publication_date
stringlengths 20
20
| doc_url
stringlengths 33
34
| id
stringlengths 33
34
| authors
stringlengths 5
20.1k
| update_date
stringlengths 20
20
| category_all
stringlengths 5
138
| abstract
stringlengths 123
3.09k
| category_primary
stringclasses 78
values | title
stringlengths 10
211
|
|---|---|---|---|---|---|---|---|---|
2024-05-28T17:40:48Z | http://arxiv.org/pdf/2405.18400v3 | http://arxiv.org/abs/2405.18400v3 | Ethan Shen, Alan Fan, Sarah M. Pratt, Jae Sung Park, Matthew Wallingford, Sham M. Kakade, Ari Holtzman, Ranjay Krishna, Ali Farhadi, Aditya Kusupati | 2024-06-25T01:49:45Z | cs.CL, cs.LG | Many applications today provide users with multiple auto-complete drafts as
they type, including GitHub's code completion, Gmail's smart compose, and
Apple's messaging auto-suggestions. Under the hood, language models support
this by running an autoregressive inference pass to provide a draft.
Consequently, providing $k$ drafts to the user requires running an expensive
language model $k$ times. To alleviate the computation cost of running $k$
inference passes, we propose Superposed Decoding, a new decoding algorithm that
generates $k$ drafts at the computation cost of one autoregressive inference
pass. We achieve this by feeding a superposition of the most recent token
embeddings from the $k$ drafts as input to the next decoding step of the
language model. At every inference step we combine the $k$ drafts with the
top-$k$ tokens to get $k^2$ new drafts and cache the $k$ most likely options,
using an n-gram interpolation with minimal compute overhead to filter out
incoherent generations. Our experiments show that $k$ drafts from Superposed
Decoding are at least as coherent and factual as Nucleus Sampling and Greedy
Decoding respectively, while being at least $2.44\times$ faster for $k\ge3$. In
a compute-normalized setting, user evaluations demonstrably favor text
generated by Superposed Decoding over Nucleus Sampling. Code and more examples
open-sourced at https://github.com/RAIVNLab/SuperposedDecoding. | cs.CL | Superposed Decoding: Multiple Generations from a Single Autoregressive Inference Pass |
2024-05-28T17:36:06Z | http://arxiv.org/pdf/2406.00049v1 | http://arxiv.org/abs/2406.00049v1 | Gonçalo R. A. Faria, Sweta Agrawal, António Farinhas, Ricardo Rei, José G. C. de Souza, André F. T. Martins | 2024-05-28T17:36:06Z | cs.CL, cs.LG | An important challenge in machine translation (MT) is to generate
high-quality and diverse translations. Prior work has shown that the estimated
likelihood from the MT model correlates poorly with translation quality. In
contrast, quality evaluation metrics (such as COMET or BLEURT) exhibit high
correlations with human judgments, which has motivated their use as rerankers
(such as quality-aware and minimum Bayes risk decoding). However, relying on a
single translation with high estimated quality increases the chances of "gaming
the metric''. In this paper, we address the problem of sampling a set of
high-quality and diverse translations. We provide a simple and effective way to
avoid over-reliance on noisy quality estimates by using them as the energy
function of a Gibbs distribution. Instead of looking for a mode in the
distribution, we generate multiple samples from high-density areas through the
Metropolis-Hastings algorithm, a simple Markov chain Monte Carlo approach. The
results show that our proposed method leads to high-quality and diverse outputs
across multiple language pairs (English$\leftrightarrow${German, Russian}) with
two strong decoder-only LLMs (Alma-7b, Tower-7b). | cs.CL | QUEST: Quality-Aware Metropolis-Hastings Sampling for Machine Translation |
2024-05-28T17:22:22Z | http://arxiv.org/pdf/2405.18380v1 | http://arxiv.org/abs/2405.18380v1 | Pengxiang Li, Lu Yin, Xiaowei Gao, Shiwei Liu | 2024-05-28T17:22:22Z | cs.LG, cs.AI, cs.CL | The rapid advancements in Large Language Models (LLMs) have revolutionized
various natural language processing tasks. However, the substantial size of
LLMs presents significant challenges in training or fine-tuning. While
parameter-efficient approaches such as low-rank adaptation (LoRA) have gained
popularity, they often compromise performance compared to full-rank
fine-tuning. In this paper, we propose Outlier-weighed Layerwise Sampled
Low-Rank Projection (OwLore), a new memory-efficient fine-tuning approach,
inspired by the layerwise outlier distribution of LLMs, which dynamically
samples pre-trained layers to fine-tune instead of adding additional adaptors.
We first interpret the outlier phenomenon through the lens of Heavy-Tailed
Self-Regularization theory (HT-SR), discovering that layers with more outliers
tend to be more heavy-tailed and consequently better trained. Inspired by this
finding, OwLore strategically assigns higher sampling probabilities to layers
with more outliers to better leverage the knowledge stored in pre-trained LLMs.
To further mitigate the memory demands of fine-tuning, we integrate gradient
low-rank projection into our approach, which facilitates each layer to be
efficiently trained in a low-rank manner. By incorporating the efficient
characteristics of low-rank and optimal layerwise sampling, OwLore
significantly improves the memory-performance trade-off in LLM pruning. Our
extensive experiments across various architectures, including LLaMa2, LLaMa3,
and Mistral, demonstrate that OwLore consistently outperforms baseline
approaches, including full fine-tuning. Specifically, it achieves up to a 1.1%
average accuracy gain on the Commonsense Reasoning benchmark, a 3.0%
improvement on MMLU, and a notable 10% boost on MT-Bench, while being more
memory efficient. OwLore allows us to fine-tune LLaMa2-7B with only 21GB of
memory. | cs.LG | OwLore: Outlier-weighed Layerwise Sampled Low-Rank Projection for Memory-Efficient LLM Fine-tuning |
2024-05-28T17:14:02Z | http://arxiv.org/pdf/2405.18375v1 | http://arxiv.org/abs/2405.18375v1 | Phakphum Artkaew | 2024-05-28T17:14:02Z | cs.CL | Commonsense reasoning is one of the important aspect of natural language
understanding, with several benchmarks developed to evaluate it. However, only
a few of these benchmarks are available in languages other than English.
Developing parallel benchmarks facilitates cross-lingual evaluation, enabling a
better understanding of different languages. This research introduces a
collection of Winograd Schemas in Thai, a novel dataset designed to evaluate
commonsense reasoning capabilities in the context of the Thai language.
Through a methodology involving native speakers, professional translators,
and thorough validation, the schemas aim to closely reflect Thai language
nuances, idioms, and cultural references while maintaining ambiguity and
commonsense challenges. We evaluate the performance of popular large language
models on this benchmark, revealing their strengths, limitations, and providing
insights into the current state-of-the-art. Results indicate that while models
like GPT-4 and Claude-3-Opus achieve high accuracy in English, their
performance significantly drops in Thai, highlighting the need for further
advancements in multilingual commonsense reasoning. | cs.CL | Thai Winograd Schemas: A Benchmark for Thai Commonsense Reasoning |
2024-05-28T17:08:31Z | http://arxiv.org/pdf/2405.18369v1 | http://arxiv.org/abs/2405.18369v1 | Eshaan Agarwal, Vivek Dani, Tanuja Ganu, Akshay Nambi | 2024-05-28T17:08:31Z | cs.CL, cs.AI, cs.LG | Large language models (LLMs) have revolutionized AI across diverse domains,
showcasing remarkable capabilities. Central to their success is the concept of
prompting, which guides model output generation. However, manual prompt
engineering is labor-intensive and domain-specific, necessitating automated
solutions. This paper introduces PromptWizard, a novel framework leveraging
LLMs to iteratively synthesize and refine prompts tailored to specific tasks.
Unlike existing approaches, PromptWizard optimizes both prompt instructions and
in-context examples, maximizing model performance. The framework iteratively
refines prompts by mutating instructions and incorporating negative examples to
deepen understanding and ensure diversity. It further enhances both
instructions and examples with the aid of a critic, synthesizing new
instructions and examples enriched with detailed reasoning steps for optimal
performance. PromptWizard offers several key features and capabilities,
including computational efficiency compared to state-of-the-art approaches,
adaptability to scenarios with varying amounts of training data, and
effectiveness with smaller LLMs. Rigorous evaluation across 35 tasks on 8
datasets demonstrates PromptWizard's superiority over existing prompt
strategies, showcasing its efficacy and scalability in prompt optimization. | cs.CL | PromptWizard: Task-Aware Agent-driven Prompt Optimization Framework |
2024-05-28T17:01:22Z | http://arxiv.org/pdf/2406.00048v1 | http://arxiv.org/abs/2406.00048v1 | Francesco Cagnetta, Matthieu Wyart | 2024-05-28T17:01:22Z | cs.CL, cond-mat.dis-nn, cs.LG | How much data is required to learn the structure of a language via next-token
prediction? We study this question for synthetic datasets generated via a
Probabilistic Context-Free Grammar (PCFG) -- a hierarchical generative model
that captures the tree-like structure of natural languages. We determine
token-token correlations analytically in our model and show that they can be
used to build a representation of the grammar's hidden variables, the longer
the range the deeper the variable. In addition, a finite training set limits
the resolution of correlations to an effective range, whose size grows with
that of the training set. As a result, a Language Model trained with
increasingly many examples can build a deeper representation of the grammar's
structure, thus reaching good performance despite the high dimensionality of
the problem. We conjecture that the relationship between training set size and
effective range of correlations holds beyond our synthetic datasets. In
particular, our conjecture predicts how the scaling law for the test loss
behaviour with training set size depends on the length of the context window,
which we confirm empirically for a collection of lines from Shakespeare's
plays. | cs.CL | Towards a theory of how the structure of language is acquired by deep neural networks |
2024-05-28T16:56:42Z | http://arxiv.org/pdf/2405.18359v1 | http://arxiv.org/abs/2405.18359v1 | Somnath Kumar, Vaibhav Balloli, Mercy Ranjit, Kabir Ahuja, Tanuja Ganu, Sunayana Sitaram, Kalika Bali, Akshay Nambi | 2024-05-28T16:56:42Z | cs.CL, cs.AI, cs.LG | Large language models (LLMs) are at the forefront of transforming numerous
domains globally. However, their inclusivity and effectiveness remain limited
for non-Latin scripts and low-resource languages. This paper tackles the
imperative challenge of enhancing the multilingual performance of LLMs without
extensive training or fine-tuning. Through systematic investigation and
evaluation of diverse languages using popular question-answering (QA) datasets,
we present novel techniques that unlock the true potential of LLMs in a
polyglot landscape. Our approach encompasses three key strategies that yield
significant improvements in multilingual proficiency. First, by meticulously
optimizing prompts tailored for polyglot LLMs, we unlock their latent
capabilities, resulting in substantial performance boosts across languages.
Second, we introduce a new hybrid approach that synergizes LLM Retrieval
Augmented Generation (RAG) with multilingual embeddings and achieves improved
multilingual task performance. Finally, we introduce a novel learning approach
that dynamically selects the optimal prompt strategy, LLM model, and embedding
model per query at run-time. This dynamic adaptation maximizes the efficacy of
LLMs across languages, outperforming best static and random strategies.
Additionally, our approach adapts configurations in both offline and online
settings, and can seamlessly adapt to new languages and datasets, leading to
substantial advancements in multilingual understanding and generation across
diverse languages. | cs.CL | Bridging the Gap: Dynamic Learning Strategies for Improving Multilingual Performance in LLMs |
2024-05-28T16:55:41Z | http://arxiv.org/pdf/2405.18358v1 | http://arxiv.org/abs/2405.18358v1 | Somnath Kumar, Yash Gadhia, Tanuja Ganu, Akshay Nambi | 2024-05-28T16:55:41Z | cs.CL, cs.AI, cs.CV, cs.LG | Recent advancements in Multi-modal Large Language Models (MLLMs) have
significantly improved their performance in tasks combining vision and
language. However, challenges persist in detailed multi-modal understanding,
comprehension of complex tasks, and reasoning over multi-modal information.
This paper introduces MMCTAgent, a novel multi-modal critical thinking agent
framework designed to address the inherent limitations of current MLLMs in
complex visual reasoning tasks. Inspired by human cognitive processes and
critical thinking, MMCTAgent iteratively analyzes multi-modal information,
decomposes queries, plans strategies, and dynamically evolves its reasoning.
Additionally, MMCTAgent incorporates critical thinking elements such as
verification of final answers and self-reflection through a novel approach that
defines a vision-based critic and identifies task-specific evaluation criteria,
thereby enhancing its decision-making abilities. Through rigorous evaluations
across various image and video understanding benchmarks, we demonstrate that
MMCTAgent (with and without the critic) outperforms both foundational MLLMs and
other tool-augmented pipelines. | cs.CL | MMCTAgent: Multi-modal Critical Thinking Agent Framework for Complex Visual Reasoning |
2024-05-28T16:55:33Z | http://arxiv.org/pdf/2405.18357v2 | http://arxiv.org/abs/2405.18357v2 | Jundong Xu, Hao Fei, Liangming Pan, Qian Liu, Mong-Li Lee, Wynne Hsu | 2024-06-11T07:41:03Z | cs.CL | While the recent Chain-of-Thought (CoT) technique enhances the reasoning
ability of large language models (LLMs) with the theory of mind, it might still
struggle in handling logical reasoning that relies much on symbolic expressions
and rigid deducing rules. To strengthen the logical reasoning capability of
LLMs, we propose a novel Symbolic Chain-of-Thought, namely SymbCoT, a fully
LLM-based framework that integrates symbolic expressions and logic rules with
CoT prompting. Technically, building upon an LLM, SymbCoT 1) first translates
the natural language context into the symbolic format, and then 2) derives a
step-by-step plan to solve the problem with symbolic logical rules, 3) followed
by a verifier to check the translation and reasoning chain. Via thorough
evaluations on 5 standard datasets with both First-Order Logic and Constraint
Optimization symbolic expressions, SymbCoT shows striking improvements over the
CoT method consistently, meanwhile refreshing the current state-of-the-art
performances. We further demonstrate that our system advances in more faithful,
flexible, and explainable logical reasoning. To our knowledge, this is the
first to combine symbolic expressions and rules into CoT for logical reasoning
with LLMs. Code is open at https://github.com/Aiden0526/SymbCoT. | cs.CL | Faithful Logical Reasoning via Symbolic Chain-of-Thought |
2024-05-28T16:48:05Z | http://arxiv.org/pdf/2405.18350v1 | http://arxiv.org/abs/2405.18350v1 | Silvia García Méndez, Milagros Fernández Gavilanes, Enrique Costa Montenegro, Jonathan Juncal Martínez, Francisco Javier González Castaño, Ehud Reiter | 2024-05-28T16:48:05Z | cs.CL, cs.AI | We present an automatic text expansion system to generate English sentences,
which performs automatic Natural Language Generation (NLG) by combining
linguistic rules with statistical approaches. Here, "automatic" means that the
system can generate coherent and correct sentences from a minimum set of words.
From its inception, the design is modular and adaptable to other languages.
This adaptability is one of its greatest advantages. For English, we have
created the highly precise aLexiE lexicon with wide coverage, which represents
a contribution on its own. We have evaluated the resulting NLG library in an
Augmentative and Alternative Communication (AAC) proof of concept, both
directly (by regenerating corpus sentences) and manually (from annotations)
using a popular corpus in the NLG field. We performed a second analysis by
comparing the quality of text expansion in English to Spanish, using an ad-hoc
Spanish-English parallel corpus. The system might also be applied to other
domains such as report and news generation. | cs.CL | A System for Automatic English Text Expansion |
2024-05-28T16:44:02Z | http://arxiv.org/pdf/2405.18348v1 | http://arxiv.org/abs/2405.18348v1 | Sweta Agrawal, António Farinhas, Ricardo Rei, André F. T. Martins | 2024-05-28T16:44:02Z | cs.CL | Automatic metrics for evaluating translation quality are typically validated
by measuring how well they correlate with human assessments. However,
correlation methods tend to capture only the ability of metrics to
differentiate between good and bad source-translation pairs, overlooking their
reliability in distinguishing alternative translations for the same source. In
this paper, we confirm that this is indeed the case by showing that current
metrics are insensitive to nuanced differences in translation quality. This
effect is most pronounced when the quality is high and the variance among
alternatives is low. Given this finding, we shift towards detecting
high-quality correct translations, an important problem in practical
decision-making scenarios where a binary check of correctness is prioritized
over a nuanced evaluation of quality. Using the MQM framework as the gold
standard, we systematically stress-test the ability of current metrics to
identify translations with no errors as marked by humans. Our findings reveal
that current metrics often over or underestimate translation quality,
indicating significant room for improvement in automatic evaluation methods. | cs.CL | Can Automatic Metrics Assess High-Quality Translations? |
2024-05-28T16:42:43Z | http://arxiv.org/pdf/2405.18344v1 | http://arxiv.org/abs/2405.18344v1 | Aryan Rangapur, Aman Rangapur | 2024-05-28T16:42:43Z | cs.CL, cs.AI, I.7, I.m | Large language models have gained considerable interest for their impressive
performance on various tasks. Within this domain, ChatGPT and GPT-4, developed
by OpenAI, and the Gemini, developed by Google, have emerged as particularly
popular among early adopters. Additionally, Mixtral by Mistral AI and Claude by
Anthropic are newly released, further expanding the landscape of advanced
language models. These models are viewed as disruptive technologies with
applications spanning customer service, education, healthcare, and finance.
