)
XAI method, recognized for its superior capability in localizing entire salient
regions in image-based classification is extended to textual data. The extended
method, SIDU-TXT, utilizes feature activation maps from 'black-box' models to
generate heatmaps at a granular, word-based level, thereby providing
explanations that highlight contextually significant textual elements crucial
for model predictions. Given the absence of a unified standard for assessing
XAI methods, this study applies a holistic three-tiered comprehensive
evaluation framework: Functionally-Grounded, Human-Grounded and
Application-Grounded, to assess the effectiveness of the proposed SIDU-TXT
across various experiments. We find that, in sentiment analysis task of a movie
review dataset, SIDU-TXT excels in both functionally and human-grounded
evaluations, demonstrating superior performance through quantitative and
qualitative analyses compared to benchmarks like Grad-CAM and LIME. In the
application-grounded evaluation within the sensitive and complex legal domain
of asylum decision-making, SIDU-TXT and Grad-CAM demonstrate comparable
performances, each with its own set of strengths and weaknesses. However, both
methods fall short of entirely fulfilling the sophisticated criteria of expert
expectations, highlighting the imperative need for additional research in XAI
methods suitable for such domains. | cs.CL | SIDU-TXT: An XAI Algorithm for NLP with a Holistic Assessment Approach |
2024-02-05T14:23:43Z | http://arxiv.org/pdf/2402.03038v1 | http://arxiv.org/abs/2402.03038v1 | Branislav Pecher, Ivan Srba, Maria Bielikova, Joaquin Vanschoren | 2024-02-05T14:23:43Z | cs.LG, cs.AI, cs.CL | In few-shot learning, such as meta-learning, few-shot fine-tuning or
in-context learning, the limited number of samples used to train a model have a
significant impact on the overall success. Although a large number of sample
selection strategies exist, their impact on the performance of few-shot
learning is not extensively known, as most of them have been so far evaluated
in typical supervised settings only. In this paper, we thoroughly investigate
the impact of 20 sample selection strategies on the performance of 5 few-shot
learning approaches over 8 image and 6 text datasets. In addition, we propose a
new method for automatic combination of sample selection strategies (ACSESS)
that leverages the strengths and complementary information of the individual
strategies. The experimental results show that our method consistently
outperforms the individual selection strategies, as well as the recently
proposed method for selecting support examples for in-context learning. We also
show a strong modality, dataset and approach dependence for the majority of
strategies as well as their dependence on the number of shots - demonstrating
that the sample selection strategies play a significant role for lower number
of shots, but regresses to random selection at higher number of shots. | cs.LG | Automatic Combination of Sample Selection Strategies for Few-Shot Learning |
2024-02-05T13:47:53Z | http://arxiv.org/pdf/2402.03009v1 | http://arxiv.org/abs/2402.03009v1 | Junjie Fang, Likai Tang, Hongzhe Bi, Yujia Qin, Si Sun, Zhenyu Li, Haolun Li, Yongjian Li, Xin Cong, Yukun Yan, Xiaodong Shi, Sen Song, Yankai Lin, Zhiyuan Liu, Maosong Sun | 2024-02-05T13:47:53Z | cs.CL, cs.AI | Long-context processing is a critical ability that constrains the
applicability of large language models. Although there exist various methods
devoted to enhancing the long-context processing ability of large language
models (LLMs), they are developed in an isolated manner and lack systematic
analysis and integration of their strengths, hindering further developments. In
this paper, we introduce UniMem, a unified framework that reformulates existing
long-context methods from the view of memory augmentation of LLMs. UniMem is
characterized by four key dimensions: Memory Management, Memory Writing, Memory
Reading, and Memory Injection, providing a systematic theory for understanding
various long-context methods. We reformulate 16 existing methods based on
UniMem and analyze four representative methods: Transformer-XL, Memorizing
Transformer, RMT, and Longformer into equivalent UniMem forms to reveal their
design principles and strengths. Based on these analyses, we propose UniMix, an
innovative approach that integrates the strengths of these algorithms.
Experimental results show that UniMix achieves superior performance in handling
long contexts with significantly lower perplexity than baselines. | cs.CL | UniMem: Towards a Unified View of Long-Context Large Language Models |
2024-02-05T13:31:28Z | http://arxiv.org/pdf/2402.02992v2 | http://arxiv.org/abs/2402.02992v2 | Tianlin Liu, Shangmin Guo, Leonardo Bianco, Daniele Calandriello, Quentin Berthet, Felipe Llinares, Jessica Hoffmann, Lucas Dixon, Michal Valko, Mathieu Blondel | 2024-05-24T08:39:07Z | cs.LG, cs.AI, cs.CL | Aligning language models with human preferences is crucial for reducing
errors and biases in these models. Alignment techniques, such as reinforcement
learning from human feedback (RLHF), are typically cast as optimizing a
tradeoff between human preference rewards and a proximity regularization term
that encourages staying close to the unaligned model. Selecting an appropriate
level of regularization is critical: insufficient regularization can lead to
reduced model capabilities due to reward hacking, whereas excessive
regularization hinders alignment. Traditional methods for finding the optimal
regularization level require retraining multiple models with varying
regularization strengths. This process, however, is resource-intensive,
especially for large models. To address this challenge, we propose
decoding-time realignment (DeRa), a simple method to explore and evaluate
different regularization strengths in aligned models without retraining. DeRa
enables control over the degree of alignment, allowing users to smoothly
transition between unaligned and aligned models. It also enhances the
efficiency of hyperparameter tuning by enabling the identification of effective
regularization strengths using a validation dataset. | cs.LG | Decoding-time Realignment of Language Models |
2024-02-05T13:18:42Z | http://arxiv.org/pdf/2402.02987v1 | http://arxiv.org/abs/2402.02987v1 | Junjie Chu, Zeyang Sha, Michael Backes, Yang Zhang | 2024-02-05T13:18:42Z | cs.CR, cs.AI, cs.CL, cs.LG | In recent times, significant advancements have been made in the field of
large language models (LLMs), represented by GPT series models. To optimize
task execution, users often engage in multi-round conversations with GPT models
hosted in cloud environments. These multi-round conversations, potentially
replete with private information, require transmission and storage within the
cloud. However, this operational paradigm introduces additional attack
surfaces. In this paper, we first introduce a specific Conversation
Reconstruction Attack targeting GPT models. Our introduced Conversation
Reconstruction Attack is composed of two steps: hijacking a session and
reconstructing the conversations. Subsequently, we offer an exhaustive
evaluation of the privacy risks inherent in conversations when GPT models are
subjected to the proposed attack. However, GPT-4 demonstrates certain
robustness to the proposed attacks. We then introduce two advanced attacks
aimed at better reconstructing previous conversations, specifically the UNR
attack and the PBU attack. Our experimental findings indicate that the PBU
attack yields substantial performance across all models, achieving semantic
similarity scores exceeding 0.60, while the UNR attack is effective solely on
GPT-3.5. Our results reveal the concern about privacy risks associated with
conversations involving GPT models and aim to draw the community's attention to
prevent the potential misuse of these models' remarkable capabilities. We will
responsibly disclose our findings to the suppliers of related large language
models. | cs.CR | Conversation Reconstruction Attack Against GPT Models |
2024-02-05T12:56:22Z | http://arxiv.org/pdf/2402.02975v1 | http://arxiv.org/abs/2402.02975v1 | Nicolò Penzo, Antonio Longa, Bruno Lepri, Sara Tonelli, Marco Guerini | 2024-02-05T12:56:22Z | cs.CL | Current text classification approaches usually focus on the content to be
classified. Contextual aspects (both linguistic and extra-linguistic) are
usually neglected, even in tasks based on online discussions. Still in many
cases the multi-party and multi-turn nature of the context from which these
elements are selected can be fruitfully exploited. In this work, we propose a
series of experiments on a large dataset for stance detection in English, in
which we evaluate the contribution of different types of contextual
information, i.e. linguistic, structural and temporal, by feeding them as
natural language input into a transformer-based model. We also experiment with
different amounts of training data and analyse the topology of local discussion
networks in a privacy-compliant way. Results show that structural information
can be highly beneficial to text classification but only under certain
circumstances (e.g. depending on the amount of training data and on discussion
chain complexity). Indeed, we show that contextual information on smaller
datasets from other classification tasks does not yield significant
improvements. Our framework, based on local discussion networks, allows the
integration of structural information, while minimising user profiling, thus
preserving their privacy. | cs.CL | Putting Context in Context: the Impact of Discussion Structure on Text Classification |
2024-02-05T12:47:19Z | http://arxiv.org/pdf/2402.02969v2 | http://arxiv.org/abs/2402.02969v2 | Simone Bombari, Marco Mondelli | 2024-05-17T15:31:34Z | stat.ML, cs.CL, cs.LG | Understanding the reasons behind the exceptional success of transformers
requires a better analysis of why attention layers are suitable for NLP tasks.