More recently, Mistral has entered the scene, captivating users with its unique
ability to generate creative content. Understanding the perspectives of these
users is crucial, as they can offer valuable insights into the potential
strengths, weaknesses, and overall success or failure of these technologies in
various domains. This research delves into the responses generated by ChatGPT,
GPT-4, Gemini, Mixtral and Claude across different Conversational QA corpora.
Evaluation scores were meticulously computed and subsequently compared to
ascertain the overall performance of these models. Our study pinpointed
instances where these models provided inaccurate answers to questions, offering
insights into potential areas where they might be susceptible to errors. In
essence, this research provides a comprehensive comparison and evaluation of
these state of-the-art language models, shedding light on their capabilities
while also highlighting potential areas for improvement | cs.CL | The Battle of LLMs: A Comparative Study in Conversational QA Tasks |
2024-05-28T16:28:58Z | http://arxiv.org/pdf/2405.18335v1 | http://arxiv.org/abs/2405.18335v1 | Silvia García Méndez, Fátima Leal, Benedita Malheiro, Juan Carlos Burguillo Rial | 2024-05-28T16:28:58Z | cs.CL, cs.AI, cs.LG | Wiki articles are created and maintained by a crowd of editors, producing a
continuous stream of reviews. Reviews can take the form of additions, reverts,
or both. This crowdsourcing model is exposed to manipulation since neither
reviews nor editors are automatically screened and purged. To protect articles
against vandalism or damage, the stream of reviews can be mined to classify
reviews and profile editors in real-time. The goal of this work is to
anticipate and explain which reviews to revert. This way, editors are informed
why their edits will be reverted. The proposed method employs stream-based
processing, updating the profiling and classification models on each incoming
event. The profiling uses side and content-based features employing Natural
Language Processing, and editor profiles are incrementally updated based on
their reviews. Since the proposed method relies on self-explainable
classification algorithms, it is possible to understand why a review has been
classified as a revert or a non-revert. In addition, this work contributes an
algorithm for generating synthetic data for class balancing, making the final
classification fairer. The proposed online method was tested with a real data
set from Wikivoyage, which was balanced through the aforementioned synthetic
data generation. The results attained near-90 % values for all evaluation
metrics (accuracy, precision, recall, and F-measure). | cs.CL | Interpretable classification of wiki-review streams |
2024-05-28T16:11:11Z | http://arxiv.org/pdf/2405.18320v1 | http://arxiv.org/abs/2405.18320v1 | Mihir Chauhan, Mohammad Abuzar Shaikh, Bina Ramamurthy, Mingchen Gao, Siwei Lyu, Sargur Srihari | 2024-05-28T16:11:11Z | cs.CV, cs.AI, cs.CL | We present SSL-HV: Self-Supervised Learning approaches applied to the task of
Handwriting Verification. This task involves determining whether a given pair
of handwritten images originate from the same or different writer distribution.
We have compared the performance of multiple generative, contrastive SSL
approaches against handcrafted feature extractors and supervised learning on
CEDAR AND dataset. We show that ResNet based Variational Auto-Encoder (VAE)
outperforms other generative approaches achieving 76.3% accuracy, while
ResNet-18 fine-tuned using Variance-Invariance-Covariance Regularization
(VICReg) outperforms other contrastive approaches achieving 78% accuracy. Using
a pre-trained VAE and VICReg for the downstream task of writer verification we
observed a relative improvement in accuracy of 6.7% and 9% over ResNet-18
supervised baseline with 10% writer labels. | cs.CV | Self-Supervised Learning Based Handwriting Verification |
2024-05-28T16:01:19Z | http://arxiv.org/pdf/2405.18308v1 | http://arxiv.org/abs/2405.18308v1 | Thomas Muller, Ryan Cotterell, Alexander Fraser, Hinrich Schütze | 2024-05-28T16:01:19Z | cs.CL | We present LEMMING, a modular log-linear model that jointly models
lemmatization and tagging and supports the integration of arbitrary global
features. It is trainable on corpora annotated with gold standard tags and
lemmata and does not rely on morphological dictionaries or analyzers. LEMMING
sets the new state of the art in token-based statistical lemmatization on six
languages; e.g., for Czech lemmatization, we reduce the error by 60%, from 4.05
to 1.58. We also give empirical evidence that jointly modeling morphological
tags and lemmata is mutually beneficial. | cs.CL | Joint Lemmatization and Morphological Tagging with LEMMING |
2024-05-28T15:47:11Z | http://arxiv.org/pdf/2405.18292v1 | http://arxiv.org/abs/2405.18292v1 | Renzhi Wang, Piji Li | 2024-05-28T15:47:11Z | cs.CL | Parameter-Efficient Fine-Tuning (PEFT) methods enable efficient adaptation of
Large Language Models (LLMs) to various downstream applications. However, the
effectiveness of the PEFT diminishes notably when downstream tasks require
accurate learning of factual knowledge. In this paper, we adopt a semantic
perspective to investigate this phenomenon, uncovering the reasons behind
PEFT's limitations in knowledge learning task. Our findings reveal that: (1)
PEFT presents a notable risk of pushing the model away from the intended
knowledge target; (2) multiple knowledge interfere with each other, and such
interference suppresses the learning and expression of knowledge features.
Based on these insights, we introduce a data filtering strategy to exclude data
that is detrimental to knowledge learning and a re-weighted learning strategy
to make the model attentive to semantic distance during knowledge learning.
Experimental results demonstrate the effectiveness of the proposed method on
open-source large language model, further validate the semantic challenge in
PEFT, thus paving the way for future research. | cs.CL | Semantic are Beacons: A Semantic Perspective for Unveiling Parameter-Efficient Fine-Tuning in Knowledge Learning |
2024-05-28T15:11:17Z | http://arxiv.org/pdf/2405.18258v1 | http://arxiv.org/abs/2405.18258v1 | Qing Zhou, Junlin Huang, Qiang Li, Junyu Gao, Qi Wang | 2024-05-28T15:11:17Z | cs.CV, cs.AI, cs.CL | From paired image-text training to text-only training for image captioning,
the pursuit of relaxing the requirements for high-cost and large-scale
annotation of good quality data remains consistent. In this paper, we propose
Text-only Synthesis for Image Captioning (ToCa), which further advances this
relaxation with fewer human labor and less computing time. Specifically, we
deconstruct caption text into structures and lexical words, which serve as the
fundamental components of the caption. By combining different structures and
lexical words as inputs to the large language model, massive captions that
contain various patterns of lexical words are generated. This method not only
approaches the target domain but also surpasses it by generating new captions,
thereby enhancing the zero-shot generalization ability of the model.
Considering the different levels of data access in the real world, we define
three synthesis scenarios: cross-domain synthesis, in-domain synthesis, and
data-efficient synthesis. Experiments in these scenarios demonstrate the
generalizability, transferability and practicability of ToCa with a nearly 5
CIDEr improvement for zero-shot cross-domain captioning and a maximum increase
of over 20 CIDEr for data-efficient captioning. | cs.CV | Text-only Synthesis for Image Captioning |
2024-05-28T14:50:22Z | http://arxiv.org/pdf/2405.18241v1 | http://arxiv.org/abs/2405.18241v1 | Wei Liu, Ming Xiang, Nai Ding | 2024-05-28T14:50:22Z | cs.CL, cs.AI | Understanding how sentences are internally represented in the human brain, as
well as in large language models (LLMs) such as ChatGPT, is a major challenge
for cognitive science. Classic linguistic theories propose that the brain
represents a sentence by parsing it into hierarchically organized constituents.
In contrast, LLMs do not explicitly parse linguistic constituents and their
latent representations remains poorly explained. Here, we demonstrate that
humans and LLMs construct similar latent representations of hierarchical
linguistic constituents by analyzing their behaviors during a novel one-shot
learning task, in which they infer which words should be deleted from a
sentence. Both humans and LLMs tend to delete a constituent, instead of a
nonconstituent word string. In contrast, a naive sequence processing model that
has access to word properties and ordinal positions does not show this
property. Based on the word deletion behaviors, we can reconstruct the latent
constituency tree representation of a sentence for both humans and LLMs. These
results demonstrate that a latent tree-structured constituency representation
can emerge in both the human brain and LLMs. | cs.CL | Active Use of Latent Constituency Representation in both Humans and Large Language Models |
2024-05-28T14:13:32Z | http://arxiv.org/pdf/2405.18208v1 | http://arxiv.org/abs/2405.18208v1 | Chengxing Xie, Difan Zou | 2024-05-28T14:13:32Z | cs.AI, cs.CL, cs.LG | Recent studies have highlighted their proficiency in some simple tasks like
writing and coding through various reasoning strategies. However, LLM agents
still struggle with tasks that require comprehensive planning, a process that
challenges current models and remains a critical research issue. In this study,
we concentrate on travel planning, a Multi-Phases planning problem, that
involves multiple interconnected stages, such as outlining, information
gathering, and planning, often characterized by the need to manage various
constraints and uncertainties. Existing reasoning approaches have struggled to
effectively address this complex task. Our research aims to address this
challenge by developing a human-like planning framework for LLM agents, i.e.,
guiding the LLM agent to simulate various steps that humans take when solving
Multi-Phases problems. Specifically, we implement several strategies to enable
LLM agents to generate a coherent outline for each travel query, mirroring
human planning patterns. Additionally, we integrate Strategy Block and
Knowledge Block into our framework: Strategy Block facilitates information
collection, while Knowledge Block provides essential information for detailed
planning. Through our extensive experiments, we demonstrate that our framework
significantly improves the planning capabilities of LLM agents, enabling them
to tackle the travel planning task with improved efficiency and effectiveness.
Our experimental results showcase the exceptional performance of the proposed
framework; when combined with GPT-4-Turbo, it attains $10\times$ the
performance gains in comparison to the baseline framework deployed on
GPT-4-Turbo. | cs.AI | A Human-Like Reasoning Framework for Multi-Phases Planning Task with Large Language Models |
2024-05-28T14:11:01Z | http://arxiv.org/pdf/2405.18203v2 | http://arxiv.org/abs/2405.18203v2 | Wei Zhu, Aaron Xuxiang Tian, Congrui Yin, Yuan Ni, Xiaoling Wang, Guotong Xie | 2024-06-07T06:41:18Z | cs.CL | Soft prompt tuning is a widely studied parameter-efficient fine-tuning
method. However, it has a clear drawback: many soft tokens must be inserted
into the input sequences to guarantee downstream performance. As a result, soft
prompt tuning is less considered than Low-rank adaptation (LoRA) in the large
language modeling (LLM) era. In this work, we propose a novel prompt tuning
method, Instruction-Aware Prompt Tuning (IAPT), that requires only four soft
tokens. First, we install a parameter-efficient soft prompt generator at each
Transformer layer to generate idiosyncratic soft prompts for each input
instruction. The generated soft prompts can be seen as a semantic summary of
the input instructions and can effectively guide the output generation. Second,
the soft prompt generators are modules with a bottleneck architecture
consisting of a self-attention pooling operation, two linear projections, and
an activation function. Pilot experiments show that prompt generators at
different Transformer layers require different activation functions. Thus, we
propose to learn the idiosyncratic activation functions for prompt generators
automatically with the help of rational functions. We have conducted
experiments on various tasks, and the experimental results demonstrate that (a)
our IAPT method can outperform the recent baselines with comparable tunable
parameters. (b) Our IAPT method is more efficient than LoRA under the
single-backbone multi-tenant setting. | cs.CL | IAPT: Instruction-Aware Prompt Tuning for Large Language Models |
2024-05-28T13:09:22Z | http://arxiv.org/pdf/2406.00046v2 | http://arxiv.org/abs/2406.00046v2 | Tong Chen, Danny Wang, Xurong Liang, Marten Risius, Gianluca Demartini, Hongzhi Yin | 2024-06-11T13:18:14Z | cs.CL, cs.LG | To counter the side effect brought by the proliferation of social media
platforms, hate speech detection (HSD) plays a vital role in halting the
dissemination of toxic online posts at an early stage. However, given the
ubiquitous topical communities on social media, a trained HSD classifier easily
becomes biased towards specific targeted groups (e.g., female and black
people), where a high rate of false positive/negative results can significantly
impair public trust in the fairness of content moderation mechanisms, and
eventually harm the diversity of online society. Although existing
fairness-aware HSD methods can smooth out some discrepancies across targeted
groups, they are mostly specific to a narrow selection of targets that are
assumed to be known and fixed. This inevitably prevents those methods from
generalizing to real-world use cases where new targeted groups constantly
emerge over time. To tackle this defect, we propose Generalizable target-aware
Fairness (GetFair), a new method for fairly classifying each post that contains
diverse and even unseen targets during inference. To remove the HSD
classifier's spurious dependence on target-related features, GetFair trains a
series of filter functions in an adversarial pipeline, so as to deceive the
discriminator that recovers the targeted group from filtered post embeddings.
To maintain scalability and generalizability, we innovatively parameterize all
filter functions via a hypernetwork that is regularized by the semantic
affinity among targets. Taking a target's pretrained word embedding as input,
the hypernetwork generates the weights used by each target-specific filter
on-the-fly without storing dedicated filter parameters. Finally, comparative
experiments on two HSD datasets have shown advantageous performance of GetFair
on out-of-sample targets. | cs.CL | Hate Speech Detection with Generalizable Target-aware Fairness |
2024-05-28T12:47:43Z | http://arxiv.org/pdf/2406.17790v1 | http://arxiv.org/abs/2406.17790v1 | Areeg Fahad Rasheed, M. Zarkoosh | 2024-05-28T12:47:43Z | cs.CL | Within few-shot learning, in-context learning (ICL) has become a potential
method for leveraging contextual information to improve model performance on
small amounts of data or in resource-constrained environments where training
models on large datasets is prohibitive. However, the quality of the selected
sample in a few shots severely limits the usefulness of ICL. The primary goal
of this paper is to enhance the performance of evaluation metrics for
in-context learning by selecting high-quality samples in few-shot learning
scenarios. We employ the chi-square test to identify high-quality samples and
compare the results with those obtained using low-quality samples. Our findings
demonstrate that utilizing high-quality samples leads to improved performance
with respect to all evaluated metrics. | cs.CL | Mashee at SemEval-2024 Task 8: The Impact of Samples Quality on the Performance of In-Context Learning for Machine Text Classification |
2024-05-28T12:23:39Z | http://arxiv.org/pdf/2405.18115v1 | http://arxiv.org/abs/2405.18115v1 | Gili Goldin, Nick Howell, Noam Ordan, Ella Rabinovich, Shuly Wintner | 2024-05-28T12:23:39Z | cs.CL, 68T50, I.2.7 | We present the Knesset Corpus, a corpus of Hebrew parliamentary proceedings
containing over 30 million sentences (over 384 million tokens) from all the
(plenary and committee) protocols held in the Israeli parliament between 1998
and 2022. Sentences are annotated with morpho-syntactic information and are
associated with detailed meta-information reflecting demographic and political
properties of the speakers, based on a large database of parliament members and
factions that we compiled. We discuss the structure and composition of the
corpus and the various processing steps we applied to it. To demonstrate the
utility of this novel dataset we present two use cases. We show that the corpus
can be used to examine historical developments in the style of political
discussions by showing a reduction in lexical richness in the proceedings over
time. We also investigate some differences between the styles of men and women
speakers. These use cases exemplify the potential of the corpus to shed light
on important trends in the Israeli society, supporting research in linguistics,
political science, communication, law, etc. | cs.CL | The Knesset Corpus: An Annotated Corpus of Hebrew Parliamentary Proceedings |
2024-05-28T12:23:16Z | http://arxiv.org/pdf/2405.18113v1 | http://arxiv.org/abs/2405.18113v1 | Hongda Sun, Hongzhan Lin, Haiyu Yan, Chen Zhu, Yang Song, Xin Gao, Shuo Shang, Rui Yan | 2024-05-28T12:23:16Z | cs.CL, cs.AI | The emergence of online recruitment services has revolutionized the
traditional landscape of job seeking and recruitment, necessitating the
development of high-quality industrial applications to improve person-job
fitting. Existing methods generally rely on modeling the latent semantics of
resumes and job descriptions and learning a matching function between them.