In particular, such tasks require predictive models to capture contextual
meaning which often depends on one or few words, even if the sentence is long.
Our work studies this key property, dubbed word sensitivity (WS), in the
prototypical setting of random features. We show that attention layers enjoy
high WS, namely, there exists a vector in the space of embeddings that largely
perturbs the random attention features map. The argument critically exploits
the role of the softmax in the attention layer, highlighting its benefit
compared to other activations (e.g., ReLU). In contrast, the WS of standard
random features is of order $1/\sqrt{n}$, $n$ being the number of words in the
textual sample, and thus it decays with the length of the context. We then
translate these results on the word sensitivity into generalization bounds: due
to their low WS, random features provably cannot learn to distinguish between
two sentences that differ only in a single word; in contrast, due to their high
WS, random attention features have higher generalization capabilities. We
validate our theoretical results with experimental evidence over the BERT-Base
word embeddings of the imdb review dataset. | stat.ML | Towards Understanding the Word Sensitivity of Attention Layers: A Study via Random Features |
2024-02-05T12:16:28Z | http://arxiv.org/pdf/2402.05964v2 | http://arxiv.org/abs/2402.05964v2 | Yehui Tang, Yunhe Wang, Jianyuan Guo, Zhijun Tu, Kai Han, Hailin Hu, Dacheng Tao | 2024-04-07T13:03:58Z | cs.LG, cs.CL, cs.CV | Transformer plays a vital role in the realms of natural language processing
(NLP) and computer vision (CV), specially for constructing large language
models (LLM) and large vision models (LVM). Model compression methods reduce
the memory and computational cost of Transformer, which is a necessary step to
implement large language/vision models on practical devices. Given the unique
architecture of Transformer, featuring alternative attention and feedforward
neural network (FFN) modules, specific compression techniques are usually
required. The efficiency of these compression methods is also paramount, as
retraining large models on the entire training dataset is usually impractical.
This survey provides a comprehensive review of recent compression methods, with
a specific focus on their application to Transformer-based models. The
compression methods are primarily categorized into pruning, quantization,
knowledge distillation, and efficient architecture design (Mamba, RetNet, RWKV,
etc.). In each category, we discuss compression methods for both language and
vision tasks, highlighting common underlying principles. Finally, we delve into
the relation between various compression methods, and discuss further
directions in this domain. | cs.LG | A Survey on Transformer Compression |
2024-02-05T11:58:56Z | http://arxiv.org/pdf/2402.05128v2 | http://arxiv.org/abs/2402.05128v2 | Hessa Abdulrahman Alawwad, Areej Alhothali, Usman Naseem, Ali Alkhathlan, Amani Jamal | 2024-02-14T10:06:54Z | cs.CL, cs.AI | Textbook question answering (TQA) is a challenging task in artificial
intelligence due to the complex nature of context and multimodal data. Although
previous research has significantly improved the task, there are still some
limitations including the models' weak reasoning and inability to capture
contextual information in the lengthy context. The introduction of large
language models (LLMs) has revolutionized the field of AI, however, directly
applying LLMs often leads to inaccurate answers. This paper proposes a
methodology that handle the out-of-domain scenario in TQA where concepts are
spread across different lessons by incorporating the retrieval augmented
generation (RAG) technique and utilize transfer learning to handle the long
context and enhance reasoning abilities. Through supervised fine-tuning of the
LLM model Llama-2 and the incorporation of RAG, our architecture outperforms
the baseline, achieving a 4.12% accuracy improvement on validation set and
9.84% on test set for non-diagram multiple-choice questions. | cs.CL | Enhancing Textbook Question Answering Task with Large Language Models and Retrieval Augmented Generation |
2024-02-05T11:47:36Z | http://arxiv.org/pdf/2402.02926v1 | http://arxiv.org/abs/2402.02926v1 | V. S. D. S. Mahesh Akavarapu, Arnab Bhattacharya | 2024-02-05T11:47:36Z | cs.CL, cs.LG, cs.SI, I.2.7 | Identification of cognates across related languages is one of the primary
problems in historical linguistics. Automated cognate identification is helpful
for several downstream tasks including identifying sound correspondences,
proto-language reconstruction, phylogenetic classification, etc. Previous
state-of-the-art methods for cognate identification are mostly based on
distributions of phonemes computed across multilingual wordlists and make
little use of the cognacy labels that define links among cognate clusters. In
this paper, we present a transformer-based architecture inspired by
computational biology for the task of automated cognate detection. Beyond a
certain amount of supervision, this method performs better than the existing
methods, and shows steady improvement with further increase in supervision,
thereby proving the efficacy of utilizing the labeled information. We also
demonstrate that accepting multiple sequence alignments as input and having an
end-to-end architecture with link prediction head saves much computation time
while simultaneously yielding superior performance. | cs.CL | Automated Cognate Detection as a Supervised Link Prediction Task with Cognate Transformer |
2024-02-05T11:32:13Z | http://arxiv.org/pdf/2402.02915v1 | http://arxiv.org/abs/2402.02915v1 | Jessica Nieder, Johann-Mattis List | 2024-02-05T11:32:13Z | cs.CL | Closely related languages show linguistic similarities that allow speakers of
one language to understand speakers of another language without having actively
learned it. Mutual intelligibility varies in degree and is typically tested in
psycholinguistic experiments. To study mutual intelligibility computationally,
we propose a computer-assisted method using the Linear Discriminative Learner,
a computational model developed to approximate the cognitive processes by which
humans learn languages, which we expand with multilingual semantic vectors and
multilingual sound classes. We test the model on cognate data from German,
Dutch, and English, three closely related Germanic languages. We find that our
model's comprehension accuracy depends on 1) the automatic trimming of
inflections and 2) the language pair for which comprehension is tested. Our
multilingual modelling approach does not only offer new methodological findings
for automatic testing of mutual intelligibility across languages but also
extends the use of Linear Discriminative Learning to multilingual settings. | cs.CL | A Computational Model for the Assessment of Mutual Intelligibility Among Closely Related Languages |
2024-02-05T11:05:20Z | http://arxiv.org/pdf/2402.02896v1 | http://arxiv.org/abs/2402.02896v1 | Ivar Frisch, Mario Giulianelli | 2024-02-05T11:05:20Z | cs.CL, cs.AI, cs.CY, cs.MA | While both agent interaction and personalisation are vibrant topics in
research on large language models (LLMs), there has been limited focus on the
effect of language interaction on the behaviour of persona-conditioned LLM
agents. Such an endeavour is important to ensure that agents remain consistent
to their assigned traits yet are able to engage in open, naturalistic
dialogues. In our experiments, we condition GPT-3.5 on personality profiles
through prompting and create a two-group population of LLM agents using a
simple variability-inducing sampling algorithm. We then administer personality
tests and submit the agents to a collaborative writing task, finding that
different profiles exhibit different degrees of personality consistency and
linguistic alignment to their conversational partners. Our study seeks to lay
the groundwork for better understanding of dialogue-based interaction between
LLMs and highlights the need for new approaches to crafting robust, more
human-like LLM personas for interactive environments. | cs.CL | LLM Agents in Interaction: Measuring Personality Consistency and Linguistic Alignment in Interacting Populations of Large Language Models |
2024-02-05T10:49:05Z | http://arxiv.org/pdf/2402.02883v1 | http://arxiv.org/abs/2402.02883v1 | Lucas Möller, Dmitry Nikolaev, Sebastian Padó | 2024-02-05T10:49:05Z | cs.CL, cs.LG | Siamese encoders such as sentence transformers are among the least understood
deep models. Established attribution methods cannot tackle this model class
since it compares two inputs rather than processing a single one. To address
this gap, we have recently proposed an attribution method specifically for
Siamese encoders (M\"oller et al., 2023). However, it requires models to be
adjusted and fine-tuned and therefore cannot be directly applied to
off-the-shelf models. In this work, we reassess these restrictions and propose
(i) a model with exact attribution ability that retains the original model's
predictive performance and (ii) a way to compute approximate attributions for
off-the-shelf models. We extensively compare approximate and exact attributions
and use them to analyze the models' attendance to different linguistic aspects.
We gain insights into which syntactic roles Siamese transformers attend to,
confirm that they mostly ignore negation, explore how they judge semantically
opposite adjectives, and find that they exhibit lexical bias. | cs.CL | Approximate Attributions for Off-the-Shelf Siamese Transformers |
2024-02-05T10:39:32Z | http://arxiv.org/pdf/2402.02872v2 | http://arxiv.org/abs/2402.02872v2 | Zeping Yu, Sophia Ananiadou | 2024-06-11T12:58:51Z | cs.CL, cs.LG | We investigate the mechanism of in-context learning (ICL) on sentence
classification tasks with semantically-unrelated labels ("foo"/"bar"). We find
intervening in only 1\% heads (named "in-context heads") significantly affects
ICL accuracy from 87.6\% to 24.4\%. To understand this phenomenon, we analyze
the value-output vectors in these heads and discover that the vectors at each
label position contain substantial information about the corresponding labels.
Furthermore, we observe that the prediction shift from "foo" to "bar" is due to
the respective reduction and increase in these heads' attention scores at "foo"
and "bar" positions. Therefore, we propose a hypothesis for ICL: in in-context
heads, the value-output matrices extract label features, while the query-key
matrices compute the similarity between the features at the last position and
those at each label position. The query and key matrices can be considered as
two towers that learn the similarity metric between the last position's
features and each demonstration at label positions. Using this hypothesis, we
explain the majority label bias and recency bias in ICL and propose two methods
to reduce these biases by 22\% and 17\%, respectively. | cs.CL | How do Large Language Models Learn In-Context? Query and Key Matrices of In-Context Heads are Two Towers for Metric Learning |
2024-02-05T10:24:40Z | http://arxiv.org/pdf/2402.02864v1 | http://arxiv.org/abs/2402.02864v1 | Axel Sorensen, Siyao Peng, Barbara Plank, Rob van der Goot | 2024-02-05T10:24:40Z | cs.CL, cs.HC | Annotation tools are the starting point for creating Natural Language
Processing (NLP) datasets. There is a wide variety of tools available; setting
up these tools is however a hindrance. We propose EEVEE, an annotation tool
focused on simplicity, efficiency, and ease of use. It can run directly in the
browser (no setup required) and uses tab-separated files (as opposed to
character offsets or task-specific formats) for annotation. It allows for
annotation of multiple tasks on a single dataset and supports four task-types:
sequence labeling, span labeling, text classification and seq2seq. | cs.CL | EEVEE: An Easy Annotation Tool for Natural Language Processing |
2024-02-05T09:57:15Z | http://arxiv.org/pdf/2402.02844v1 | http://arxiv.org/abs/2402.02844v1 | Juraj Vladika, Florian Matthes | 2024-02-05T09:57:15Z | cs.CL, cs.AI, cs.IR | The increasing rate at which scientific knowledge is discovered and health
claims shared online has highlighted the importance of developing efficient
fact-checking systems for scientific claims. The usual setting for this task in
the literature assumes that the documents containing the evidence for claims
are already provided and annotated or contained in a limited corpus. This
renders the systems unrealistic for real-world settings where knowledge sources
with potentially millions of documents need to be queried to find relevant
evidence. In this paper, we perform an array of experiments to test the
performance of open-domain claim verification systems. We test the final
verdict prediction of systems on four datasets of biomedical and health claims
in different settings. While keeping the pipeline's evidence selection and
verdict prediction parts constant, document retrieval is performed over three
common knowledge sources (PubMed, Wikipedia, Google) and using two different
information retrieval techniques. We show that PubMed works better with
specialized biomedical claims, while Wikipedia is more suited for everyday
health concerns. Likewise, BM25 excels in retrieval precision, while semantic
search in recall of relevant evidence. We discuss the results, outline frequent
retrieval patterns and challenges, and provide promising future directions. | cs.CL | Comparing Knowledge Sources for Open-Domain Scientific Claim Verification |
2024-02-05T09:48:07Z | http://arxiv.org/pdf/2402.02837v1 | http://arxiv.org/abs/2402.02837v1 | Amandine Decker, Maxime Amblard | 2024-02-05T09:48:07Z | cs.CL | Topics play an important role in the global organisation of a conversation as
what is currently discussed constrains the possible contributions of the
participant. Understanding the way topics are organised in interaction would
provide insight on the structure of dialogue beyond the sequence of utterances.
However, studying this high-level structure is a complex task that we try to
approach by first segmenting dialogues into smaller topically coherent sets of
utterances. Understanding the interactions between these segments would then
enable us to propose a model of topic organisation at a dialogue level. In this
paper we work with open-domain conversations and try to reach a comparable
level of accuracy as recent machine learning based topic segmentation models
but with a formal approach. The features we identify as meaningful for this
task help us understand better the topical structure of a conversation. | cs.CL | With a Little Help from my (Linguistic) Friends: Topic Segmentation of Multi-party Casual Conversations |
2024-02-05T09:44:49Z | http://arxiv.org/pdf/2402.02834v2 | http://arxiv.org/abs/2402.02834v2 | Bo-Kyeong Kim, Geonmin Kim, Tae-Ho Kim, Thibault Castells, Shinkook Choi, Junho Shin, Hyoung-Kyu Song | 2024-06-23T08:45:33Z | cs.LG, cs.CL | Structured pruning of modern large language models (LLMs) has emerged as a
way of decreasing their high computational needs. Width pruning reduces the
size of projection weight matrices (e.g., by removing attention heads) while
maintaining the number of layers. Depth pruning, in contrast, removes entire
layers or blocks, while keeping the size of the remaining weights unchanged.
Most current research focuses on either width-only or a blend of width and
depth pruning, with little comparative analysis between the two units (width
vs. depth) concerning their impact on LLM inference efficiency. In this work,
we show that simple depth pruning can effectively compress LLMs while achieving
comparable or superior performance to recent width pruning studies. Our pruning
method boosts inference speeds, especially under memory-constrained conditions
that require limited batch sizes for running LLMs, where width pruning is
ineffective. In retraining pruned models for quality recovery, continued
pretraining on a large corpus markedly outperforms LoRA-based tuning,
particularly at severe pruning ratios. We hope this work can help build compact
yet capable LLMs. Code and models can be found at:
https://github.com/Nota-NetsPresso/shortened-llm | cs.LG | Shortened LLaMA: Depth Pruning for Large Language Models with Comparison of Retraining Methods |
2024-02-05T09:10:32Z | http://arxiv.org/pdf/2402.02823v2 | http://arxiv.org/abs/2402.02823v2 | Jasper Dekoninck, Mark Niklas Müller, Maximilian Baader, Marc Fischer, Martin Vechev | 2024-02-12T17:50:07Z | cs.LG, cs.AI, cs.CL, cs.CR | Large language models are widespread, with their performance on benchmarks
frequently guiding user preferences for one model over another. However, the
vast amount of data these models are trained on can inadvertently lead to
contamination with public benchmarks, thus compromising performance
measurements. While recently developed contamination detection methods try to
address this issue, they overlook the possibility of deliberate contamination
by malicious model providers aiming to evade detection. We argue that this
setting is of crucial importance as it casts doubt on the reliability of public
benchmarks. To more rigorously study this issue, we propose a categorization of
both model providers and contamination detection methods. This reveals
vulnerabilities in existing methods that we exploit with EAL, a simple yet
effective contamination technique that significantly inflates benchmark
performance while completely evading current detection methods. | cs.LG | Evading Data Contamination Detection for Language Models is (too) Easy |
2024-02-05T08:35:33Z | http://arxiv.org/pdf/2402.02807v3 | http://arxiv.org/abs/2402.02807v3 | Luise Häuser, Gerhard Jäger, Taraka Rama, Johann-Mattis List, Alexandros Stamatakis | 2024-05-14T07:38:26Z | cs.CL, cs.SD, eess.AS | In traditional studies on language evolution, scholars often emphasize the
importance of sound laws and sound correspondences for phylogenetic inference
of language family trees. However, to date, computational approaches have
typically not taken this potential into account. Most computational studies
still rely on lexical cognates as major data source for phylogenetic
reconstruction in linguistics, although there do exist a few studies in which
authors praise the benefits of comparing words at the level of sound sequences.