Inspired by the powerful role-playing capabilities of Large Language Models
(LLMs), we propose to introduce a mock interview process between LLM-played
interviewers and candidates. The mock interview conversations can provide
additional evidence for candidate evaluation, thereby augmenting traditional
person-job fitting based solely on resumes and job descriptions. However,
characterizing these two roles in online recruitment still presents several
challenges, such as developing the skills to raise interview questions,
formulating appropriate answers, and evaluating two-sided fitness. To this end,
we propose MockLLM, a novel applicable framework that divides the person-job
matching process into two modules: mock interview generation and two-sided
evaluation in handshake protocol, jointly enhancing their performance through
collaborative behaviors between interviewers and candidates. We design a
role-playing framework as a multi-role and multi-behavior paradigm to enable a
single LLM agent to effectively behave with multiple functions for both
parties. Moreover, we propose reflection memory generation and dynamic prompt
modification techniques to refine the behaviors of both sides, enabling
continuous optimization of the augmented additional evidence. Extensive
experimental results show that MockLLM can achieve the best performance on
person-job matching accompanied by high mock interview quality, envisioning its
emerging application in real online recruitment in the future. | cs.CL | Facilitating Multi-Role and Multi-Behavior Collaboration of Large Language Models for Online Job Seeking and Recruiting |
2024-05-28T12:18:50Z | http://arxiv.org/pdf/2405.18111v2 | http://arxiv.org/abs/2405.18111v2 | Junda Zhu, Lingyong Yan, Haibo Shi, Dawei Yin, Lei Sha | 2024-06-16T12:30:32Z | cs.CL | Large language models (LLMs) are proven to benefit a lot from
retrieval-augmented generation (RAG) in alleviating hallucinations confronted
with knowledge-intensive questions. RAG adopts information retrieval techniques
to inject external knowledge from semantic-relevant documents as input
contexts. However, due to today's Internet being flooded with numerous noisy
and fabricating content, it is inevitable that RAG systems are vulnerable to
these noises and prone to respond incorrectly. To this end, we propose to
optimize the retrieval-augmented Generator with a Adversarial Tuning
Multi-agent system (ATM). The ATM steers the Generator to have a robust
perspective of useful documents for question answering with the help of an
auxiliary Attacker agent. The Generator and the Attacker are tuned
adversarially for several iterations. After rounds of multi-agent iterative
tuning, the Generator can eventually better discriminate useful documents
amongst fabrications. The experimental results verify the effectiveness of ATM
and we also observe that the Generator can achieve better performance compared
to state-of-the-art baselines. | cs.CL | ATM: Adversarial Tuning Multi-agent System Makes a Robust Retrieval-Augmented Generator |
2024-05-28T11:19:39Z | http://arxiv.org/pdf/2405.18061v1 | http://arxiv.org/abs/2405.18061v1 | Neha Sharma, Kairit Sirts | 2024-05-28T11:19:39Z | cs.CL | Research exploring linguistic markers in individuals with depression has
demonstrated that language usage can serve as an indicator of mental health.
This study investigates the impact of discussion topic as context on linguistic
markers and emotional expression in depression, using a Reddit dataset to
explore interaction effects. Contrary to common findings, our sentiment
analysis revealed a broader range of emotional intensity in depressed
individuals, with both higher negative and positive sentiments than controls.
This pattern was driven by posts containing no emotion words, revealing the
limitations of the lexicon based approaches in capturing the full emotional
context. We observed several interesting results demonstrating the importance
of contextual analyses. For instance, the use of 1st person singular pronouns
and words related to anger and sadness correlated with increased positive
sentiments, whereas a higher rate of present-focused words was associated with
more negative sentiments. Our findings highlight the importance of discussion
contexts while interpreting the language used in depression, revealing that the
emotional intensity and meaning of linguistic markers can vary based on the
topic of discussion. | cs.CL | Context is Important in Depressive Language: A Study of the Interaction Between the Sentiments and Linguistic Markers in Reddit Discussions |
2024-05-28T11:19:13Z | http://arxiv.org/pdf/2405.18060v1 | http://arxiv.org/abs/2405.18060v1 | Mehrimah Amirpour, Reza Azmi | 2024-05-28T11:19:13Z | cs.CL | The PRFashion24 dataset is a comprehensive Persian dataset collected from
various online fashion stores, spanning from April 2020 to March 2024. With
767,272 reviews, it is the first dataset in its kind that encompasses diverse
categories within the fashion industry in the Persian language. The goal of
this study is to harness deep learning techniques, specifically Long Short-Term
Memory (LSTM) networks and a combination of Bidirectional LSTM and
Convolutional Neural Network (BiLSTM-CNN), to analyze and reveal sentiments
towards online fashion shopping. The LSTM model yielded an accuracy of 81.23%,
while the BiLSTM-CNN model reached 82.89%. This research aims not only to
introduce a diverse dataset in the field of fashion but also to enhance the
public's understanding of opinions on online fashion shopping, which
predominantly reflect a positive sentiment. Upon publication, both the
optimized models and the PRFashion24 dataset will be available on GitHub. | cs.CL | PRFashion24: A Dataset for Sentiment Analysis of Fashion Products Reviews in Persian |
2024-05-28T11:13:40Z | http://arxiv.org/pdf/2406.17789v1 | http://arxiv.org/abs/2406.17789v1 | Irene Plaza, Nina Melero, Cristina del Pozo, Javier Conde, Pedro Reviriego, Marina Mayor-Rocher, María Grandury | 2024-05-28T11:13:40Z | cs.CL, cs.AI | The evaluation of Large Language Models (LLMs) is a key element in their
continuous improvement process and many benchmarks have been developed to
assess the performance of LLMs in different tasks and topics. As LLMs become
adopted worldwide, evaluating them in languages other than English is
increasingly important. However, most LLM benchmarks are simply translated
using an automated tool and then run in the target language. This means that
the results depend not only on the LLM performance in that language but also on
the quality of the translation. In this paper, we consider the case of the
well-known Massive Multitask Language Understanding (MMLU) benchmark. Selected
categories of the benchmark are translated into Spanish using Azure Translator
and ChatGPT4 and run on ChatGPT4. Next, the results are processed to identify
the test items that produce different answers in Spanish and English. Those are
then analyzed manually to understand if the automatic translation caused the
change. The results show that a significant fraction of the failing items can
be attributed to mistakes in the translation of the benchmark. These results
make a strong case for improving benchmarks in languages other than English by
at least revising the translations of the items and preferably by adapting the
tests to the target language by experts. | cs.CL | Spanish and LLM Benchmarks: is MMLU Lost in Translation? |
2024-05-28T10:39:10Z | http://arxiv.org/pdf/2405.18035v2 | http://arxiv.org/abs/2405.18035v2 | Guangmin Zheng, Jin Wang, Liang-Chih Yu, Xuejie Zhang | 2024-05-29T06:35:46Z | cs.CL | Aspect-based sentiment analysis (ABSA) identifies sentiment information
related to specific aspects and provides deeper market insights to businesses
and organizations. With the emergence of large language models (LMs), recent
studies have proposed using fixed examples for instruction tuning to
reformulate ABSA as a generation task. However, the performance is sensitive to
the selection of in-context examples; several retrieval methods are based on
surface similarity and are independent of the LM generative objective. This
study proposes an instruction learning method with retrieval-based example
ranking for ABSA tasks. For each target sample, an LM was applied as a scorer
to estimate the likelihood of the output given the input and a candidate
example as the prompt, and training examples were labeled as positive or
negative by ranking the scores. An alternating training schema is proposed to
train both the retriever and LM. Instructional prompts can be constructed using
high-quality examples. The LM is used for both scoring and inference, improving
the generation efficiency without incurring additional computational costs or
training difficulties. Extensive experiments on three ABSA subtasks verified
the effectiveness of the proposed method, demonstrating its superiority over
various strong baseline models. Code and data are released at
https://github.com/zgMin/IT-RER-ABSA. | cs.CL | Instruction Tuning with Retrieval-based Examples Ranking for Aspect-based Sentiment Analysis |
2024-05-28T10:20:29Z | http://arxiv.org/pdf/2405.18028v1 | http://arxiv.org/abs/2405.18028v1 | Aryo Pradipta Gema, Chaeeun Lee, Pasquale Minervini, Luke Daines, T. Ian Simpson, Beatrice Alex | 2024-05-28T10:20:29Z | cs.CL, cs.AI | The MEDIQA-CORR 2024 shared task aims to assess the ability of Large Language
Models (LLMs) to identify and correct medical errors in clinical notes. In this
study, we evaluate the capability of general LLMs, specifically GPT-3.5 and
GPT-4, to identify and correct medical errors with multiple prompting
strategies. Recognising the limitation of LLMs in generating accurate
corrections only via prompting strategies, we propose incorporating error-span
predictions from a smaller, fine-tuned model in two ways: 1) by presenting it
as a hint in the prompt and 2) by framing it as multiple-choice questions from
which the LLM can choose the best correction. We found that our proposed
prompting strategies significantly improve the LLM's ability to generate
corrections. Our best-performing solution with 8-shot + CoT + hints ranked
sixth in the shared task leaderboard. Additionally, our comprehensive analyses
show the impact of the location of the error sentence, the prompted role, and
the position of the multiple-choice option on the accuracy of the LLM. This
prompts further questions about the readiness of LLM to be implemented in
real-world clinical settings. | cs.CL | Edinburgh Clinical NLP at MEDIQA-CORR 2024: Guiding Large Language Models with Hints |
2024-05-28T10:19:18Z | http://arxiv.org/pdf/2405.18027v1 | http://arxiv.org/abs/2405.18027v1 | Jaewoo Ahn, Taehyun Lee, Junyoung Lim, Jin-Hwa Kim, Sangdoo Yun, Hwaran Lee, Gunhee Kim | 2024-05-28T10:19:18Z | cs.CL | While Large Language Models (LLMs) can serve as agents to simulate human
behaviors (i.e., role-playing agents), we emphasize the importance of
point-in-time role-playing. This situates characters at specific moments in the
narrative progression for three main reasons: (i) enhancing users' narrative
immersion, (ii) avoiding spoilers, and (iii) fostering engagement in fandom
role-playing. To accurately represent characters at specific time points,
agents must avoid character hallucination, where they display knowledge that
contradicts their characters' identities and historical timelines. We introduce
TimeChara, a new benchmark designed to evaluate point-in-time character
hallucination in role-playing LLMs. Comprising 10,895 instances generated
through an automated pipeline, this benchmark reveals significant hallucination
issues in current state-of-the-art LLMs (e.g., GPT-4o). To counter this
challenge, we propose Narrative-Experts, a method that decomposes the reasoning
steps and utilizes narrative experts to reduce point-in-time character
hallucinations effectively. Still, our findings with TimeChara highlight the
ongoing challenges of point-in-time character hallucination, calling for
further study. | cs.CL | TimeChara: Evaluating Point-in-Time Character Hallucination of Role-Playing Large Language Models |
2024-05-28T09:57:28Z | http://arxiv.org/pdf/2405.18015v1 | http://arxiv.org/abs/2405.18015v1 | Xiang Dai, Sarvnaz Karimi, Abeed Sarker, Ben Hachey, Cecile Paris | 2024-05-28T09:57:28Z | cs.CL | Objective. Active adverse event surveillance monitors Adverse Drug Events
(ADE) from different data sources, such as electronic health records, medical
literature, social media and search engine logs. Over years, many datasets are
created, and shared tasks are organised to facilitate active adverse event
surveillance. However, most-if not all-datasets or shared tasks focus on
extracting ADEs from a particular type of text. Domain generalisation-the
ability of a machine learning model to perform well on new, unseen domains
(text types)-is under-explored. Given the rapid advancements in natural
language processing, one unanswered question is how far we are from having a
single ADE extraction model that are effective on various types of text, such
as scientific literature and social media posts}. Methods. We contribute to
answering this question by building a multi-domain benchmark for adverse drug
event extraction, which we named MultiADE. The new benchmark comprises several
existing datasets sampled from different text types and our newly created
dataset-CADECv2, which is an extension of CADEC (Karimi, et al., 2015),
covering online posts regarding more diverse drugs than CADEC. Our new dataset
is carefully annotated by human annotators following detailed annotation
guidelines. Conclusion. Our benchmark results show that the generalisation of
the trained models is far from perfect, making it infeasible to be deployed to
process different types of text. In addition, although intermediate transfer
learning is a promising approach to utilising existing resources, further
investigation is needed on methods of domain adaptation, particularly
cost-effective methods to select useful training instances. | cs.CL | MultiADE: A Multi-domain Benchmark for Adverse Drug Event Extraction |
2024-05-28T09:50:46Z | http://arxiv.org/pdf/2405.18009v1 | http://arxiv.org/abs/2405.18009v1 | Zican Dong, Junyi Li, Xin Men, Wayne Xin Zhao, Bingbing Wang, Zhen Tian, Weipeng Chen, Ji-Rong Wen | 2024-05-28T09:50:46Z | cs.CL, cs.LG | Transformer-based large language models (LLMs) typically have a limited
context window, resulting in significant performance degradation when
processing text beyond the length of the context window. Extensive studies have
been proposed to extend the context window and achieve length extrapolation of
LLMs, but there is still a lack of in-depth interpretation of these approaches.
In this study, we explore the positional information within and beyond the
context window for deciphering the underlying mechanism of LLMs. By using a
mean-based decomposition method, we disentangle positional vectors from hidden
states of LLMs and analyze their formation and effect on attention.
Furthermore, when texts exceed the context window, we analyze the change of
positional vectors in two settings, i.e., direct extrapolation and context
window extension. Based on our findings, we design two training-free context
window extension methods, positional vector replacement and attention window
extension. Experimental results show that our methods can effectively extend
the context window length. | cs.CL | Exploring Context Window of Large Language Models via Decomposed Positional Vectors |
2024-05-28T09:34:50Z | http://arxiv.org/pdf/2405.17998v1 | http://arxiv.org/abs/2405.17998v1 | Yuqi Zhou, Sunhao Dai, Liang Pang, Gang Wang, Zhenhua Dong, Jun Xu, Ji-Rong Wen | 2024-05-28T09:34:50Z | cs.IR, cs.AI, cs.CL | Recently, researchers have uncovered that neural retrieval models prefer
AI-generated content (AIGC), called source bias. Compared to active search
behavior, recommendation represents another important means of information
acquisition, where users are more prone to source bias. Furthermore, delving
into the recommendation scenario, as AIGC becomes integrated within the
feedback loop involving users, data, and the recommender system, it
progressively contaminates the candidate items, the user interaction history,
and ultimately, the data used to train the recommendation models. How and to
what extent the source bias affects the neural recommendation models within
feedback loop remains unknown. In this study, we extend the investigation of
source bias into the realm of recommender systems, specifically examining its
impact across different phases of the feedback loop. We conceptualize the
progression of AIGC integration into the recommendation content ecosystem in
three distinct phases-HGC dominate, HGC-AIGC coexist, and AIGC dominance-each
representing past, present, and future states, respectively. Through extensive
experiments across three datasets from diverse domains, we demonstrate the
prevalence of source bias and reveal a potential digital echo chamber with
source bias amplification throughout the feedback loop. This trend risks
creating a recommender ecosystem with limited information source, such as AIGC,
being disproportionately recommended. To counteract this bias and prevent its
escalation in the feedback loop, we introduce a black-box debiasing method that
maintains model impartiality towards both HGC and AIGC. Our experimental
results validate the effectiveness of the proposed debiasing method, confirming
its potential to disrupt the feedback loop. | cs.IR | Source Echo Chamber: Exploring the Escalation of Source Bias in User, Data, and Recommender System Feedback Loop |
2024-05-28T09:24:52Z | http://arxiv.org/pdf/2405.17992v1 | http://arxiv.org/abs/2405.17992v1 | Laurent Bonnasse-Gahot, Christophe Pallier | 2024-05-28T09:24:52Z | cs.CL, cs.AI, q-bio.NC | Over the past decade, studies of naturalistic language processing where
participants are scanned while listening to continuous text have flourished.
Using word embeddings at first, then large language models, researchers have
created encoding models to analyze the brain signals. Presenting these models
with the same text as the participants allows to identify brain areas where
there is a significant correlation between the functional magnetic resonance
imaging (fMRI) time series and the ones predicted by the models' artificial
neurons. One intriguing finding from these studies is that they have revealed
highly symmetric bilateral activation patterns, somewhat at odds with the
well-known left lateralization of language processing. Here, we report analyses
of an fMRI dataset where we manipulate the complexity of large language models,
testing 28 pretrained models from 8 different families, ranging from 124M to
14.2B parameters. First, we observe that the performance of models in
predicting brain responses follows a scaling law, where the fit with brain
activity increases linearly with the logarithm of the number of parameters of
the model (and its performance on natural language processing tasks). Second,
we show that a left-right asymmetry gradually appears as model size increases,
and that the difference in left-right brain correlations also follows a scaling
law. Whereas the smallest models show no asymmetry, larger models fit better
and better left hemispheric activations than right hemispheric ones. This
finding reconciles computational analyses of brain activity using large
language models with the classic observation from aphasic patients showing left
hemisphere dominance for language. | cs.CL | fMRI predictors based on language models of increasing complexity recover brain left lateralization |
2024-05-28T09:12:44Z | http://arxiv.org/pdf/2405.17980v1 | http://arxiv.org/abs/2405.17980v1 | Anirudh Phukan, Shwetha Somasundaram, Apoorv Saxena, Koustava Goswami, Balaji Vasan Srinivasan | 2024-05-28T09:12:44Z | cs.CL | With the enhancement in the field of generative artificial intelligence (AI),
contextual question answering has become extremely relevant. Attributing model
generations to the input source document is essential to ensure trustworthiness
and reliability. We observe that when large language models (LLMs) are used for
contextual question answering, the output answer often consists of text copied
verbatim from the input prompt which is linked together with "glue text"
generated by the LLM. Motivated by this, we propose that LLMs have an inherent
awareness from where the text was copied, likely captured in the hidden states
of the LLM. We introduce a novel method for attribution in contextual question
answering, leveraging the hidden state representations of LLMs. Our approach
bypasses the need for extensive model retraining and retrieval model overhead,
offering granular attributions and preserving the quality of generated answers.