Building on (a) ten diverse datasets from different language families, and (b)
state-of-the-art methods for automated cognate and sound correspondence
detection, we test, for the first time, the performance of sound-based versus
cognate-based approaches to phylogenetic reconstruction. Our results show that
phylogenies reconstructed from lexical cognates are topologically closer, by
approximately one third with respect to the generalized quartet distance on
average, to the gold standard phylogenies than phylogenies reconstructed from
sound correspondences. | cs.CL | Are Sounds Sound for Phylogenetic Reconstruction? |
2024-02-05T08:26:33Z | http://arxiv.org/pdf/2402.02805v2 | http://arxiv.org/abs/2402.02805v2 | Fangru Lin, Emanuele La Malfa, Valentin Hofmann, Elle Michelle Yang, Anthony Cohn, Janet B. Pierrehumbert | 2024-06-03T13:07:06Z | cs.AI, cs.CL, cs.LG | Planning is a fundamental property of human intelligence. Reasoning about
asynchronous plans is challenging since it requires sequential and parallel
planning to optimize time costs. Can large language models (LLMs) succeed at
this task? Here, we present the first large-scale study investigating this
question. We find that a representative set of closed and open-source LLMs,
including GPT-4 and LLaMA-2, behave poorly when not supplied with illustrations
about the task-solving process in our benchmark AsyncHow. We propose a novel
technique called Plan Like a Graph (PLaG) that combines graphs with natural
language prompts and achieves state-of-the-art results. We show that although
PLaG can boost model performance, LLMs still suffer from drastic degradation
when task complexity increases, highlighting the limits of utilizing LLMs for
simulating digital devices. We see our study as an exciting step towards using
LLMs as efficient autonomous agents. Our code and data are available at
https://github.com/fangru-lin/graph-llm-asynchow-plan. | cs.AI | Graph-enhanced Large Language Models in Asynchronous Plan Reasoning |
2024-02-05T08:19:56Z | http://arxiv.org/pdf/2402.02801v2 | http://arxiv.org/abs/2402.02801v2 | Fei Yuan, Chang Ma, Shuai Yuan, Qiushi Sun, Lei Li | 2024-06-03T07:35:25Z | cs.CL, cs.AI, cs.LG | The lottery ticket hypothesis posits the existence of ``winning tickets''
within a randomly initialized neural network. Do winning tickets exist for LLMs
in fine-tuning scenarios? How can we find such winning tickets? In this paper,
we propose KS-Lottery, a method to identify a small subset of LLM parameters
highly effective in multilingual fine-tuning. Our key idea is to use
Kolmogorov-Smirnov Test to analyze the distribution shift of parameters before
and after fine-tuning. We further theoretically prove that KS-Lottery can find
the certified winning tickets in the embedding layer, fine-tuning on the found
parameters is guaranteed to perform as well as full fine-tuning. Comparing
KS-Lottery with other parameter-efficient tuning algorithms on translation
tasks, the experimental results show that KS-Lottery finds a much smaller set
of parameters for fine-tuning while achieving the comparable performance as
full fine-tuning LLM. Surprisingly, we find that fine-tuning 18 tokens'
embedding of LLaMA suffices to reach the fine-tuning translation
performance~\footnote{https://github.com/CONE-MT/KS-Lottery.}. | cs.CL | KS-Lottery: Finding Certified Lottery Tickets for Multilingual Language Models |
2024-02-05T07:59:38Z | http://arxiv.org/pdf/2402.02791v2 | http://arxiv.org/abs/2402.02791v2 | Yehui Tang, Fangcheng Liu, Yunsheng Ni, Yuchuan Tian, Zheyuan Bai, Yi-Qi Hu, Sichao Liu, Shangling Jui, Kai Han, Yunhe Wang | 2024-02-06T05:38:26Z | cs.CL, cs.AI, cs.LG | The power of large language models (LLMs) has been demonstrated through
numerous data and computing resources. However, the application of language
models on mobile devices is facing huge challenge on the computation and memory
costs, that is, tiny language models with high performance are urgently
required. Limited by the highly complex training process, there are many
details for optimizing language models that are seldom studied carefully. In
this study, based on a tiny language model with 1B parameters, we carefully
design a series of empirical study to analyze the effect of each component.
Three perspectives are mainly discussed, \ie, neural architecture, parameter
initialization, and optimization strategy. Several design formulas are
empirically proved especially effective for tiny language models, including
tokenizer compression, architecture tweaking, parameter inheritance and
multiple-round training. Then we train PanGu-$\pi$-1B Pro and PanGu-$\pi$-1.5B
Pro on 1.6T multilingual corpora, following the established formulas.
Experimental results demonstrate the improved optimization and architecture
yield a notable average improvement of 8.87 on benchmark evaluation sets for
PanGu-$\pi$-1B Pro. Besides, PanGu-$\pi$-1.5B Pro surpasses a range of SOTA
models with larger model sizes, validating its superior performance. The code
is available at https://github.com/YuchuanTian/RethinkTinyLM. | cs.CL | Rethinking Optimization and Architecture for Tiny Language Models |
2024-02-05T07:33:25Z | http://arxiv.org/pdf/2402.02782v1 | http://arxiv.org/abs/2402.02782v1 | Ana Ezquerro, Carlos Gómez-Rodríguez, David Vilares | 2024-02-05T07:33:25Z | cs.CL | We study incremental constituent parsers to assess their capacity to output
trees based on prefix representations alone. Guided by strictly left-to-right
generative language models and tree-decoding modules, we build parsers that
adhere to a strong definition of incrementality across languages. This builds
upon work that asserted incrementality, but that mostly only enforced it on
either the encoder or the decoder. Finally, we conduct an analysis against
non-incremental and partially incremental models. | cs.CL | From Partial to Strictly Incremental Constituent Parsing |
2024-02-05T07:30:32Z | http://arxiv.org/pdf/2402.02781v1 | http://arxiv.org/abs/2402.02781v1 | Yang Xiao, Rohan Kumar Das | 2024-02-05T07:30:32Z | cs.SD, cs.AI, cs.CL, cs.LG, eess.AS | Sound event detection (SED) is essential for recognizing specific sounds and
their temporal locations within acoustic signals. This becomes challenging
particularly for on-device applications, where computational resources are
limited. To address this issue, we introduce a novel framework referred to as
dual knowledge distillation for developing efficient SED systems in this work.