Our experimental results demonstrate that our method performs on par or better
than GPT-4 at identifying verbatim copied segments in LLM generations and in
attributing these segments to their source. Importantly, our method shows
robust performance across various LLM architectures, highlighting its broad
applicability. Additionally, we present Verifiability-granular, an attribution
dataset which has token level annotations for LLM generations in the contextual
question answering setup. | cs.CL | Peering into the Mind of Language Models: An Approach for Attribution in Contextual Question Answering |
2024-05-28T09:06:38Z | http://arxiv.org/pdf/2405.17978v1 | http://arxiv.org/abs/2405.17978v1 | Xiaobao Wu, Thong Nguyen, Delvin Ce Zhang, William Yang Wang, Anh Tuan Luu | 2024-05-28T09:06:38Z | cs.CL, cs.AI | Topic models have been evolving rapidly over the years, from conventional to
recent neural models. However, existing topic models generally struggle with
either effectiveness, efficiency, or stability, highly impeding their practical
applications. In this paper, we propose FASTopic, a fast, adaptive, stable, and
transferable topic model. FASTopic follows a new paradigm: Dual
Semantic-relation Reconstruction (DSR). Instead of previous conventional,
neural VAE-based or clustering-based methods, DSR discovers latent topics by
reconstruction through modeling the semantic relations among document, topic,
and word embeddings. This brings about a neat and efficient topic modeling
framework. We further propose a novel Embedding Transport Plan (ETP) method.
Rather than early straightforward approaches, ETP explicitly regularizes the
semantic relations as optimal transport plans. This addresses the relation bias
issue and thus leads to effective topic modeling. Extensive experiments on
benchmark datasets demonstrate that our FASTopic shows superior effectiveness,
efficiency, adaptivity, stability, and transferability, compared to
state-of-the-art baselines across various scenarios. Our code is available at
https://github.com/bobxwu/FASTopic . | cs.CL | FASTopic: A Fast, Adaptive, Stable, and Transferable Topic Modeling Paradigm |
2024-05-28T09:06:18Z | http://arxiv.org/pdf/2405.17977v1 | http://arxiv.org/abs/2405.17977v1 | Seongyun Lee, Sue Hyun Park, Seungone Kim, Minjoon Seo | 2024-05-28T09:06:18Z | cs.CL | Although humans inherently have diverse values, current large language model
(LLM) alignment methods often assume that aligning LLMs with the general
public's preferences is optimal. A major challenge in adopting a more
individualized approach to LLM alignment is its lack of scalability, as it
involves repeatedly acquiring preference data and training new reward models
and LLMs for each individual's preferences. To address these challenges, we
propose a new paradigm where users specify what they value most within the
system message, steering the LLM's generation behavior to better align with the
user's intentions. However, a naive application of such an approach is
non-trivial since LLMs are typically trained on a uniform system message (e.g.,
"You are a helpful assistant") which limits their ability to generalize to
diverse, unseen system messages. To improve this generalization, we create the
Multifaceted Collection, a preference dataset with 192k combinations of values
beyond generic helpfulness and harmlessness, spanning 65k user instructions.
Using this dataset, we train a 7B LLM called Janus and test it on 921 prompts
from 5 benchmarks (AlpacaEval 2.0, FLASK, Koala, MT-Bench, and Self-Instruct)
by adding various unseen system messages that reflect user preferences. Janus
achieves tie+win rate of 75.2%, 72.4%, and 66.4% against Mistral 7B Instruct
v0.2, GPT-3.5 Turbo, and GPT-4, respectively. Unexpectedly, on three benchmarks
focused on response helpfulness (AlpacaEval 2.0, MT-Bench, Arena Hard Auto
v0.1), Janus also outperforms LLaMA 3 8B Instruct by a +4.0%, +0.1%, +3.0%
margin, underscoring that training with a vast array of system messages could
also enhance alignment to the general public's preference as well. Our code,
dataset, benchmark, and models are available at
https://github.com/kaistAI/Janus. | cs.CL | Aligning to Thousands of Preferences via System Message Generalization |
2024-05-28T09:05:08Z | http://arxiv.org/pdf/2405.17976v2 | http://arxiv.org/abs/2405.17976v2 | Shaohua Wu, Jiangang Luo, Xi Chen, Lingjun Li, Xudong Zhao, Tong Yu, Chao Wang, Yue Wang, Fei Wang, Weixu Qiao, Houbo He, Zeru Zhang, Zeyu Sun, Junxiong Mao, Chong Shen | 2024-05-29T07:19:58Z | cs.AI, cs.CL | Yuan 2.0-M32, with a similar base architecture as Yuan-2.0 2B, uses a
mixture-of-experts architecture with 32 experts of which 2 experts are active.
A new router network, Attention Router, is proposed and adopted for a more
efficient selection of experts, which improves the accuracy compared to the
model with classical router network. Yuan 2.0-M32 is trained with 2000B tokens
from scratch, and the training computation consumption is only 9.25% of a dense
model at the same parameter scale. Yuan 2.0-M32 demonstrates competitive
capability on coding, math, and various domains of expertise, with only 3.7B
active parameters of 40B in total, and 7.4 GFlops forward computation per
token, both of which are only 1/19 of Llama3-70B. Yuan 2.0-M32 surpass
Llama3-70B on MATH and ARC-Challenge benchmark, with accuracy of 55.89 and 95.8
respectively. The models and source codes of Yuan 2.0-M32 are released at
Github1. | cs.AI | Yuan 2.0-M32: Mixture of Experts with Attention Router |
2024-05-28T09:04:13Z | http://arxiv.org/pdf/2405.17974v1 | http://arxiv.org/abs/2405.17974v1 | Yi-Pei Chen, Noriki Nishida, Hideki Nakayama, Yuji Matsumoto | 2024-05-28T09:04:13Z | cs.CL, cs.AI | Enhancing user engagement through personalization in conversational agents
has gained significance, especially with the advent of large language models
that generate fluent responses. Personalized dialogue generation, however, is
multifaceted and varies in its definition -- ranging from instilling a persona
in the agent to capturing users' explicit and implicit cues. This paper seeks
to systemically survey the recent landscape of personalized dialogue
generation, including the datasets employed, methodologies developed, and
evaluation metrics applied. Covering 22 datasets, we highlight benchmark
datasets and newer ones enriched with additional features. We further analyze
17 seminal works from top conferences between 2021-2023 and identify five
distinct types of problems. We also shed light on recent progress by LLMs in
personalized dialogue generation. Our evaluation section offers a comprehensive
summary of assessment facets and metrics utilized in these works. In
conclusion, we discuss prevailing challenges and envision prospect directions
for future research in personalized dialogue generation. | cs.CL | Recent Trends in Personalized Dialogue Generation: A Review of Datasets, Methodologies, and Evaluations |
2024-05-28T08:56:33Z | http://arxiv.org/pdf/2405.17969v1 | http://arxiv.org/abs/2405.17969v1 | Yunzhi Yao, Ningyu Zhang, Zekun Xi, Mengru Wang, Ziwen Xu, Shumin Deng, Huajun Chen | 2024-05-28T08:56:33Z | cs.CL, cs.AI, cs.CV, cs.IR, cs.LG | The remarkable capabilities of modern large language models are rooted in
their vast repositories of knowledge encoded within their parameters, enabling
them to perceive the world and engage in reasoning. The inner workings of how
these models store knowledge have long been a subject of intense interest and
investigation among researchers. To date, most studies have concentrated on
isolated components within these models, such as the Multilayer Perceptrons and
attention head. In this paper, we delve into the computation graph of the
language model to uncover the knowledge circuits that are instrumental in
articulating specific knowledge. The experiments, conducted with GPT2 and
TinyLLAMA, has allowed us to observe how certain information heads, relation
heads, and Multilayer Perceptrons collaboratively encode knowledge within the
model. Moreover, we evaluate the impact of current knowledge editing techniques
on these knowledge circuits, providing deeper insights into the functioning and
constraints of these editing methodologies. Finally, we utilize knowledge
circuits to analyze and interpret language model behaviors such as
hallucinations and in-context learning. We believe the knowledge circuit holds
potential for advancing our understanding of Transformers and guiding the
improved design of knowledge editing. Code and data are available in
https://github.com/zjunlp/KnowledgeCircuits. | cs.CL | Knowledge Circuits in Pretrained Transformers |
2024-05-28T08:48:08Z | http://arxiv.org/pdf/2405.17964v1 | http://arxiv.org/abs/2405.17964v1 | Teodor-George Marchitan, Claudiu Creanga, Liviu P. Dinu | 2024-05-28T08:48:08Z | cs.CL | This paper describes the approach of the UniBuc - NLP team in tackling the
SemEval 2024 Task 8: Multigenerator, Multidomain, and Multilingual Black-Box
Machine-Generated Text Detection. We explored transformer-based and hybrid deep
learning architectures. For subtask B, our transformer-based model achieved a
strong \textbf{second-place} out of $77$ teams with an accuracy of
\textbf{86.95\%}, demonstrating the architecture's suitability for this task.
However, our models showed overfitting in subtask A which could potentially be
fixed with less fine-tunning and increasing maximum sequence length. For
subtask C (token-level classification), our hybrid model overfit during
training, hindering its ability to detect transitions between human and
machine-generated text. | cs.CL | Transformer and Hybrid Deep Learning Based Models for Machine-Generated Text Detection |
2024-05-28T08:39:49Z | http://arxiv.org/pdf/2405.17957v1 | http://arxiv.org/abs/2405.17957v1 | Xiaobao Wu, Xinshuai Dong, Liangming Pan, Thong Nguyen, Anh Tuan Luu | 2024-05-28T08:39:49Z | cs.CL, cs.AI | Dynamic topic models track the evolution of topics in sequential documents,
which have derived various applications like trend analysis and opinion mining.
However, existing models suffer from repetitive topic and unassociated topic
issues, failing to reveal the evolution and hindering further applications. To
address these issues, we break the tradition of simply chaining topics in
existing work and propose a novel neural \modelfullname. We introduce a new
evolution-tracking contrastive learning method that builds the similarity
relations among dynamic topics. This not only tracks topic evolution but also
maintains topic diversity, mitigating the repetitive topic issue. To avoid
unassociated topics, we further present an unassociated word exclusion method
that consistently excludes unassociated words from discovered topics. Extensive
experiments demonstrate our model significantly outperforms state-of-the-art
baselines, tracking topic evolution with high-quality topics, showing better
performance on downstream tasks, and remaining robust to the hyperparameter for
evolution intensities. Our code is available at https://github.com/bobxwu/CFDTM . | cs.CL | Modeling Dynamic Topics in Chain-Free Fashion by Evolution-Tracking Contrastive Learning and Unassociated Word Exclusion |
2024-05-28T08:01:26Z | http://arxiv.org/pdf/2405.17935v2 | http://arxiv.org/abs/2405.17935v2 | Changle Qu, Sunhao Dai, Xiaochi Wei, Hengyi Cai, Shuaiqiang Wang, Dawei Yin, Jun Xu, Ji-Rong Wen | 2024-05-30T11:01:10Z | cs.CL, cs.AI | Recently, tool learning with large language models (LLMs) has emerged as a
promising paradigm for augmenting the capabilities of LLMs to tackle highly
complex problems. Despite growing attention and rapid advancements in this
field, the existing literature remains fragmented and lacks systematic
organization, posing barriers to entry for newcomers. This gap motivates us to
conduct a comprehensive survey of existing works on tool learning with LLMs. In
this survey, we focus on reviewing existing literature from the two primary
aspects (1) why tool learning is beneficial and (2) how tool learning is
implemented, enabling a comprehensive understanding of tool learning with LLMs.
We first explore the "why" by reviewing both the benefits of tool integration
and the inherent benefits of the tool learning paradigm from six specific
aspects. In terms of "how", we systematically review the literature according
to a taxonomy of four key stages in the tool learning workflow: task planning,
tool selection, tool calling, and response generation. Additionally, we provide
a detailed summary of existing benchmarks and evaluation methods, categorizing
them according to their relevance to different stages. Finally, we discuss
current challenges and outline potential future directions, aiming to inspire
both researchers and industrial developers to further explore this emerging and
promising area. We also maintain a GitHub repository to continually keep track
of the relevant papers and resources in this rising area at
\url{https://github.com/quchangle1/LLM-Tool-Survey}. | cs.CL | Tool Learning with Large Language Models: A Survey |
2024-05-28T07:53:40Z | http://arxiv.org/pdf/2405.17931v1 | http://arxiv.org/abs/2405.17931v1 | Keming Lu, Bowen Yu, Fei Huang, Yang Fan, Runji Lin, Chang Zhou | 2024-05-28T07:53:40Z | cs.CL, cs.LG | Effectively aligning Large Language Models (LLMs) with human-centric values
while preventing the degradation of abilities acquired through Pre-training and
Supervised Fine-tuning (SFT) poses a central challenge in Reinforcement
Learning from Human Feedback (RLHF). In this paper, we first discover that
interpolating RLHF and SFT model parameters can adjust the trade-off between
human preference and basic capabilities, thereby reducing the alignment tax at
the cost of alignment reward. Inspired by this, we propose integrating the RL
policy and SFT models at each optimization step in RLHF to continuously
regulate the training direction, introducing the Online Merging Optimizer.
Specifically, we merge gradients with the parameter differences between SFT and
pretrained models, effectively steering the gradient towards maximizing rewards
in the direction of SFT optimization. We demonstrate that our optimizer works
well with different LLM families, such as Qwen and LLaMA, across various model
sizes ranging from 1.8B to 8B, various RLHF algorithms like DPO and KTO, and
existing model merging methods. It significantly enhances alignment reward
while mitigating alignment tax, achieving higher overall performance across 14
benchmarks. | cs.CL | Online Merging Optimizers for Boosting Rewards and Mitigating Tax in Alignment |
2024-05-28T07:48:15Z | http://arxiv.org/pdf/2405.17927v1 | http://arxiv.org/abs/2405.17927v1 | Shakti N. Wadekar, Abhishek Chaurasia, Aman Chadha, Eugenio Culurciello | 2024-05-28T07:48:15Z | cs.AI, cs.CL, cs.CV, cs.LG, eess.AS | This work uniquely identifies and characterizes four prevalent multimodal
model architectural patterns in the contemporary multimodal landscape.
Systematically categorizing models by architecture type facilitates monitoring
of developments in the multimodal domain. Distinct from recent survey papers
that present general information on multimodal architectures, this research
conducts a comprehensive exploration of architectural details and identifies
four specific architectural types. The types are distinguished by their
respective methodologies for integrating multimodal inputs into the deep neural
network model. The first two types (Type A and B) deeply fuses multimodal
inputs within the internal layers of the model, whereas the following two types
(Type C and D) facilitate early fusion at the input stage. Type-A employs
standard cross-attention, whereas Type-B utilizes custom-designed layers for
modality fusion within the internal layers. On the other hand, Type-C utilizes
modality-specific encoders, while Type-D leverages tokenizers to process the
modalities at the model's input stage. The identified architecture types aid
the monitoring of any-to-any multimodal model development. Notably, Type-C and
Type-D are currently favored in the construction of any-to-any multimodal
models. Type-C, distinguished by its non-tokenizing multimodal model
architecture, is emerging as a viable alternative to Type-D, which utilizes
input-tokenizing techniques. To assist in model selection, this work highlights
the advantages and disadvantages of each architecture type based on data and
compute requirements, architecture complexity, scalability, simplification of
adding modalities, training objectives, and any-to-any multimodal generation
capability. | cs.AI | The Evolution of Multimodal Model Architectures |
2024-05-28T07:36:56Z | http://arxiv.org/pdf/2405.17915v1 | http://arxiv.org/abs/2405.17915v1 | Longze Chen, Ziqiang Liu, Wanwei He, Yunshui Li, Run Luo, Min Yang | 2024-05-28T07:36:56Z | cs.CL | Long-context modeling capabilities are important for large language models
(LLMs) in various applications. However, directly training LLMs with long
context windows is insufficient to enhance this capability since some training
samples do not exhibit strong semantic dependencies across long contexts. In
this study, we propose a data mining framework \textbf{ProLong} that can assign
each training sample with a long dependency score, which can be used to rank
and filter samples that are more advantageous for enhancing long-context
modeling abilities in LLM training. Specifically, we first use delta perplexity
scores to measure the \textit{Dependency Strength} between text segments in a
given document. Then we refine this metric based on the \textit{Dependency
Distance} of these segments to incorporate spatial relationships across
long-contexts. Final results are calibrated with a \textit{Dependency
Specificity} metric to prevent trivial dependencies introduced by repetitive
patterns. Moreover, a random sampling approach is proposed to optimize the
computational efficiency of ProLong. Comprehensive experiments on multiple
benchmarks indicate that ProLong effectively identifies documents that carry
long dependencies and LLMs trained on these documents exhibit significantly
enhanced long-context modeling capabilities. | cs.CL | Long Context is Not Long at All: A Prospector of Long-Dependency Data for Large Language Models |
2024-05-28T07:24:20Z | http://arxiv.org/pdf/2405.17902v2 | http://arxiv.org/abs/2405.17902v2 | Yaoyao Xu, Xinjian Zhao, Xiaozhuang Song, Benyou Wang, Tianshu Yu | 2024-06-29T07:07:49Z | cs.AI, cs.CL, cs.LG | We introduce a pioneering methodology for boosting large language models in
the domain of protein representation learning. Our primary contribution lies in
the refinement process for correlating the over-reliance on co-evolution
knowledge, in a way that networks are trained to distill invaluable insights
from negative samples, constituted by protein pairs sourced from disparate
categories. By capitalizing on this novel approach, our technique steers the
training of transformer-based models within the attention score space. This
advanced strategy not only amplifies performance but also reflects the nuanced
biological behaviors exhibited by proteins, offering aligned evidence with
traditional biological mechanisms such as protein-protein interaction. We
experimentally observed improved performance on various tasks over datasets, on
top of several well-established large protein models. This innovative paradigm
opens up promising horizons for further progress in the realms of protein
research and computational biology. | cs.AI | Boosting Protein Language Models with Negative Sample Mining |
2024-05-28T07:22:30Z | http://arxiv.org/pdf/2405.17900v1 | http://arxiv.org/abs/2405.17900v1 | Haoxiang Shi, Xulong Zhang, Ning Cheng, Yong Zhang, Jun Yu, Jing Xiao, Jianzong Wang | 2024-05-28T07:22:30Z | cs.CL | The purpose of emotion recognition in conversation (ERC) is to identify the
emotion category of an utterance based on contextual information. Previous ERC
methods relied on simple connections for cross-modal fusion and ignored the
information differences between modalities, resulting in the model being unable
to focus on modality-specific emotional information. At the same time, the
shared information between modalities was not processed to generate emotions.
Information redundancy problem. To overcome these limitations, we propose a
cross-modal fusion emotion prediction network based on vector connections. The
network mainly includes two stages: the multi-modal feature fusion stage based
on connection vectors and the emotion classification stage based on fused
features. Furthermore, we design a supervised inter-class contrastive learning
module based on emotion labels. Experimental results confirm the effectiveness
of the proposed method, demonstrating excellent performance on the IEMOCAP and
MELD datasets. | cs.CL | Enhancing Emotion Recognition in Conversation through Emotional Cross-Modal Fusion and Inter-class Contrastive Learning |
2024-05-28T07:13:25Z | http://arxiv.org/pdf/2405.17893v1 | http://arxiv.org/abs/2405.17893v1 | Xiaocheng Yang, Bingsen Chen, Yik-Cheung Tam | 2024-05-28T07:13:25Z | cs.CL, cs.AI | Instructing large language models (LLMs) to solve elementary school math
problems has shown great success using Chain of Thought (CoT). However, the CoT
approach relies on an LLM to generate a sequence of arithmetic calculations
which can be prone to cascaded calculation errors. We hypothesize that an LLM
should focus on extracting predicates and generating symbolic formulas from the
math problem description so that the underlying calculation can be done via an
external code interpreter. We investigate using LLM to generate Prolog programs
to solve mathematical questions. Experimental results show that our
Prolog-based arithmetic problem-solving outperforms CoT generation in the GSM8K
benchmark across three distinct LLMs. In addition, given the insensitive
ordering of predicates and symbolic formulas in Prolog, we propose to permute
the ground truth predicates for more robust LLM training via data augmentation. | cs.CL | Arithmetic Reasoning with LLM: Prolog Generation & Permutation |
2024-05-28T07:12:06Z | http://arxiv.org/pdf/2405.17890v1 | http://arxiv.org/abs/2405.17890v1 | Wujiang Xu, Zujie Liang, Jiaojiao Han, Xuying Ning, Wenfang Lin, Linxun Chen, Feng Wei, Yongfeng Zhang | 2024-05-28T07:12:06Z | cs.IR, cs.CL, cs.LG | The sequential Recommendation (SR) task involves predicting the next item a
user is likely to interact with, given their past interactions. The SR models
examine the sequence of a user's actions to discern more complex behavioral
patterns and temporal dynamics. Recent research demonstrates the great impact
of LLMs on sequential recommendation systems, either viewing sequential
recommendation as language modeling or serving as the backbone for user
representation. Although these methods deliver outstanding performance, there
is scant evidence of the necessity of a large language model and how large the
language model is needed, especially in the sequential recommendation scene.
Meanwhile, due to the huge size of LLMs, it is inefficient and impractical to
apply a LLM-based model in real-world platforms that often need to process
billions of traffic logs daily. In this paper, we explore the influence of
LLMs' depth by conducting extensive experiments on large-scale industry
datasets. Surprisingly, we discover that most intermediate layers of LLMs are
redundant. Motivated by this insight, we empower small language models for SR,
namely SLMRec, which adopt a simple yet effective knowledge distillation
method. Moreover, SLMRec is orthogonal to other post-training efficiency
techniques, such as quantization and pruning, so that they can be leveraged in
combination. Comprehensive experimental results illustrate that the proposed
SLMRec model attains the best performance using only 13% of the parameters
found in LLM-based recommendation models, while simultaneously achieving up to
6.6x and 8.0x speedups in training and inference time costs, respectively. | cs.IR | SLMRec: Empowering Small Language Models for Sequential Recommendation |
2024-05-28T06:44:13Z | http://arxiv.org/pdf/2405.17871v1 | http://arxiv.org/abs/2405.17871v1 | Xin Xiao, Bohong Wu, Jiacong Wang, Chunyuan Li, Xun Zhou, Haoyuan Guo | 2024-05-28T06:44:13Z | cs.CV, cs.AI, cs.CL | Existing image-text modality alignment in Vision Language Models (VLMs)
treats each text token equally in an autoregressive manner. Despite being
simple and effective, this method results in sub-optimal cross-modal alignment
by over-emphasizing the text tokens that are less correlated with or even
contradictory with the input images. In this paper, we advocate for assigning
distinct contributions for each text token based on its visual correlation.
Specifically, we present by contrasting image inputs, the difference in
prediction logits on each text token provides strong guidance of visual
correlation. We therefore introduce Contrastive ALignment (CAL), a simple yet
effective re-weighting strategy that prioritizes training visually correlated
tokens. Our experimental results demonstrate that CAL consistently improves
different types of VLMs across different resolutions and model sizes on various
benchmark datasets. Importantly, our method incurs minimal additional
computational overhead, rendering it highly efficient compared to alternative
data scaling strategies. Codes are available at
https://github.com/foundation-multimodal-models/CAL. | cs.CV | Seeing the Image: Prioritizing Visual Correlation by Contrastive Alignment |
2024-05-28T05:56:11Z | http://arxiv.org/pdf/2405.17849v2 | http://arxiv.org/abs/2405.17849v2 | Xing Hu, Yuan Cheng, Dawei Yang, Zhihang Yuan, Jiangyong Yu, Chen Xu, Sifan Zhou | 2024-06-05T15:26:58Z | cs.LG, cs.AI, cs.CL | Post-training quantization (PTQ) serves as a potent technique to accelerate
the inference of large language models (LLMs). Nonetheless, existing works
still necessitate a considerable number of floating-point (FP) operations
during inference, including additional quantization and de-quantization, as
well as non-linear operators such as RMSNorm and Softmax. This limitation
hinders the deployment of LLMs on the edge and cloud devices. In this paper, we
identify the primary obstacle to integer-only quantization for LLMs lies in the
large fluctuation of activations across channels and tokens in both linear and
non-linear operations. To address this issue, we propose I-LLM, a novel
integer-only fully-quantized PTQ framework tailored for LLMs. Specifically, (1)
we develop Fully-Smooth Block-Reconstruction (FSBR) to aggressively smooth
inter-channel variations of all activations and weights. (2) to alleviate
degradation caused by inter-token variations, we introduce a novel approach
called Dynamic Integer-only MatMul (DI-MatMul). This method enables dynamic
quantization in full-integer matrix multiplication by dynamically quantizing
the input and outputs with integer-only operations. (3) we design
DI-ClippedSoftmax, DI-Exp, and DI-Normalization, which utilize bit shift to
execute non-linear operators efficiently while maintaining accuracy. The
experiment shows that our I-LLM achieves comparable accuracy to the FP baseline
and outperforms non-integer quantization methods. For example, I-LLM can
operate at W4A4 with negligible loss of accuracy. To our knowledge, we are the
first to bridge the gap between integer-only quantization and LLMs. We've
published our code on anonymous.4open.science, aiming to contribute to the
advancement of this field. | cs.LG | I-LLM: Efficient Integer-Only Inference for Fully-Quantized Low-Bit Large Language Models |
2024-05-28T05:33:13Z | http://arxiv.org/pdf/2405.17840v2 | http://arxiv.org/abs/2405.17840v2 | Andrew H. Lee, Sina J. Semnani, Galo Castillo-López, Gäel de Chalendar, Monojit Choudhury, Ashna Dua, Kapil Rajesh Kavitha, Sungkyun Kim, Prashant Kodali, Ponnurangam Kumaraguru, Alexis Lombard, Mehrad Moradshahi, Gihyun Park, Nasredine Semmar, Jiwon Seo, Tianhao Shen, Manish Shrivastava, Deyi Xiong, Monica S. Lam | 2024-06-16T20:45:23Z | cs.CL | Creating multilingual task-oriented dialogue (TOD) agents is challenging due
to the high cost of training data acquisition. Following the research trend of
improving training data efficiency, we show for the first time, that in-context
learning is sufficient to tackle multilingual TOD.
To handle the challenging dialogue state tracking (DST) subtask, we break it
down to simpler steps that are more compatible with in-context learning where
only a handful of few-shot examples are used. We test our approach on the
multilingual TOD dataset X-RiSAWOZ, which has 12 domains in Chinese, English,
French, Korean, Hindi, and code-mixed Hindi-English. Our turn-by-turn DST
accuracy on the 6 languages range from 55.6% to 80.3%, seemingly worse than the
SOTA results from fine-tuned models that achieve from 60.7% to 82.8%; our BLEU
scores in the response generation (RG) subtask are also significantly lower
than SOTA.
However, after manual evaluation of the validation set, we find that by
correcting gold label errors and improving dataset annotation schema, GPT-4
with our prompts can achieve (1) 89.6%-96.8% accuracy in DST, and (2) more than
99% correct response generation across different languages. This leads us to
conclude that current automatic metrics heavily underestimate the effectiveness
of in-context learning. | cs.CL | Benchmarks Underestimate the Readiness of Multi-lingual Dialogue Agents |
2024-05-28T05:10:40Z | http://arxiv.org/pdf/2406.00045v1 | http://arxiv.org/abs/2406.00045v1 | Yuanpu Cao, Tianrong Zhang, Bochuan Cao, Ziyi Yin, Lu Lin, Fenglong Ma, Jinghui Chen | 2024-05-28T05:10:40Z | cs.CL, cs.LG | Researchers have been studying approaches to steer the behavior of Large
Language Models (LLMs) and build personalized LLMs tailored for various
applications. While fine-tuning seems to be a direct solution, it requires
substantial computational resources and may significantly affect the utility of
the original LLM. Recent endeavors have introduced more lightweight strategies,
focusing on extracting "steering vectors" to guide the model's output toward
desired behaviors by adjusting activations within specific layers of the LLM's
transformer architecture. However, such steering vectors are directly extracted
from the activations of human preference data and thus often lead to suboptimal
results and occasional failures, especially in alignment-related scenarios.
This work proposes an innovative approach that could produce more effective
steering vectors through bi-directional preference optimization. Our method is
designed to allow steering vectors to directly influence the generation
probability of contrastive human preference data pairs, thereby offering a more
precise representation of the target behavior. By carefully adjusting the
direction and magnitude of the steering vector, we enabled personalized control
over the desired behavior across a spectrum of intensities. Extensive
experimentation across various open-ended generation tasks, particularly
focusing on steering AI personas, has validated the efficacy of our approach.
Moreover, we comprehensively investigate critical alignment-concerning
scenarios, such as managing truthfulness, mitigating hallucination, and
addressing jailbreaking attacks. Remarkably, our method can still demonstrate
outstanding steering effectiveness across these scenarios. Furthermore, we
showcase the transferability of our steering vectors across different
models/LoRAs and highlight the synergistic benefits of applying multiple
vectors simultaneously. | cs.CL | Personalized Steering of Large Language Models: Versatile Steering Vectors Through Bi-directional Preference Optimization |
2024-05-28T05:00:12Z | http://arxiv.org/pdf/2405.17830v1 | http://arxiv.org/abs/2405.17830v1 | Chengyuan Liu, Shihang Wang, Yangyang Kang, Lizhi Qing, Fubang Zhao, Changlong Sun, Kun Kuang, Fei Wu | 2024-05-28T05:00:12Z | cs.CL | The performance on general tasks decreases after Large Language Models (LLMs)
are fine-tuned on domain-specific tasks, the phenomenon is known as
Catastrophic Forgetting (CF). However, this paper presents a further challenge
for real application of domain-specific LLMs beyond CF, called General
Capabilities Integration (GCI), which necessitates the integration of both the
general capabilities and domain knowledge within a single instance. The
objective of GCI is not merely to retain previously acquired general
capabilities alongside new domain knowledge, but to harmonize and utilize both
sets of skills in a cohesive manner to enhance performance on domain-specific
tasks. Taking legal domain as an example, we carefully design three groups of
training and testing tasks without lacking practicability, and construct the
corresponding datasets. To better incorporate general capabilities across
domain-specific scenarios, we introduce ALoRA, which utilizes a multi-head
attention module upon LoRA, facilitating direct information transfer from
preceding tokens to the current one. This enhancement permits the
representation to dynamically switch between domain-specific knowledge and
general competencies according to the attention. Extensive experiments are
conducted on the proposed tasks. The results exhibit the significance of our
setting, and the effectiveness of our method. | cs.CL | More Than Catastrophic Forgetting: Integrating General Capabilities For Domain-Specific LLMs |
2024-05-28T04:46:52Z | http://arxiv.org/pdf/2405.17822v1 | http://arxiv.org/abs/2405.17822v1 | Zhenyu Pan, Haozheng Luo, Manling Li, Han Liu | 2024-05-28T04:46:52Z | cs.CL, cs.AI | We present a Conversational Chain-of-Action (Conv-CoA) framework for
Open-domain Conversational Question Answering (OCQA). Compared with literature,
Conv-CoA addresses three major challenges: (i) unfaithful hallucination that is
inconsistent with real-time or domain facts, (ii) weak reasoning performance in
conversational scenarios, and (iii) unsatisfying performance in conversational
information retrieval. Our key contribution is a dynamic reasoning-retrieval
mechanism that extracts the intent of the question and decomposes it into a
reasoning chain to be solved via systematic prompting, pre-designed actions,
updating the Contextual Knowledge Set (CKS), and a novel Hopfield-based
retriever. Methodologically, we propose a resource-efficiency Hopfield
retriever to enhance the efficiency and accuracy of conversational information
retrieval within our actions. Additionally, we propose a
conversational-multi-reference faith score (Conv-MRFS) to verify and resolve
conflicts between retrieved knowledge and answers in conversations.
Empirically, we conduct comparisons between our framework and 23
state-of-the-art methods across five different research directions and two
public benchmarks. These comparisons demonstrate that our Conv-CoA outperforms
other methods in both the accuracy and efficiency dimensions. | cs.CL | Conv-CoA: Improving Open-domain Question Answering in Large Language Models via Conversational Chain-of-Action |
2024-05-28T04:36:15Z | http://arxiv.org/pdf/2406.03198v1 | http://arxiv.org/abs/2406.03198v1 | Jacy Anthis, Kristian Lum, Michael Ekstrand, Avi Feller, Alexander D'Amour, Chenhao Tan | 2024-05-28T04:36:15Z | cs.CL, cs.HC, cs.LG, stat.AP, stat.ML | The need for fair AI is increasingly clear in the era of general-purpose
systems such as ChatGPT, Gemini, and other large language models (LLMs).