Our proposed dual knowledge distillation commences with temporal-averaging
knowledge distillation (TAKD), utilizing a mean student model derived from the
temporal averaging of the student model's parameters. This allows the student
model to indirectly learn from a pre-trained teacher model, ensuring a stable
knowledge distillation. Subsequently, we introduce embedding-enhanced feature
distillation (EEFD), which involves incorporating an embedding distillation
layer within the student model to bolster contextual learning. On DCASE 2023
Task 4A public evaluation dataset, our proposed SED system with dual knowledge
distillation having merely one-third of the baseline model's parameters,
demonstrates superior performance in terms of PSDS1 and PSDS2. This highlights
the importance of proposed dual knowledge distillation for compact SED systems,
which can be ideal for edge devices. | cs.SD | Dual Knowledge Distillation for Efficient Sound Event Detection |
2024-02-05T06:52:53Z | http://arxiv.org/pdf/2402.02764v1 | http://arxiv.org/abs/2402.02764v1 | Shicheng Xu, Liang Pang, Jun Xu, Huawei Shen, Xueqi Cheng | 2024-02-05T06:52:53Z | cs.IR, cs.AI, cs.CL | The results of information retrieval (IR) are usually presented in the form
of a ranked list of candidate documents, such as web search for humans and
retrieval-augmented generation for large language models (LLMs). List-aware
retrieval aims to capture the list-level contextual features to return a better
list, mainly including reranking and truncation. Reranking finely re-scores the
documents in the list. Truncation dynamically determines the cut-off point of
the ranked list to achieve the trade-off between overall relevance and avoiding
misinformation from irrelevant documents. Previous studies treat them as two
separate tasks and model them separately. However, the separation is not
optimal. First, it is hard to share the contextual information of the ranking
list between the two tasks. Second, the separate pipeline usually meets the
error accumulation problem, where the small error from the reranking stage can
largely affect the truncation stage. To solve these problems, we propose a
Reranking-Truncation joint model (GenRT) that can perform the two tasks
concurrently. GenRT integrates reranking and truncation via generative paradigm
based on encoder-decoder architecture. We also design the novel loss functions
for joint optimization to make the model learn both tasks. Sharing parameters
by the joint model is conducive to making full use of the common modeling
information of the two tasks. Besides, the two tasks are performed concurrently
and co-optimized to solve the error accumulation problem between separate
stages. Experiments on public learning-to-rank benchmarks and open-domain Q\&A
tasks show that our method achieves SOTA performance on both reranking and
truncation tasks for web search and retrieval-augmented LLMs. | cs.IR | List-aware Reranking-Truncation Joint Model for Search and Retrieval-augmented Generation |
2024-02-05T06:12:29Z | http://arxiv.org/pdf/2402.06147v2 | http://arxiv.org/abs/2402.06147v2 | James Y. Huang, Sailik Sengupta, Daniele Bonadiman, Yi-an Lai, Arshit Gupta, Nikolaos Pappas, Saab Mansour, Katrin Kirchhoff, Dan Roth | 2024-02-21T02:25:32Z | cs.AI, cs.CL | Large Language Models (LLMs) are nowadays expected to generate content
aligned with human preferences. Current work focuses on alignment at model
training time, through techniques such as Reinforcement Learning with Human
Feedback (RLHF). However, it is unclear if such methods are an effective choice
to teach alignment objectives to the model. First, the inability to incorporate
multiple, custom rewards and reliance on a model developer's view of universal
and static principles are key limitations. Second, the residual gaps in model
training and the reliability of such approaches are also questionable (e.g.
susceptibility to jail-breaking even after safety training). To address these,
we propose DeAL, a framework that allows the user to customize reward functions
and enables Decoding-time Alignment of LLMs (DeAL). At its core, we view
decoding as a heuristic-guided search process and facilitate the use of a wide
variety of alignment objectives. Our experiments with programmatic constraints
such as keyword and length constraints (studied widely in the pre-LLM era) and
abstract objectives such as harmlessness and helpfulness (proposed in the
post-LLM era) show that we can DeAL with fine-grained trade-offs, improve
adherence to alignment objectives, and address residual gaps in LLMs. Lastly,
while DeAL can be effectively paired with RLHF and prompting techniques, its
generality makes decoding slower, an optimization we leave for future work. | cs.AI | DeAL: Decoding-time Alignment for Large Language Models |
2024-02-05T06:08:06Z | http://arxiv.org/pdf/2402.05127v1 | http://arxiv.org/abs/2402.05127v1 | Aryan Agrawal | 2024-02-05T06:08:06Z | cs.CL, cs.AI, cs.LG | This paper introduces a novel paradigm for depression detection and treatment
using advanced Large Language Models (LLMs): Generative Pre-trained Transformer
4 (GPT-4), Llama 2 chat, and Gemini. These LLMs are fine-tuned with specialized
prompts to diagnose, explain, and suggest therapeutic interventions for
depression. A unique few-shot prompting method enhances the models' ability to
analyze and explain depressive symptoms based on the DSM-5 criteria. In the
interaction phase, the models engage in empathetic dialogue management, drawing
from resources like PsychDB and a Cognitive Behavioral Therapy (CBT) Guide,
fostering supportive interactions with individuals experiencing major
depressive disorders. Additionally, the research introduces the Illuminate
Database, enriched with various CBT modules, aiding in personalized therapy
recommendations. The study evaluates LLM performance using metrics such as F1
scores, Precision, Recall, Cosine similarity, and Recall-Oriented Understudy
for Gisting Evaluation (ROUGE) across different test sets, demonstrating their
effectiveness. This comprehensive approach blends cutting-edge AI with
established psychological methods, offering new possibilities in mental health
care and showcasing the potential of LLMs in revolutionizing depression
diagnosis and treatment strategies. | cs.CL | Illuminate: A novel approach for depression detection with explainable analysis and proactive therapy using prompt engineering |
2024-02-05T06:06:47Z | http://arxiv.org/pdf/2402.02750v1 | http://arxiv.org/abs/2402.02750v1 | Zirui Liu, Jiayi Yuan, Hongye Jin, Shaochen Zhong, Zhaozhuo Xu, Vladimir Braverman, Beidi Chen, Xia Hu | 2024-02-05T06:06:47Z | cs.CL, cs.LG, cs.PF | Efficiently serving large language models (LLMs) requires batching many
requests together to reduce the cost per request. Yet, the key-value (KV)
cache, which stores attention keys and values to avoid re-computations,
significantly increases memory demands and becomes the new bottleneck in speed
and memory usage. This memory demand increases with larger batch sizes and
longer context lengths. Additionally, the inference speed is limited by the
size of KV cache, as the GPU's SRAM must load the entire KV cache from the main
GPU memory for each token generated, causing the computational core to be idle
during this process. A straightforward and effective solution to reduce KV
cache size is quantization, which decreases the total bytes taken by KV cache.
However, there is a lack of in-depth studies that explore the element
distribution of KV cache to understand the hardness and limitation of KV cache
quantization. To fill the gap, we conducted a comprehensive study on the
element distribution in KV cache of popular LLMs. Our findings indicate that
the key cache should be quantized per-channel, i.e., group elements along the
channel dimension and quantize them together. In contrast, the value cache
should be quantized per-token. From this analysis, we developed a tuning-free
2bit KV cache quantization algorithm, named KIVI. With the hardware-friendly
implementation, KIVI can enable Llama (Llama-2), Falcon, and Mistral models to
maintain almost the same quality while using $\mathbf{2.6\times}$ less peak
memory usage (including the model weight). This reduction in memory usage
enables up to $\mathbf{4\times}$ larger batch size, bringing
$\mathbf{2.35\times \sim 3.47\times}$ throughput on real LLM inference
workload. The source code is available at https://github.com/jy-yuan/KIVI. | cs.CL | KIVI: A Tuning-Free Asymmetric 2bit Quantization for KV Cache |
2024-02-05T04:25:24Z | http://arxiv.org/pdf/2402.02716v1 | http://arxiv.org/abs/2402.02716v1 | Xu Huang, Weiwen Liu, Xiaolong Chen, Xingmei Wang, Hao Wang, Defu Lian, Yasheng Wang, Ruiming Tang, Enhong Chen | 2024-02-05T04:25:24Z | cs.AI, cs.CL, cs.LG | As Large Language Models (LLMs) have shown significant intelligence, the
progress to leverage LLMs as planning modules of autonomous agents has
attracted more attention. This survey provides the first systematic view of
LLM-based agents planning, covering recent works aiming to improve planning
ability. We provide a taxonomy of existing works on LLM-Agent planning, which
can be categorized into Task Decomposition, Plan Selection, External Module,
Reflection and Memory. Comprehensive analyses are conducted for each direction,
and further challenges for the field of research are discussed. | cs.AI | Understanding the planning of LLM agents: A survey |
2024-02-05T03:15:26Z | http://arxiv.org/pdf/2402.02695v2 | http://arxiv.org/abs/2402.02695v2 | Raha Moraffah, Huan Liu | 2024-02-21T00:09:00Z | cs.CL, cs.AI, cs.CR, cs.LG | Sentence-level attacks craft adversarial sentences that are synonymous with
correctly-classified sentences but are misclassified by the text classifiers.
Under the black-box setting, classifiers are only accessible through their
feedback to queried inputs, which is predominately available in the form of
class probabilities. Even though utilizing class probabilities results in
stronger attacks, due to the challenges of using them for sentence-level
attacks, existing attacks use either no feedback or only the class labels.
Overcoming the challenges, we develop a novel algorithm that uses class
probabilities for black-box sentence-level attacks, investigate the
effectiveness of using class probabilities on the attack's success, and examine
the question if it is worthy or practical to use class probabilities by
black-box sentence-level attacks. We conduct extensive evaluations of our
attack comparing with the baselines across various classifiers and benchmark
datasets. | cs.CL | Exploiting Class Probabilities for Black-box Sentence-level Attacks |
2024-02-05T03:00:44Z | http://arxiv.org/pdf/2402.05126v1 | http://arxiv.org/abs/2402.05126v1 | Imaad Zaffar Khan, Amaan Aijaz Sheikh, Utkarsh Sinha | 2024-02-05T03:00:44Z | cs.CL, cs.LG | With the abundance of data and information in todays time, it is nearly
impossible for man, or, even machine, to go through all of the data line by
line. What one usually does is to try to skim through the lines and retain the
absolutely important information, that in a more formal term is called
summarization. Text summarization is an important task that aims to compress
lengthy documents or articles into shorter, coherent representations while
preserving the core information and meaning. This project introduces an
innovative approach to text summarization, leveraging the capabilities of Graph
Neural Networks (GNNs) and Named Entity Recognition (NER) systems. GNNs, with
their exceptional ability to capture and process the relational data inherent
in textual information, are adept at understanding the complex structures
within large documents. Meanwhile, NER systems contribute by identifying and
emphasizing key entities, ensuring that the summarization process maintains a
focus on the most critical aspects of the text. By integrating these two
technologies, our method aims to enhances the efficiency of summarization and
also tries to ensures a high degree relevance in the condensed content. This
project, therefore, offers a promising direction for handling the ever
increasing volume of textual data in an information-saturated world. | cs.CL | Graph Neural Network and NER-Based Text Summarization |
2024-02-05T02:36:41Z | http://arxiv.org/pdf/2402.06655v1 | http://arxiv.org/abs/2402.06655v1 | Raha Moraffah, Shubh Khandelwal, Amrita Bhattacharjee, Huan Liu | 2024-02-05T02:36:41Z | cs.CR, cs.AI, cs.CL, cs.LG | Adversarial purification is a defense mechanism for safeguarding classifiers
against adversarial attacks without knowing the type of attacks or training of
the classifier. These techniques characterize and eliminate adversarial
perturbations from the attacked inputs, aiming to restore purified samples that
retain similarity to the initially attacked ones and are correctly classified
by the classifier. Due to the inherent challenges associated with
characterizing noise perturbations for discrete inputs, adversarial text
purification has been relatively unexplored. In this paper, we investigate the
effectiveness of adversarial purification methods in defending text
classifiers. We propose a novel adversarial text purification that harnesses
the generative capabilities of Large Language Models (LLMs) to purify
adversarial text without the need to explicitly characterize the discrete noise
perturbations. We utilize prompt engineering to exploit LLMs for recovering the
purified examples for given adversarial examples such that they are
semantically similar and correctly classified. Our proposed method demonstrates
remarkable performance over various classifiers, improving their accuracy under
the attack by over 65% on average. | cs.CR | Adversarial Text Purification: A Large Language Model Approach for Defense |
2024-02-05T02:32:09Z | http://arxiv.org/pdf/2402.02680v1 | http://arxiv.org/abs/2402.02680v1 | Rohin Manvi, Samar Khanna, Marshall Burke, David Lobell, Stefano Ermon | 2024-02-05T02:32:09Z | cs.CL, cs.AI, cs.CY, cs.LG | Large Language Models (LLMs) inherently carry the biases contained in their
training corpora, which can lead to the perpetuation of societal harm. As the
impact of these foundation models grows, understanding and evaluating their
biases becomes crucial to achieving fairness and accuracy. We propose to study
what LLMs know about the world we live in through the lens of geography. This
approach is particularly powerful as there is ground truth for the numerous
aspects of human life that are meaningfully projected onto geographic space
such as culture, race, language, politics, and religion. We show various
problematic geographic biases, which we define as systemic errors in geospatial
predictions. Initially, we demonstrate that LLMs are capable of making accurate
zero-shot geospatial predictions in the form of ratings that show strong
monotonic correlation with ground truth (Spearman's $\rho$ of up to 0.89). We
then show that LLMs exhibit common biases across a range of objective and
subjective topics. In particular, LLMs are clearly biased against locations
with lower socioeconomic conditions (e.g. most of Africa) on a variety of
sensitive subjective topics such as attractiveness, morality, and intelligence
(Spearman's $\rho$ of up to 0.70). Finally, we introduce a bias score to
quantify this and find that there is significant variation in the magnitude of
bias across existing LLMs. | cs.CL | Large Language Models are Geographically Biased |
2024-02-05T01:14:07Z | http://arxiv.org/pdf/2402.02662v1 | http://arxiv.org/abs/2402.02662v1 | Eric Yang Yu, Christopher Liao, Sathvik Ravi, Theodoros Tsiligkaridis, Brian Kulis | 2024-02-05T01:14:07Z | cs.CV, cs.CL, cs.LG | Recent advances in vision-language models have combined contrastive
approaches with generative methods to achieve state-of-the-art (SOTA) on
downstream inference tasks like zero-shot image classification. However, a
persistent issue of these models for image classification is their
out-of-distribution (OOD) generalization capabilities. We first show that when
an OOD data point is misclassified, the correct class can be typically found in
the Top-K predicted classes. In order to steer the model prediction toward the
correct class within the top predicted classes, we propose the Image-Caption
Encoding (ICE) method, a straightforward approach that directly enforces
consistency between the image-conditioned and caption-conditioned predictions
at evaluation time only. Intuitively, we take advantage of unique properties of
the generated captions to guide our local search for the correct class label
within the Top-K predicted classes. We show that our method can be easily
combined with other SOTA methods to enhance Top-1 OOD accuracies by 0.5% on
average and up to 3% on challenging datasets. Our code:
https://github.com/Chris210634/ice | cs.CV | Image-Caption Encoding for Improving Zero-Shot Generalization |
2024-02-05T00:57:51Z | http://arxiv.org/pdf/2402.02658v1 | http://arxiv.org/abs/2402.02658v1 | Zihan Wang, Yunxuan Li, Yuexin Wu, Liangchen Luo, Le Hou, Hongkun Yu, Jingbo Shang | 2024-02-05T00:57:51Z | cs.AI, cs.CL, cs.LG | Process supervision, using a trained verifier to evaluate the intermediate
steps generated by reasoner, has demonstrated significant improvements in
multi-step problem solving. In this paper, to avoid expensive human annotation
effort on the verifier training data, we introduce Model-induced Process
Supervision (MiPS), a novel method for automating data curation. MiPS annotates
an intermediate step by sampling completions of this solution through the
reasoning model, and obtaining an accuracy defined as the proportion of correct
completions. Errors in the reasoner would cause MiPS to underestimate the
accuracy of intermediate steps, therefore, we suggest and empirically show that
verification focusing on high predicted scores of the verifier shall be
preferred over that of low predicted scores, contrary to prior work. Our
approach significantly improves the performance of PaLM 2 on math and coding
tasks (accuracy +0.67% on GSM8K, +4.16% on MATH, +0.92% on MBPP compared with
an output supervision trained verifier). Additionally, our study demonstrates
that the verifier exhibits strong generalization ability across different
reasoning models. | cs.AI | Multi-step Problem Solving Through a Verifier: An Empirical Analysis on Model-induced Process Supervision |