However, the increasing complexity of human-AI interaction and its social
impacts have raised questions of how fairness standards could be applied. Here,
we review the technical frameworks that machine learning researchers have used
to evaluate fairness, such as group fairness and fair representations, and find
that their application to LLMs faces inherent limitations. We show that each
framework either does not logically extend to LLMs or presents a notion of
fairness that is intractable for LLMs, primarily due to the multitudes of
populations affected, sensitive attributes, and use cases. To address these
challenges, we develop guidelines for the more realistic goal of achieving
fairness in particular use cases: the criticality of context, the
responsibility of LLM developers, and the need for stakeholder participation in
an iterative process of design and evaluation. Moreover, it may eventually be
possible and even necessary to use the general-purpose capabilities of AI
systems to address fairness challenges as a form of scalable AI-assisted
alignment. | cs.CL | The Impossibility of Fair LLMs |
2024-05-28T04:22:28Z | http://arxiv.org/pdf/2406.01609v1 | http://arxiv.org/abs/2406.01609v1 | Akshat Mohan Dasula, Hrushitha Tigulla, Preethika Bhukya | 2024-05-28T04:22:28Z | cs.IR, cs.CL, cs.LG | Traditionally in the domain of legal research, the retrieval of pertinent
citations from intricate case descriptions has demanded manual effort and
keyword-based search applications that mandate expertise in understanding legal
jargon. Legal case descriptions hold pivotal information for legal
professionals and researchers, necessitating more efficient and automated
approaches. We propose a methodology that combines natural language processing
(NLP) and machine learning techniques to enhance the organization and
utilization of legal case descriptions. This approach revolves around the
creation of textual embeddings with the help of state-of-art embedding models.
Our methodology addresses two primary objectives: unsupervised clustering and
supervised citation retrieval, both designed to automate the citation
extraction process. Although the proposed methodology can be used for any
dataset, we employed the Supreme Court of The United States (SCOTUS) dataset,
yielding remarkable results. Our methodology achieved an impressive accuracy
rate of 90.9%. By automating labor-intensive processes, we pave the way for a
more efficient, time-saving, and accessible landscape in legal research,
benefiting legal professionals, academics, and researchers. | cs.IR | Judgement Citation Retrieval using Contextual Similarity |
2024-05-28T04:11:37Z | http://arxiv.org/pdf/2405.17809v1 | http://arxiv.org/abs/2405.17809v1 | Chenyang Le, Yao Qian, Dongmei Wang, Long Zhou, Shujie Liu, Xiaofei Wang, Midia Yousefi, Yanmin Qian, Jinyu Li, Sheng Zhao, Michael Zeng | 2024-05-28T04:11:37Z | cs.CL, cs.AI, cs.SD, eess.AS | There is a rising interest and trend in research towards directly translating
speech from one language to another, known as end-to-end speech-to-speech
translation. However, most end-to-end models struggle to outperform cascade
models, i.e., a pipeline framework by concatenating speech recognition, machine
translation and text-to-speech models. The primary challenges stem from the
inherent complexities involved in direct translation tasks and the scarcity of
data. In this study, we introduce a novel model framework TransVIP that
leverages diverse datasets in a cascade fashion yet facilitates end-to-end
inference through joint probability. Furthermore, we propose two separated
encoders to preserve the speaker's voice characteristics and isochrony from the
source speech during the translation process, making it highly suitable for
scenarios such as video dubbing. Our experiments on the French-English language
pair demonstrate that our model outperforms the current state-of-the-art
speech-to-speech translation model. | cs.CL | TransVIP: Speech to Speech Translation System with Voice and Isochrony Preservation |
2024-05-28T04:04:40Z | http://arxiv.org/pdf/2405.17804v1 | http://arxiv.org/abs/2405.17804v1 | Wei Li, Houfeng Wang | 2024-05-28T04:04:40Z | cs.CL | Grammatical error correction (GEC) is a task dedicated to rectifying texts
with minimal edits, which can be decoupled into two components: detection and
correction. However, previous works have predominantly focused on direct
correction, with no prior efforts to integrate both into a single model.
Moreover, the exploration of the detection-correction paradigm by large
language models (LLMs) remains underdeveloped. This paper introduces an
integrated detection-correction structure, named DeCoGLM, based on the General
Language Model (GLM). The detection phase employs a fault-tolerant detection
template, while the correction phase leverages autoregressive mask infilling
for localized error correction. Through the strategic organization of input
tokens and modification of attention masks, we facilitate multi-task learning
within a single model. Our model demonstrates competitive performance against
the state-of-the-art models on English and Chinese GEC datasets. Further
experiments present the effectiveness of the detection-correction structure in
LLMs, suggesting a promising direction for GEC. | cs.CL | Detection-Correction Structure via General Language Model for Grammatical Error Correction |
2024-05-28T03:49:54Z | http://arxiv.org/pdf/2405.17799v1 | http://arxiv.org/abs/2405.17799v1 | Yudong Wang, Damai Dai, Zhifang Sui | 2024-05-28T03:49:54Z | cs.LG, cs.CL | Most work treats large language models as black boxes without in-depth
understanding of their internal working mechanism. In order to explain the
internal representations of LLMs, we propose a gradient-based metric to assess
the activation level of model parameters. Based on this metric, we obtain three
preliminary findings. (1) When the inputs are in the same domain, parameters in
the shallow layers will be activated densely, which means a larger portion of
parameters will have great impacts on the outputs. In contrast, parameters in
the deep layers are activated sparsely. (2) When the inputs are across
different domains, parameters in shallow layers exhibit higher similarity in
the activation behavior than deep layers. (3) In deep layers, the similarity of
the distributions of activated parameters is positively correlated to the
empirical data relevance. Further, we develop three validation experiments to
solidify these findings. (1) Firstly, starting from the first finding, we
attempt to configure different prune ratios for different layers, and find this
method can benefit model pruning. (2) Secondly, we find that a pruned model
based on one calibration set can better handle tasks related to the calibration
task than those not related, which validate the second finding. (3) Thirdly,
Based on the STS-B and SICK benchmark, we find that two sentences with
consistent semantics tend to share similar parameter activation patterns in
deep layers, which aligns with our third finding. Our work sheds light on the
behavior of parameter activation in LLMs, and we hope these findings will have
the potential to inspire more practical applications. | cs.LG | Exploring Activation Patterns of Parameters in Language Models |
2024-05-28T02:46:11Z | http://arxiv.org/pdf/2405.17767v1 | http://arxiv.org/abs/2405.17767v1 | Robert Wu, Vardan Papyan | 2024-05-28T02:46:11Z | cs.LG, cs.CL, stat.ML, 68T07 (Primary) 68T50 (Secondary), I.2.6; I.2.7 | Neural collapse ($\mathcal{NC}$) is a phenomenon observed in classification
tasks where top-layer representations collapse into their class means, which
become equinorm, equiangular and aligned with the classifiers. These behaviors
-- associated with generalization and robustness -- would manifest under
specific conditions: models are trained towards zero loss, with noise-free
labels belonging to balanced classes, which do not outnumber the model's hidden
dimension. Recent studies have explored $\mathcal{NC}$ in the absence of one or
more of these conditions to extend and capitalize on the associated benefits of
ideal geometries. Language modeling presents a curious frontier, as
\textit{training by token prediction} constitutes a classification task where
none of the conditions exist: the vocabulary is imbalanced and exceeds the
embedding dimension; different tokens might correspond to similar contextual
embeddings; and large language models (LLMs) in particular are typically only
trained for a few epochs. This paper empirically investigates the impact of
scaling the architectures and training of causal language models (CLMs) on
their progression towards $\mathcal{NC}$. We find that $\mathcal{NC}$
properties that develop with scaling are linked to generalization. Moreover,
there is evidence of some relationship between $\mathcal{NC}$ and
generalization independent of scale. Our work therefore underscores the
generality of $\mathcal{NC}$ as it extends to the novel and more challenging
setting of language modeling. Downstream, we seek to inspire further research
on the phenomenon to deepen our understanding of LLMs -- and neural networks at
large -- and improve existing architectures based on $\mathcal{NC}$-related
properties. | cs.LG | Linguistic Collapse: Neural Collapse in (Large) Language Models |
2024-05-28T02:33:38Z | http://arxiv.org/pdf/2405.17764v2 | http://arxiv.org/abs/2405.17764v2 | Tianhao Zhang, Zhexiao Lin, Zhecheng Sheng, Chen Jiang, Dongyeop Kang | 2024-06-15T04:08:53Z | cs.CL, cs.AI, math.ST, stat.TH | Modeling and analyzing long sequences of text is an essential task for
Natural Language Processing. Success in capturing long text dynamics using
neural language models will facilitate many downstream tasks such as coherence
evaluation, text generation, machine translation and so on. This paper presents
a novel approach to model sequences through a stochastic process. We introduce
a likelihood-based training objective for the text encoder and design a more
thorough measurement (score) for long text evaluation compared to the previous
approach. The proposed training objective effectively preserves the sequence
coherence, while the new score comprehensively captures both temporal and
spatial dependencies. Theoretical properties of our new score show its
advantages in sequence evaluation. Experimental results show superior
performance in various sequence evaluation tasks, including global and local
discrimination within and between documents of different lengths. We also
demonstrate the encoder achieves competitive results on discriminating human
and AI written text. | cs.CL | On the Sequence Evaluation based on Stochastic Processes |
2024-05-28T02:12:35Z | http://arxiv.org/pdf/2405.17755v1 | http://arxiv.org/abs/2405.17755v1 | Shengnan Wang, Youhui Bai, Lin Zhang, Pingyi Zhou, Shixiong Zhao, Gong Zhang, Sen Wang, Renhai Chen, Hua Xu, Hongwei Sun | 2024-05-28T02:12:35Z | cs.CL, cs.AI | Length generalization failure problem, namely the large language model (LLM)
fails to generalize to texts longer than its maximum training length, greatly
restricts the application of LLM in the scenarios with streaming long inputs.
To address this problem, the existing methods either require substantial costs
or introduce precision loss. In this paper, we empirically find that the
accuracy of the LLM's prediction is highly correlated to its certainty. Based
on this, we propose an efficient training free framework, named XL3M (it means
extra-long large language model), which enables the LLMs trained on short
sequences to reason extremely long sequence without any further training or
fine-tuning. Under the XL3M framework, the input context will be firstly
decomposed into multiple short sub-contexts, where each sub-context contains an
independent segment and a common ``question'' which is a few tokens from the
end of the original context. Then XL3M gives a method to measure the relevance
between each segment and the ``question'', and constructs a concise key context
by splicing all the relevant segments in chronological order. The key context
is further used instead of the original context to complete the inference task.
Evaluations on comprehensive benchmarks show the superiority of XL3M. Using our
framework, a Llama2-7B model is able to reason 20M long sequences on an 8-card
Huawei Ascend 910B NPU machine with 64GB memory per card. | cs.CL | XL3M: A Training-free Framework for LLM Length Extension Based on Segment-wise Inference |
2024-05-28T01:55:35Z | http://arxiv.org/pdf/2405.17743v2 | http://arxiv.org/abs/2405.17743v2 | Zhengyang Tang, Chenyu Huang, Xin Zheng, Shixi Hu, Zizhuo Wang, Dongdong Ge, Benyou Wang | 2024-05-30T02:12:05Z | cs.CL, cs.AI, cs.CE, cs.LG | Large Language Models (LLMs) have emerged as powerful tools for tackling
complex Operations Research (OR) problem by providing the capacity in
automating optimization modeling. However, current methodologies heavily rely
on prompt engineering (e.g., multi-agent cooperation) with proprietary LLMs,
raising data privacy concerns that could be prohibitive in industry
applications. To tackle this issue, we propose training open-source LLMs for
optimization modeling. We identify four critical requirements for the training
dataset of OR LLMs, design and implement OR-Instruct, a semi-automated process
for creating synthetic data tailored to specific requirements. We also
introduce the IndustryOR benchmark, the first industrial benchmark for testing
LLMs on solving real-world OR problems. We apply the data from OR-Instruct to
various open-source LLMs of 7b size (termed as ORLMs), resulting in a
significantly improved capability for optimization modeling. Our
best-performing ORLM achieves state-of-the-art performance on the NL4OPT, MAMO,
and IndustryOR benchmarks. Our code and data are available at
\url{https://github.com/Cardinal-Operations/ORLM}. | cs.CL | ORLM: Training Large Language Models for Optimization Modeling |
2024-05-28T01:53:16Z | http://arxiv.org/pdf/2405.17740v1 | http://arxiv.org/abs/2405.17740v1 | Srijata Maji, Moghis Fereidouni, Vinaik Chhetri, Umar Farooq, A. B. Siddique | 2024-05-28T01:53:16Z | cs.CL | Existing recommendation systems have focused on two paradigms: 1- historical
user-item interaction-based recommendations and 2- conversational
recommendations. Conversational recommendation systems facilitate natural
language dialogues between users and the system, allowing the system to solicit
users' explicit needs while enabling users to inquire about recommendations and
provide feedback. Due to substantial advancements in natural language
processing, conversational recommendation systems have gained prominence.
Existing conversational recommendation datasets have greatly facilitated
research in their respective domains. Despite the exponential growth in mobile
users and apps in recent years, research in conversational mobile app
recommender systems has faced substantial constraints. This limitation can
primarily be attributed to the lack of high-quality benchmark datasets
specifically tailored for mobile apps. To facilitate research for
conversational mobile app recommendations, we introduce MobileConvRec.
MobileConvRec simulates conversations by leveraging real user interactions with
mobile apps on the Google Play store, originally captured in large-scale mobile
app recommendation dataset MobileRec. The proposed conversational
recommendation dataset synergizes sequential user-item interactions, which
reflect implicit user preferences, with comprehensive multi-turn conversations
to effectively grasp explicit user needs. MobileConvRec consists of over 12K
multi-turn recommendation-related conversations spanning 45 app categories.
Moreover, MobileConvRec presents rich metadata for each app such as permissions
data, security and privacy-related information, and binary executables of apps,
among others. We demonstrate that MobileConvRec can serve as an excellent
testbed for conversational mobile app recommendation through a comparative
study of several pre-trained large language models. | cs.CL | MobileConvRec: A Conversational Dataset for Mobile Apps Recommendations |
2024-05-28T01:23:58Z | http://arxiv.org/pdf/2405.17732v2 | http://arxiv.org/abs/2405.17732v2 | Jiahuan Cao, Yongxin Shi, Dezhi Peng, Yang Liu, Lianwen Jin | 2024-05-30T11:32:05Z | cs.CL | Classical Chinese Understanding (CCU) holds significant value in preserving
and exploration of the outstanding traditional Chinese culture. Recently,
researchers have attempted to leverage the potential of Large Language Models
(LLMs) for CCU by capitalizing on their remarkable comprehension and semantic
capabilities. However, no comprehensive benchmark is available to assess the
CCU capabilities of LLMs. To fill this gap, this paper introduces C$^{3}$bench,
a Comprehensive Classical Chinese understanding benchmark, which comprises
50,000 text pairs for five primary CCU tasks, including classification,
retrieval, named entity recognition, punctuation, and translation. Furthermore,
the data in C$^{3}$bench originates from ten different domains, covering most
of the categories in classical Chinese. Leveraging the proposed C$^{3}$bench,
we extensively evaluate the quantitative performance of 15 representative LLMs
on all five CCU tasks. Our results not only establish a public leaderboard of
LLMs' CCU capabilities but also gain some findings. Specifically, existing LLMs
are struggle with CCU tasks and still inferior to supervised models.
Additionally, the results indicate that CCU is a task that requires special
attention. We believe this study could provide a standard benchmark,
comprehensive baselines, and valuable insights for the future advancement of
LLM-based CCU research. The evaluation pipeline and dataset are available at
\url{https://github.com/SCUT-DLVCLab/C3bench}. | cs.CL | C$^{3}$Bench: A Comprehensive Classical Chinese Understanding Benchmark for Large Language Models |
2024-05-28T01:15:21Z | http://arxiv.org/pdf/2405.18448v1 | http://arxiv.org/abs/2405.18448v1 | Boammani Aser Lompo, Thanh-Dung Le | 2024-05-28T01:15:21Z | cs.CL, eess.SP | This research aims to classify numerical values extracted from medical
documents across seven distinct physiological categories, employing
CamemBERT-bio. Previous studies suggested that transformer-based models might
not perform as well as traditional NLP models in such tasks. To enhance
CamemBERT-bio's performances, we introduce two main innovations: integrating
keyword embeddings into the model and adopting a number-agnostic strategy by
excluding all numerical data from the text. The implementation of label
embedding techniques refines the attention mechanisms, while the technique of
using a `numerical-blind' dataset aims to bolster context-centric learning.
Another key component of our research is determining the criticality of
extracted numerical data. To achieve this, we utilized a simple approach that
involves verifying if the value falls within the established standard ranges.
Our findings are encouraging, showing substantial improvements in the
effectiveness of CamemBERT-bio, surpassing conventional methods with an F1
score of 0.89. This represents an over 20\% increase over the 0.73 $F_1$ score
of traditional approaches and an over 9\% increase over the 0.82 $F_1$ score of
state-of-the-art approaches. All this was achieved despite using small and
imbalanced training datasets. | cs.CL | Multi-objective Representation for Numbers in Clinical Narratives Using CamemBERT-bio |
2024-05-28T00:08:29Z | http://arxiv.org/pdf/2405.17712v1 | http://arxiv.org/abs/2405.17712v1 | Ahatsham Hayat, Mohammad Rashedul Hasan | 2024-05-28T00:08:29Z | cs.CL, cs.LG | This paper introduces the Contextual Language model for Accurate Imputation
Method (CLAIM), a novel strategy that capitalizes on the expansive knowledge
and reasoning capabilities of pre-trained large language models (LLMs) to
address missing data challenges in tabular datasets. Unlike traditional
imputation methods, which predominantly rely on numerical estimations, CLAIM
utilizes contextually relevant natural language descriptors to fill missing
values. This approach transforms datasets into natural language contextualized
formats that are inherently more aligned with LLMs' capabilities, thereby
facilitating the dual use of LLMs: first, to generate missing value
descriptors, and then, to fine-tune the LLM on the enriched dataset for
improved performance in downstream tasks. Our evaluations across diverse
datasets and missingness patterns reveal CLAIM's superior performance over
existing imputation techniques. Furthermore, our investigation into the
effectiveness of context-specific versus generic descriptors for missing data
highlights the importance of contextual accuracy in enhancing LLM performance
for data imputation. The results underscore CLAIM's potential to markedly
improve the reliability and quality of data analysis and machine learning
models, offering a more nuanced and effective solution for handling missing
data. | cs.CL | CLAIM Your Data: Enhancing Imputation Accuracy with Contextual Large Language Models |
2024-05-28T00:02:38Z | http://arxiv.org/pdf/2406.00044v1 | http://arxiv.org/abs/2406.00044v1 | Xu Wang, Yuan Wu | 2024-05-28T00:02:38Z | cs.CL, cs.LG | Adversarial training has been instrumental in advancing multi-domain text
classification (MDTC). Traditionally, MDTC methods employ a shared-private
paradigm, with a shared feature extractor for domain-invariant knowledge and
individual private feature extractors for domain-specific knowledge. Despite
achieving state-of-the-art results, these methods grapple with the escalating
model parameters due to the continuous addition of new domains. To address this
challenge, we introduce the Stochastic Adversarial Network (SAN), which
innovatively models the parameters of the domain-specific feature extractor as
a multivariate Gaussian distribution, as opposed to a traditional weight
vector. This design allows for the generation of numerous domain-specific
feature extractors without a substantial increase in model parameters,
maintaining the model's size on par with that of a single domain-specific
extractor. Furthermore, our approach integrates domain label smoothing and
robust pseudo-label regularization to fortify the stability of adversarial
training and to refine feature discriminability, respectively. The performance
of our SAN, evaluated on two leading MDTC benchmarks, demonstrates its
competitive edge against the current state-of-the-art methodologies. The code
is available at https://github.com/wangxu0820/SAN. | cs.CL | Stochastic Adversarial Networks for Multi-Domain Text Classification |
2024-05-28T00:00:04Z | http://arxiv.org/pdf/2405.17710v2 | http://arxiv.org/abs/2405.17710v2 | Paiheng Xu, Louiqa Raschid, Vanessa Frias-Martinez | 2024-06-28T06:24:37Z | cs.SI, cs.CL | Social media platforms like Twitter (now X) have been pivotal in information
dissemination and public engagement, especially during COVID-19. A key goal for
public health experts was to encourage prosocial behavior that could impact
local outcomes such as masking and social distancing. Given the importance of
local news and guidance during COVID-19, the objective of our research is to
analyze the effect of localized engagement, on social media conversations. This
study examines the impact of geographic co-location, as a proxy for localized
engagement between public health experts (PHEs) and the public, on social
media. We analyze a Twitter conversation dataset from January 2020 to November
2021, comprising over 19 K tweets from nearly five hundred PHEs, along with
approximately 800 K replies from 350 K participants. Our findings reveal that
geo-co-location is associated with higher engagement rates, especially in
conversations on topics including masking, lockdowns, and education, and in
conversations with academic and medical professionals. Lexical features
associated with emotion and personal experiences were more common in
geo-co-located contexts. This research provides insights into how geographic
co-location influences social media engagement and can inform strategies to
improve public health messaging. | cs.SI | Does Geo-co-location Matter? A Case Study of Public Health Conversations during COVID-19 |
2024-05-27T23:22:23Z | http://arxiv.org/pdf/2405.17703v1 | http://arxiv.org/abs/2405.17703v1 | Jason Li | 2024-05-27T23:22:23Z | cs.LG, cs.CL | Recent research (arXiv:2310.11453, arXiv:2402.17764) has proposed binary and
ternary transformer networks as a way to significantly reduce memory and
improve inference speed in Large Language Models (LLMs) while maintaining
accuracy. In this work, we apply techniques from mechanistic interpretability
to investigate whether such networks learn distinctly different or similar
algorithms when compared to full-precision transformer networks. In particular,
we reverse engineer the algorithms learned for the toy problem of modular
addition where we find that binary and ternary networks learn similar
algorithms as full precision networks. This provides evidence against the
possibility of using binary and ternary networks as a more interpretable
alternative in the LLM setting. | cs.LG | Mechanistic Interpretability of Binary and Ternary Transformers |
2024-05-27T21:03:26Z | http://arxiv.org/pdf/2405.17658v1 | http://arxiv.org/abs/2405.17658v1 | Kaustubh D. Dhole, Ramraj Chandradevan, Eugene Agichtein | 2024-05-27T21:03:26Z | cs.IR, cs.CL, H.3.3; I.2.7 | Query Reformulation (QR) is a set of techniques used to transform a user's
original search query to a text that better aligns with the user's intent and
improves their search experience. Recently, zero-shot QR has been a promising
approach due to its ability to exploit knowledge inherent in large language
models. Inspired by the success of ensemble prompting strategies which have
benefited other tasks, we investigate if they can improve query reformulation.
In this context, we propose two ensemble-based prompting techniques,
GenQREnsemble and GenQRFusion which leverage paraphrases of a zero-shot
instruction to generate multiple sets of keywords to improve retrieval
performance ultimately. We further introduce their post-retrieval variants to
incorporate relevance feedback from a variety of sources, including an oracle
simulating a human user and a "critic" LLM. We demonstrate that an ensemble of
query reformulations can improve retrieval effectiveness by up to 18% on
nDCG@10 in pre-retrieval settings and 9% on post-retrieval settings on multiple
benchmarks, outperforming all previously reported SOTA results. We perform
subsequent analyses to investigate the effects of feedback documents,
incorporate domain-specific instructions, filter reformulations, and generate
fluent reformulations that might be more beneficial to human searchers.
Together, the techniques and the results presented in this paper establish a
new state of the art in automated query reformulation for retrieval and suggest
promising directions for future research. | cs.IR | Generative Query Reformulation Using Ensemble Prompting, Document Fusion, and Relevance Feedback |
2024-05-27T20:53:22Z | http://arxiv.org/pdf/2405.17653v2 | http://arxiv.org/abs/2405.17653v2 | Xinting Huang, Madhur Panwar, Navin Goyal, Michael Hahn | 2024-06-02T08:50:02Z | cs.LG, cs.AI, cs.CL | The inner workings of neural networks can be better understood if we can
fully decipher the information encoded in neural activations. In this paper, we
argue that this information is embodied by the subset of inputs that give rise
to similar activations. Computing such subsets is nontrivial as the input space
is exponentially large. We propose InversionView, which allows us to
practically inspect this subset by sampling from a trained decoder model
conditioned on activations. This helps uncover the information content of
activation vectors, and facilitates understanding of the algorithms implemented
by transformer models. We present three case studies where we investigate
models ranging from small transformers to GPT-2. In these studies, we
demonstrate the characteristics of our method, show the distinctive advantages
it offers, and provide causally verified circuits. | cs.LG | InversionView: A General-Purpose Method for Reading Information from Neural Activations |
2024-05-27T20:29:13Z | http://arxiv.org/pdf/2406.02575v1 | http://arxiv.org/abs/2406.02575v1 | Trishna Chakraborty, Erfan Shayegani, Zikui Cai, Nael Abu-Ghazaleh, M. Salman Asif, Yue Dong, Amit K. Roy-Chowdhury, Chengyu Song | 2024-05-27T20:29:13Z | cs.CL, cs.CR, cs.LG | Recent studies reveal that integrating new modalities into Large Language
Models (LLMs), such as Vision-Language Models (VLMs), creates a new attack
surface that bypasses existing safety training techniques like Supervised
Fine-tuning (SFT) and Reinforcement Learning with Human Feedback (RLHF). While
further SFT and RLHF-based safety training can be conducted in multi-modal
settings, collecting multi-modal training datasets poses a significant
challenge. Inspired by the structural design of recent multi-modal models,
where, regardless of the combination of input modalities, all inputs are
ultimately fused into the language space, we aim to explore whether unlearning
solely in the textual domain can be effective for cross-modality safety
alignment. Our evaluation across six datasets empirically demonstrates the
transferability -- textual unlearning in VLMs significantly reduces the Attack
Success Rate (ASR) to less than 8\% and in some cases, even as low as nearly
2\% for both text-based and vision-text-based attacks, alongside preserving the
utility. Moreover, our experiments show that unlearning with a multi-modal
dataset offers no potential benefits but incurs significantly increased
computational demands, possibly up to 6 times higher. | cs.CL | Cross-Modal Safety Alignment: Is textual unlearning all you need? |
2024-05-27T20:00:38Z | http://arxiv.org/pdf/2405.17633v1 | http://arxiv.org/abs/2405.17633v1 | Jocelyn Shen, Joel Mire, Hae Won Park, Cynthia Breazeal, Maarten Sap | 2024-05-27T20:00:38Z | cs.CL | Empathy serves as a cornerstone in enabling prosocial behaviors, and can be
evoked through sharing of personal experiences in stories. While empathy is
influenced by narrative content, intuitively, people respond to the way a story
is told as well, through narrative style. Yet the relationship between empathy
and narrative style is not fully understood. In this work, we empirically
examine and quantify this relationship between style and empathy using LLMs and
large-scale crowdsourcing studies. We introduce a novel, theory-based taxonomy,
HEART (Human Empathy and Narrative Taxonomy) that delineates elements of
narrative style that can lead to empathy with the narrator of a story. We
establish the performance of LLMs in extracting narrative elements from HEART,
showing that prompting with our taxonomy leads to reasonable, human-level
annotations beyond what prior lexicon-based methods can do. To show empirical
use of our taxonomy, we collect a dataset of empathy judgments of stories via a
large-scale crowdsourcing study with N=2,624 participants. We show that
narrative elements extracted via LLMs, in particular, vividness of emotions and
plot volume, can elucidate the pathways by which narrative style cultivates
empathy towards personal stories. Our work suggests that such models can be
used for narrative analyses that lead to human-centered social and behavioral
insights. | cs.CL | HEART-felt Narratives: Tracing Empathy and Narrative Style in Personal Stories with LLMs |
2024-05-27T19:22:41Z | http://arxiv.org/pdf/2405.17613v1 | http://arxiv.org/abs/2405.17613v1 | Divyam Madaan, Taro Makino, Sumit Chopra, Kyunghyun Cho | 2024-05-27T19:22:41Z | cs.CV, cs.CL, cs.LG | Supervised multi-modal learning involves mapping multiple modalities to a
target label. Previous studies in this field have concentrated on capturing in
isolation either the inter-modality dependencies (the relationships between
different modalities and the label) or the intra-modality dependencies (the
relationships within a single modality and the label). We argue that these
conventional approaches that rely solely on either inter- or intra-modality
dependencies may not be optimal in general. We view the multi-modal learning
problem from the lens of generative models where we consider the target as a
source of multiple modalities and the interaction between them. Towards that
end, we propose inter- & intra-modality modeling (I2M2) framework, which
captures and integrates both the inter- and intra-modality dependencies,
leading to more accurate predictions. We evaluate our approach using real-world
healthcare and vision-and-language datasets with state-of-the-art models,
demonstrating superior performance over traditional methods focusing only on
one type of modality dependency. | cs.CV | A Framework for Multi-modal Learning: Jointly Modeling Inter- & Intra-Modality Dependencies |
2024-05-27T19:16:42Z | http://arxiv.org/pdf/2405.17610v1 | http://arxiv.org/abs/2405.17610v1 | Francisco de Arriba-Pérez, Silvia García-Méndez, Francisco J. González-Castaño, Jaime González-González | 2024-05-27T19:16:42Z | cs.CL, cs.AI, cs.LG | Artificial Intelligence techniques such as Machine Learning (ML) have not
been exploited to their maximum potential in the legal domain. This has been
partially due to the insufficient explanations they provided about their
decisions. Automatic expert systems with explanatory capabilities can be
specially useful when legal practitioners search jurisprudence to gather
contextual knowledge for their cases. Therefore, we propose a hybrid system
that applies ML for multi-label classification of judgements (sentences) and
visual and natural language descriptions for explanation purposes, boosted by
Natural Language Processing techniques and deep legal reasoning to identify the
entities, such as the parties, involved. We are not aware of any prior work on
automatic multi-label classification of legal judgements also providing natural
language explanations to the end-users with comparable overall quality. Our
solution achieves over 85 % micro precision on a labelled data set annotated by
legal experts. This endorses its interest to relieve human experts from
monotonous labour-intensive legal classification tasks. | cs.CL | Explainable machine learning multi-label classification of Spanish legal judgements |
2024-05-27T19:07:13Z | http://arxiv.org/pdf/2405.17604v1 | http://arxiv.org/abs/2405.17604v1 | Klaudia Bałazy, Mohammadreza Banaei, Karl Aberer, Jacek Tabor | 2024-05-27T19:07:13Z | cs.LG, cs.AI, cs.CL | The recent trend in scaling language models has led to a growing demand for
parameter-efficient tuning (PEFT) methods such as LoRA (Low-Rank Adaptation).
LoRA consistently matches or surpasses the full fine-tuning baseline with fewer
parameters. However, handling numerous task-specific or user-specific LoRA
modules on top of a base model still presents significant storage challenges.
To address this, we introduce LoRA-XS (Low-Rank Adaptation with eXtremely Small
number of parameters), a novel approach leveraging Singular Value Decomposition
(SVD) for parameter-efficient fine-tuning. LoRA-XS introduces a small r x r
weight matrix between frozen LoRA matrices, which are constructed by SVD of the
original weight matrix. Training only r x r weight matrices ensures
independence from model dimensions, enabling more parameter-efficient
fine-tuning, especially for larger models. LoRA-XS achieves a remarkable
reduction of trainable parameters by over 100x in 7B models compared to LoRA.
Our benchmarking across various scales, including GLUE, GSM8k, and MATH
benchmarks, shows that our approach outperforms LoRA and recent
state-of-the-art approaches like VeRA in terms of parameter efficiency while
maintaining competitive performance. | cs.LG | LoRA-XS: Low-Rank Adaptation with Extremely Small Number of Parameters |
2024-05-27T18:40:49Z | http://arxiv.org/pdf/2405.17587v1 | http://arxiv.org/abs/2405.17587v1 | Emile Contal, Garrin McGoldrick | 2024-05-27T18:40:49Z | cs.IR, cs.CL, cs.LG | Large Language Models (LLM) hold immense promise for real-world applications,
but their generic knowledge often falls short of domain-specific needs.
Fine-tuning, a common approach, can suffer from catastrophic forgetting and
hinder generalizability. In-Context Learning (ICL) offers an alternative, which
can leverage Retrieval-Augmented Generation (RAG) to provide LLMs with relevant
demonstrations for few-shot learning tasks. This paper explores the desired
qualities of a demonstration retrieval system for ICL. We argue that ICL
retrieval in this context resembles item-cold-start recommender systems,
prioritizing discovery and maximizing information gain over strict relevance.
We propose a novel evaluation method that measures the LLM's subsequent
performance on NLP tasks, eliminating the need for subjective diversity scores.
Our findings demonstrate the critical role of diversity and quality bias in
retrieved demonstrations for effective ICL, and highlight the potential of
recommender system techniques in this domain. | cs.IR | RAGSys: Item-Cold-Start Recommender as RAG System |
2024-05-27T17:59:56Z | http://arxiv.org/pdf/2405.17430v1 | http://arxiv.org/abs/2405.17430v1 | Mu Cai, Jianwei Yang, Jianfeng Gao, Yong Jae Lee | 2024-05-27T17:59:56Z | cs.CV, cs.AI, cs.CL, cs.LG | Large Multimodal Models (LMMs) such as LLaVA have shown strong performance in
visual-linguistic reasoning. These models first embed images into a fixed large
number of visual tokens and then feed them into a Large Language Model (LLM).