2024-02-05T00:56:30Z | http://arxiv.org/pdf/2402.02656v1 | http://arxiv.org/abs/2402.02656v1 | Satpreet Harcharan Singh, Kevin Jiang, Kanchan Bhasin, Ashutosh Sabharwal, Nidal Moukaddam, Ankit B Patel | 2024-02-05T00:56:30Z | cs.CL, q-bio.QM | Semi-structured interviews (SSIs) are a commonly employed data-collection
method in healthcare research, offering in-depth qualitative insights into
subject experiences. Despite their value, the manual analysis of SSIs is
notoriously time-consuming and labor-intensive, in part due to the difficulty
of extracting and categorizing emotional responses, and challenges in scaling
human evaluation for large populations. In this study, we develop RACER, a
Large Language Model (LLM) based expert-guided automated pipeline that
efficiently converts raw interview transcripts into insightful domain-relevant
themes and sub-themes. We used RACER to analyze SSIs conducted with 93
healthcare professionals and trainees to assess the broad personal and
professional mental health impacts of the COVID-19 crisis. RACER achieves
moderately high agreement with two human evaluators (72%), which approaches the
human inter-rater agreement (77%). Interestingly, LLMs and humans struggle with
similar content involving nuanced emotional, ambivalent/dialectical, and
psychological statements. Our study highlights the opportunities and challenges
in using LLMs to improve research efficiency and opens new avenues for scalable
analysis of SSIs in healthcare research. | cs.CL | RACER: An LLM-powered Methodology for Scalable Analysis of Semi-structured Mental Health Interviews |
2024-02-05T00:54:40Z | http://arxiv.org/pdf/2402.02655v2 | http://arxiv.org/abs/2402.02655v2 | Thinh Phuoc Ngo, Khoa Tran Anh Dang, Son T. Luu, Kiet Van Nguyen, Ngan Luu-Thuy Nguyen | 2024-04-06T04:29:58Z | cs.CL | This paper presents the development process of a Vietnamese spoken language
corpus for machine reading comprehension (MRC) tasks and provides insights into
the challenges and opportunities associated with using real-world data for
machine reading comprehension tasks. The existing MRC corpora in Vietnamese
mainly focus on formal written documents such as Wikipedia articles, online
newspapers, or textbooks. In contrast, the VlogQA consists of 10,076
question-answer pairs based on 1,230 transcript documents sourced from YouTube
-- an extensive source of user-uploaded content, covering the topics of food
and travel. By capturing the spoken language of native Vietnamese speakers in
natural settings, an obscure corner overlooked in Vietnamese research, the
corpus provides a valuable resource for future research in reading
comprehension tasks for the Vietnamese language. Regarding performance
evaluation, our deep-learning models achieved the highest F1 score of 75.34% on
the test set, indicating significant progress in machine reading comprehension
for Vietnamese spoken language data. In terms of EM, the highest score we
accomplished is 53.97%, which reflects the challenge in processing spoken-based
content and highlights the need for further improvement. | cs.CL | VlogQA: Task, Dataset, and Baseline Models for Vietnamese Spoken-Based Machine Reading Comprehension |
2024-02-05T00:44:28Z | http://arxiv.org/pdf/2402.02648v2 | http://arxiv.org/abs/2402.02648v2 | Jinwoo Ahn, Kyuseung Shin | 2024-03-01T10:46:01Z | cs.CL, cs.AI | Large Language Models (LLMs) frequently struggle with complex reasoning
tasks, failing to construct logically sound steps towards the solution. In
response to this behavior, users often try prompting the LLMs repeatedly in
hopes of reaching a better response. This paper studies such repetitive
behavior and its effect by defining a novel setting, Chain-of-Feedback (CoF).
The setting takes questions that require multi-step reasoning as an input. Upon
response, we repetitively prompt meaningless feedback (e.g. 'make another
attempt') requesting additional trials. Surprisingly, our preliminary results
show that repeated meaningless feedback gradually decreases the quality of the
responses, eventually leading to a larger deviation from the intended outcome.
To alleviate these troubles, we propose a novel method, Recursive
Chain-of-Feedback (R-CoF). Following the logic of recursion in computer
science, R-CoF recursively revises the initially incorrect response by breaking
down each incorrect reasoning step into smaller individual problems. Our
preliminary results show that majority of questions that LLMs fail to respond
correctly can be answered using R-CoF without any sample data outlining the
logical process. | cs.CL | Recursive Chain-of-Feedback Prevents Performance Degradation from Redundant Prompting |
2024-02-05T00:06:08Z | http://arxiv.org/pdf/2402.05125v3 | http://arxiv.org/abs/2402.05125v3 | Michael Wornow, Alejandro Lozano, Dev Dash, Jenelle Jindal, Kenneth W. Mahaffey, Nigam H. Shah | 2024-04-10T05:37:26Z | cs.CL, cs.AI | Matching patients to clinical trials is a key unsolved challenge in bringing
new drugs to market. Today, identifying patients who meet a trial's eligibility
criteria is highly manual, taking up to 1 hour per patient. Automated screening
is challenging, however, as it requires understanding unstructured clinical
text. Large language models (LLMs) offer a promising solution. In this work, we
explore their application to trial matching. First, we design an LLM-based
system which, given a patient's medical history as unstructured clinical text,
evaluates whether that patient meets a set of inclusion criteria (also
specified as free text). Our zero-shot system achieves state-of-the-art scores
on the n2c2 2018 cohort selection benchmark. Second, we improve the data and
cost efficiency of our method by identifying a prompting strategy which matches
patients an order of magnitude faster and more cheaply than the status quo, and
develop a two-stage retrieval pipeline that reduces the number of tokens
processed by up to a third while retaining high performance. Third, we evaluate
the interpretability of our system by having clinicians evaluate the natural
language justifications generated by the LLM for each eligibility decision, and
show that it can output coherent explanations for 97% of its correct decisions
and 75% of its incorrect ones. Our results establish the feasibility of using
LLMs to accelerate clinical trial operations. | cs.CL | Zero-Shot Clinical Trial Patient Matching with LLMs |
2024-02-04T23:42:02Z | http://arxiv.org/pdf/2402.02643v1 | http://arxiv.org/abs/2402.02643v1 | Xuanhe Zhou, Xinyang Zhao, Guoliang Li | 2024-02-04T23:42:02Z | cs.DB, cs.AI, cs.CL, cs.LG | Machine learning (ML) techniques for optimizing data management problems have
been extensively studied and widely deployed in recent five years. However
traditional ML methods have limitations on generalizability (adapting to
different scenarios) and inference ability (understanding the context).
Fortunately, large language models (LLMs) have shown high generalizability and
human-competitive abilities in understanding context, which are promising for
data management tasks (e.g., database diagnosis, database tuning). However,
existing LLMs have several limitations: hallucination, high cost, and low
accuracy for complicated tasks. To address these challenges, we design LLMDB,
an LLM-enhanced data management paradigm which has generalizability and high
inference ability while avoiding hallucination, reducing LLM cost, and
achieving high accuracy. LLMDB embeds domain-specific knowledge to avoid
hallucination by LLM fine-tuning and prompt engineering. LLMDB reduces the high
cost of LLMs by vector databases which provide semantic search and caching
abilities. LLMDB improves the task accuracy by LLM agent which provides
multiple-round inference and pipeline executions. We showcase three real-world
scenarios that LLMDB can well support, including query rewrite, database
diagnosis and data analytics. We also summarize the open research challenges of
LLMDB. | cs.DB | LLM-Enhanced Data Management |
2024-02-04T23:23:51Z | http://arxiv.org/pdf/2402.02639v2 | http://arxiv.org/abs/2402.02639v2 | Ned Cooper, Courtney Heldreth, Ben Hutchinson | 2024-02-06T02:50:48Z | cs.CL | Indigenous languages are historically under-served by Natural Language
Processing (NLP) technologies, but this is changing for some languages with the
recent scaling of large multilingual models and an increased focus by the NLP
community on endangered languages. This position paper explores ethical
considerations in building NLP technologies for Indigenous languages, based on
the premise that such projects should primarily serve Indigenous communities.
We report on interviews with 17 researchers working in or with Aboriginal
and/or Torres Strait Islander communities on language technology projects in
Australia. Drawing on insights from the interviews, we recommend practices for
NLP researchers to increase attention to the process of engagements with
Indigenous communities, rather than focusing only on decontextualised
artefacts. | cs.CL | "It's how you do things that matters": Attending to Process to Better Serve Indigenous Communities with Language Technologies |
2024-02-04T23:04:02Z | http://arxiv.org/pdf/2402.02636v1 | http://arxiv.org/abs/2402.02636v1 | Gaël Gendron, Bao Trung Nguyen, Alex Yuxuan Peng, Michael Witbrock, Gillian Dobbie | 2024-02-04T23:04:02Z | cs.CL, cs.AI, cs.IT, cs.LG, math.IT, I.2.3; I.2.6; I.2.7; G.3 | Despite impressive performance on language modelling and complex reasoning
tasks, Large Language Models (LLMs) fall short on the same tasks in uncommon
settings or with distribution shifts, exhibiting some lack of generalisation
ability. This issue has usually been alleviated by feeding more training data
into the LLM. However, this method is brittle, as the scope of tasks may not be
readily predictable or may evolve, and updating the model with new data
generally requires extensive additional training. By contrast, systems, such as
causal models, that learn abstract variables and causal relationships can
demonstrate increased robustness against changes in the distribution. One
reason for this success is the existence and use of Independent Causal
Mechanisms (ICMs) representing high-level concepts that only sparsely interact.
In this work, we apply two concepts from causality to learn ICMs within LLMs.
We develop a new LLM architecture composed of multiple sparsely interacting
language modelling modules. We introduce a routing scheme to induce
specialisation of the network into domain-specific modules. We also present a
Mutual Information minimisation objective that trains a separate module to
learn abstraction and domain-invariant mechanisms. We show that such causal
constraints can improve out-of-distribution performance on abstract and causal
reasoning tasks. | cs.CL | Can Large Language Models Learn Independent Causal Mechanisms? |
2024-02-04T22:56:56Z | http://arxiv.org/pdf/2402.02633v1 | http://arxiv.org/abs/2402.02633v1 | Eric Khiu, Hasti Toossi, David Anugraha, Jinyu Liu, Jiaxu Li, Juan Armando Parra Flores, Leandro Acros Roman, A. Seza Doğruöz, En-Shiun Annie Lee | 2024-02-04T22:56:56Z | cs.CL, cs.LG | Fine-tuning and testing a multilingual large language model is expensive and
challenging for low-resource languages (LRLs). While previous studies have
predicted the performance of natural language processing (NLP) tasks using
machine learning methods, they primarily focus on high-resource languages,
overlooking LRLs and shifts across domains. Focusing on LRLs, we investigate
three factors: the size of the fine-tuning corpus, the domain similarity
between fine-tuning and testing corpora, and the language similarity between
source and target languages. We employ classical regression models to assess
how these factors impact the model's performance. Our results indicate that
domain similarity has the most critical impact on predicting the performance of
Machine Translation models. | cs.CL | Predicting Machine Translation Performance on Low-Resource Languages: The Role of Domain Similarity |
2024-02-04T22:53:38Z | http://arxiv.org/pdf/2402.02632v2 | http://arxiv.org/abs/2402.02632v2 | Nafiseh Nikeghbal, Amir Hossein Kargaran, Abbas Heydarnoori | 2024-02-08T14:24:40Z | cs.SE, cs.CL | Platforms such as GitHub and GitLab introduce Issue Report Templates (IRTs)
to enable more effective issue management and better alignment with developer
expectations. However, these templates are not widely adopted in most
repositories, and there is currently no tool available to aid developers in
generating them. In this work, we introduce GIRT-Model, an assistant language
model that automatically generates IRTs based on the developer's instructions
regarding the structure and necessary fields. We create GIRT-Instruct, a
dataset comprising pairs of instructions and IRTs, with the IRTs sourced from
GitHub repositories. We use GIRT-Instruct to instruction-tune a T5-base model
to create the GIRT-Model. In our experiments, GIRT-Model outperforms general
language models (T5 and Flan-T5 with different parameter sizes) in IRT
generation by achieving significantly higher scores in ROUGE, BLEU, METEOR, and
human evaluation. Additionally, we analyze the effectiveness of GIRT-Model in a
user study in which participants wrote short IRTs with GIRT-Model. Our results
show that the participants find GIRT-Model useful in the automated generation
of templates. We hope that through the use of GIRT-Model, we can encourage more
developers to adopt IRTs in their repositories. We publicly release our code,
dataset, and model at https://github.com/ISE-Research/girt-model. | cs.SE | GIRT-Model: Automated Generation of Issue Report Templates |
2024-02-04T22:48:05Z | http://arxiv.org/pdf/2402.03396v1 | http://arxiv.org/abs/2402.03396v1 | Yifeng He, Jiabo Huang, Yuyang Rong, Yiwen Guo, Ethan Wang, Hao Chen | 2024-02-04T22:48:05Z | cs.SE, cs.AI, cs.CL, cs.CR, cs.LG | The remarkable capability of large language models (LLMs) in generating
high-quality code has drawn increasing attention in the software testing
community. However, existing code LLMs often demonstrate unsatisfactory
capabilities in generating accurate and complete tests since they were trained
on code snippets collected without differentiating between code for testing
purposes and other code. In this paper, we present a large-scale dataset
UniTSyn, which is capable of enhancing the prowess of LLMs for Unit Test
Synthesis. Associating tests with the tested functions is crucial for LLMs to
infer the expected behavior and the logic paths to be verified. By leveraging
Language Server Protocol, UniTSyn achieves the challenging goal of collecting
focal-test pairs without per-project execution setups or per-language
heuristics that tend to be fragile and difficult to scale. It contains 2.7
million focal-test pairs across five mainstream programming languages, making
it possible to be utilized for enhancing the test generation ability of LLMs.
The details of UniTSyn can be found in Table 1. Our experiments demonstrate
that, by building an autoregressive model based on UniTSyn, we can achieve
significant benefits in learning and understanding unit test representations,
resulting in improved generation accuracy and code coverage across all
evaluated programming languages. Code and data will be publicly available. | cs.SE | UniTSyn: A Large-Scale Dataset Capable of Enhancing the Prowess of Large Language Models for Program Testing |
2024-02-04T22:12:29Z | http://arxiv.org/pdf/2402.02625v1 | http://arxiv.org/abs/2402.02625v1 | Razvan-Gabriel Dumitru, Darius Peteleaza, Mihai Surdeanu | 2024-02-04T22:12:29Z | cs.LG, cs.AI, cs.CL, I.2.0; I.2.7 | We introduce the concept of multiple temporal perspectives, a novel approach
applicable to Recurrent Neural Network (RNN) architectures for enhancing their
understanding of sequential data. This method involves maintaining diverse
temporal views of previously encountered text, significantly enriching the
language models' capacity to interpret context. To show the efficacy of this
approach, we incorporate it into the Receptance Weighted Key Value (RWKV)
architecture, addressing its inherent challenge of retaining all historical
information within a single hidden state. Notably, this improvement is achieved
with a minimal increase in the number of parameters --even as little as
$0.04\%$ of the original number of parameters. Further, the additional
parameters necessary for the multiple temporal perspectives are fine-tuned with
minimal computational overhead, avoiding the need for a full pre-training. The
resulting model maintains linear computational complexity during prompt
inference, ensuring consistent efficiency across various sequence lengths. The
empirical results and ablation studies included in our research validate the
effectiveness of our approach, showcasing improved performance across multiple
benchmarks. The code, model weights and datasets are open-sourced at:
https://github.com/RazvanDu/TemporalRNNs. | cs.LG | Enhancing Transformer RNNs with Multiple Temporal Perspectives |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.