However, this design causes an excessive number of tokens for dense visual
scenarios such as high-resolution images and videos, leading to great
inefficiency. While token pruning/merging methods do exist, they produce a
single length output for each image and do not afford flexibility in trading
off information density v.s. efficiency. Inspired by the concept of Matryoshka
Dolls, we propose M3: Matryoshka Multimodal Models, which learns to represent
visual content as nested sets of visual tokens that capture information across
multiple coarse-to-fine granularities. Our approach offers several unique
benefits for LMMs: (1) One can explicitly control the visual granularity per
test instance during inference, e.g. , adjusting the number of tokens used to
represent an image based on the anticipated complexity or simplicity of the
content; (2) M3 provides a framework for analyzing the granularity needed for
existing datasets, where we find that COCO-style benchmarks only need around ~9
visual tokens to obtain accuracy similar to that of using all 576 tokens; (3)
Our approach provides a foundation to explore the best trade-off between
performance and visual token length at sample level, where our investigation
reveals that a large gap exists between the oracle upper bound and current
fixed-scale representations. | cs.CV | Matryoshka Multimodal Models |
2024-05-27T17:59:45Z | http://arxiv.org/pdf/2405.17428v1 | http://arxiv.org/abs/2405.17428v1 | Chankyu Lee, Rajarshi Roy, Mengyao Xu, Jonathan Raiman, Mohammad Shoeybi, Bryan Catanzaro, Wei Ping | 2024-05-27T17:59:45Z | cs.CL, cs.AI, cs.IR, cs.LG | Decoder-only large language model (LLM)-based embedding models are beginning
to outperform BERT or T5-based embedding models in general-purpose text
embedding tasks, including dense vector-based retrieval. In this work, we
introduce the NV-Embed model with a variety of architectural designs and
training procedures to significantly enhance the performance of LLM as a
versatile embedding model, while maintaining its simplicity and
reproducibility. For model architecture, we propose a latent attention layer to
obtain pooled embeddings, which consistently improves retrieval and downstream
task accuracy compared to mean pooling or using the last <EOS> token embedding
from LLMs. To enhance representation learning, we remove the causal attention
mask of LLMs during contrastive training. For model training, we introduce a
two-stage contrastive instruction-tuning method. It first applies contrastive
training with instructions on retrieval datasets, utilizing in-batch negatives
and curated hard negative examples. At stage-2, it blends various non-retrieval
datasets into instruction tuning, which not only enhances non-retrieval task
accuracy but also improves retrieval performance. Combining these techniques,
our NV-Embed model, using only publicly available data, has achieved a
record-high score of 69.32, ranking No. 1 on the Massive Text Embedding
Benchmark (MTEB) (as of May 24, 2024), with 56 tasks, encompassing retrieval,
reranking, classification, clustering, and semantic textual similarity tasks.
Notably, our model also attains the highest score of 59.36 on 15 retrieval
tasks in the MTEB benchmark (also known as BEIR). We will open-source the model
at: https://huggingface.co/nvidia/NV-Embed-v1. | cs.CL | NV-Embed: Improved Techniques for Training LLMs as Generalist Embedding Models |
2024-05-27T17:59:25Z | http://arxiv.org/pdf/2405.17423v1 | http://arxiv.org/abs/2405.17423v1 | Laurens Samson, Nimrod Barazani, Sennay Ghebreab, Yuki M. Asano | 2024-05-27T17:59:25Z | cs.CV, cs.CL | This paper aims to advance our understanding of how Visual Language Models
(VLMs) handle privacy-sensitive information, a crucial concern as these
technologies become integral to everyday life. To this end, we introduce a new
benchmark PrivBench, which contains images from 8 sensitive categories such as
passports, or fingerprints. We evaluate 10 state-of-the-art VLMs on this
benchmark and observe a generally limited understanding of privacy,
highlighting a significant area for model improvement. Based on this we
introduce PrivTune, a new instruction-tuning dataset aimed at equipping VLMs
with knowledge about visual privacy. By tuning two pretrained VLMs, TinyLLaVa
and MiniGPT-v2, on this small dataset, we achieve strong gains in their ability
to recognize sensitive content, outperforming even GPT4-V. At the same time, we
show that privacy-tuning only minimally affects the VLMs performance on
standard benchmarks such as VQA. Overall, this paper lays out a crucial
challenge for making VLMs effective in handling real-world data safely and
provides a simple recipe that takes the first step towards building
privacy-aware VLMs. | cs.CV | Privacy-Aware Visual Language Models |
2024-05-27T17:57:48Z | http://arxiv.org/pdf/2405.17537v1 | http://arxiv.org/abs/2405.17537v1 | ZeMing Gong, Austin T. Wang, Joakim Bruslund Haurum, Scott C. Lowe, Graham W. Taylor, Angel X. Chang | 2024-05-27T17:57:48Z | cs.AI, cs.CL, cs.CV | Measuring biodiversity is crucial for understanding ecosystem health. While
prior works have developed machine learning models for the taxonomic
classification of photographic images and DNA separately, in this work, we
introduce a multimodal approach combining both, using CLIP-style contrastive
learning to align images, DNA barcodes, and textual data in a unified embedding
space. This allows for accurate classification of both known and unknown insect
species without task-specific fine-tuning, leveraging contrastive learning for
the first time to fuse DNA and image data. Our method surpasses previous
single-modality approaches in accuracy by over 11% on zero-shot learning tasks,
showcasing its effectiveness in biodiversity studies. | cs.AI | BIOSCAN-CLIP: Bridging Vision and Genomics for Biodiversity Monitoring at Scale |
2024-05-27T17:55:36Z | http://arxiv.org/pdf/2406.00041v2 | http://arxiv.org/abs/2406.00041v2 | Rui Guo, Greg Farnan, Niall McLaughlin, Barry Devereux | 2024-06-27T14:31:22Z | cs.CL | The BioNLP ACL'24 Shared Task on Streamlining Discharge Documentation aims to
reduce the administrative burden on clinicians by automating the creation of
critical sections of patient discharge letters. This paper presents our
approach using the Llama3 8B quantized model to generate the "Brief Hospital
Course" and "Discharge Instructions" sections. We employ a zero-shot method
combined with Retrieval-Augmented Generation (RAG) to produce concise,
contextually accurate summaries. Our contributions include the development of a
curated template-based approach to ensure reliability and consistency, as well
as the integration of RAG for word count prediction. We also describe several
unsuccessful experiments to provide insights into our pathway for the
competition. Our results demonstrate the effectiveness and efficiency of our
approach, achieving high scores across multiple evaluation metrics. | cs.CL | QUB-Cirdan at "Discharge Me!": Zero shot discharge letter generation by open-source LLM |
2024-05-27T17:51:24Z | http://arxiv.org/pdf/2405.17402v1 | http://arxiv.org/abs/2405.17402v1 | Philip Schroeder, Nathaniel Morgan, Hongyin Luo, James Glass | 2024-05-27T17:51:24Z | cs.CL | Large language models (LLMs) have shown impressive capabilities across
diverse settings, but still struggle as the length and complexity of the
context increases. To address this challenge, we propose Thinking Recursively
and Dynamically (ThReaD). THREAD frames model generation as a thread of
execution that, based on the context, can run to completion or dynamically
spawn new threads. By spawning, threads can offload work (e.g., thinking,
retrieving information) to child threads, which only return tokens needed for
the parent thread to do its work. In effect, this enables the model to adapt,
as needed, the amount of intermediate work used to produce tokens. We apply
THREAD in the settings of LLM task solving and question answering, where the
dynamic threading allows the model to recursively decompose the given task or
question into progressively simpler sub-problems that can be solved by separate
child threads. We test THREAD, implemented using a few-shot learning approach,
on diverse benchmarks for agent tasks and data-grounded question answering.
THREAD achieves state-of-the-art performance with GPT-4 and GPT-3.5 on these
benchmarks, including ALFWorld, TextCraft, and WebShop, along with two new
benchmarks, DataCommons QA and MIMIC-III ICU QA. In addition, THREAD
outperforms existing frameworks by 10% to 50% absolute points with smaller
models, including Llama-3-8b and CodeLlama-7b. | cs.CL | THREAD: Thinking Deeper with Recursive Spawning |
2024-05-27T17:46:57Z | http://arxiv.org/pdf/2405.17394v2 | http://arxiv.org/abs/2405.17394v2 | Yash Sarrof, Yana Veitsman, Michael Hahn | 2024-06-02T19:43:55Z | cs.CL, cs.FL, cs.LG | Recently, recurrent models based on linear state space models (SSMs) have
shown promising performance in language modeling (LM), competititve with
transformers. However, there is little understanding of the in-principle
abilities of such models, which could provide useful guidance to the search for
better LM architectures. We present a comprehensive theoretical study of the
capacity of such SSMs as it compares to that of transformers and traditional
RNNs. We find that SSMs and transformers have overlapping but distinct
strengths. In star-free state tracking, SSMs implement straightforward and
exact solutions to problems that transformers struggle to represent exactly.
They can also model bounded hierarchical structure with optimal memory even
without simulating a stack. On the other hand, we identify a design choice in
current SSMs that limits their expressive power. We discuss implications for
SSM and LM research, and verify results empirically on a recent SSM, Mamba. | cs.CL | The Expressive Capacity of State Space Models: A Formal Language Perspective |
2024-05-27T17:42:54Z | http://arxiv.org/pdf/2405.17390v1 | http://arxiv.org/abs/2405.17390v1 | Nimol Thuon, Wangrui Zhang, Sada Thuon | 2024-05-27T17:42:54Z | cs.IR, cs.CL | This paper introduces KSW, a Khmer-specific approach to keyword extraction
that leverages a specialized stop word dictionary. Due to the limited
availability of natural language processing resources for the Khmer language,
effective keyword extraction has been a significant challenge. KSW addresses
this by developing a tailored stop word dictionary and implementing a
preprocessing methodology to remove stop words, thereby enhancing the
extraction of meaningful keywords. Our experiments demonstrate that KSW
achieves substantial improvements in accuracy and relevance compared to
previous methods, highlighting its potential to advance Khmer text processing
and information retrieval. The KSW resources, including the stop word
dictionary, are available at the following GitHub repository:
(https://github.com/back-kh/KSWv2-Khmer-Stop-Word-based-Dictionary-for-Keyword-Extraction.git). | cs.IR | KSW: Khmer Stop Word based Dictionary for Keyword Extraction |
2024-05-27T17:41:54Z | http://arxiv.org/pdf/2405.17386v1 | http://arxiv.org/abs/2405.17386v1 | Zixian Huang, Wenhao Zhu, Gong Cheng, Lei Li, Fei Yuan | 2024-05-27T17:41:54Z | cs.CL, cs.AI | Reasoning capabilities are crucial for Large Language Models (LLMs), yet a
notable gap exists between English and non-English languages. To bridge this
disparity, some works fine-tune LLMs to relearn reasoning capabilities in
non-English languages, while others replace non-English inputs with an external
model's outputs such as English translation text to circumvent the challenge of
LLM understanding non-English. Unfortunately, these methods often underutilize
the built-in skilled reasoning and useful language understanding capabilities
of LLMs. In order to better utilize the minds of reasoning and language
understanding in LLMs, we propose a new method, namely MindMerger, which merges
LLMs with the external language understanding capabilities from multilingual
models to boost the multilingual reasoning performance. Furthermore, a two-step
training scheme is introduced to first train to embeded the external
capabilities into LLMs and then train the collaborative utilization of the
external capabilities and the built-in capabilities in LLMs. Experiments on
three multilingual reasoning datasets and a language understanding dataset
demonstrate that MindMerger consistently outperforms all baselines, especially
in low-resource languages. Without updating the parameters of LLMs, the average
accuracy improved by 6.7% and 8.0% across all languages and low-resource
languages on the MGSM dataset, respectively. | cs.CL | MindMerger: Efficient Boosting LLM Reasoning in non-English Languages |
2024-05-27T17:38:55Z | http://arxiv.org/pdf/2405.17383v1 | http://arxiv.org/abs/2405.17383v1 | Zhen Qin, Xuyang Shen, Dong Li, Weigao Sun, Stan Birchfield, Richard Hartley, Yiran Zhong | 2024-05-27T17:38:55Z | cs.CL | We present the Linear Complexity Sequence Model (LCSM), a comprehensive
solution that unites various sequence modeling techniques with linear
complexity, including linear attention, state space model, long convolution,
and linear RNN, within a single framework. The goal is to enhance comprehension
of these models by analyzing the impact of each component from a cohesive and
streamlined viewpoint. Specifically, we segment the modeling processes of these
models into three distinct stages: Expand, Oscillation, and Shrink (EOS), with
each model having its own specific settings. The Expand stage involves
projecting the input signal onto a high-dimensional memory state. This is
followed by recursive operations performed on the memory state in the
Oscillation stage. Finally, the memory state is projected back to a
low-dimensional space in the Shrink stage. We perform comprehensive experiments
to analyze the impact of different stage settings on language modeling and
retrieval tasks. Our results show that data-driven methods are crucial for the
effectiveness of the three stages in language modeling, whereas hand-crafted
methods yield better performance in retrieval tasks. | cs.CL | Unlocking the Secrets of Linear Complexity Sequence Model from A Unified Perspective |
2024-05-27T17:38:33Z | http://arxiv.org/pdf/2405.17382v1 | http://arxiv.org/abs/2405.17382v1 | Hyunseok Lee, Jihoon Tack, Jinwoo Shin | 2024-05-27T17:38:33Z | cs.LG, cs.CL | The remarkable capabilities and easy accessibility of large language models
(LLMs) have significantly increased societal risks (e.g., fake news
generation), necessitating the development of LLM-generated text (LGT)
detection methods for safe usage. However, detecting LGTs is challenging due to
the vast number of LLMs, making it impractical to account for each LLM
individually; hence, it is crucial to identify the common characteristics
shared by these models. In this paper, we draw attention to a common feature of
recent powerful LLMs, namely the alignment training, i.e., training LLMs to
generate human-preferable texts. Our key finding is that as these aligned LLMs
are trained to maximize the human preferences, they generate texts with higher
estimated preferences even than human-written texts; thus, such texts are
easily detected by using the reward model (i.e., an LLM trained to model human
preference distribution). Based on this finding, we propose two training
schemes to further improve the detection ability of the reward model, namely
(i) continual preference fine-tuning to make the reward model prefer aligned
LGTs even further and (ii) reward modeling of Human/LLM mixed texts (a
rephrased texts from human-written texts using aligned LLMs), which serves as a
median preference text corpus between LGTs and human-written texts to learn the
decision boundary better. We provide an extensive evaluation by considering six
text domains across twelve aligned LLMs, where our method demonstrates
state-of-the-art results. Code is available at
https://github.com/hyunseoklee-ai/reward_llm_detect. | cs.LG | ReMoDetect: Reward Models Recognize Aligned LLM's Generations |
2024-05-27T17:38:13Z | http://arxiv.org/pdf/2405.17381v2 | http://arxiv.org/abs/2405.17381v2 | Zhen Qin, Weigao Sun, Dong Li, Xuyang Shen, Weixuan Sun, Yiran Zhong | 2024-06-20T09:12:42Z | cs.CL | We present Lightning Attention, the first linear attention implementation
that maintains a constant training speed for various sequence lengths under
fixed memory consumption. Due to the issue with cumulative summation operations
(cumsum), previous linear attention implementations cannot achieve their
theoretical advantage in a casual setting. However, this issue can be
effectively solved by utilizing different attention calculation strategies to
compute the different parts of attention. Specifically, we split the attention
calculation into intra-blocks and inter-blocks and use conventional attention
computation for intra-blocks and linear attention kernel tricks for
inter-blocks. This eliminates the need for cumsum in the linear attention
calculation. Furthermore, a tiling technique is adopted through both forward
and backward procedures to take full advantage of the GPU hardware. To enhance
accuracy while preserving efficacy, we introduce TransNormerLLM (TNL), a new
architecture that is tailored to our lightning attention. We conduct rigorous
testing on standard and self-collected datasets with varying model sizes and
sequence lengths. TNL is notably more efficient than other language models. In
addition, benchmark results indicate that TNL performs on par with
state-of-the-art LLMs utilizing conventional transformer structures. The source
code is released at github.com/OpenNLPLab/TransnormerLLM. | cs.CL | Various Lengths, Constant Speed: Efficient Language Modeling with Lightning Attention |
2024-05-27T17:32:37Z | http://arxiv.org/pdf/2405.17376v1 | http://arxiv.org/abs/2405.17376v1 | Mohamed Nabih Ali, Alessio Brutti, Daniele Falavigna | 2024-05-27T17:32:37Z | cs.CL | Automatic speech recognition models require large amounts of speech
recordings for training. However, the collection of such data often is
cumbersome and leads to privacy concerns. Federated learning has been widely
used as an effective decentralized technique that collaboratively learns a
shared prediction model while keeping the data local on different clients.
Unfortunately, client devices often feature limited computation and
communication resources leading to practical difficulties for large models. In
addition, the heterogeneity that characterizes edge devices makes it
sub-optimal to generate a single model that fits all of them. Differently from
the recent literature, where multiple models with different architectures are
used, in this work, we propose using dynamical architectures which, employing
early-exit solutions, can adapt their processing (i.e. traversed layers)
depending on the input and on the operation conditions. This solution falls in
the realm of partial training methods and brings two benefits: a single model
is used on a variety of devices; federating the models after local training is
straightforward. Experiments on public datasets show that our proposed approach
is effective and can be combined with basic federated learning strategies. | cs.CL | Federating Dynamic Models using Early-Exit Architectures for Automatic Speech Recognition on Heterogeneous Clients |
2024-05-27T17:10:04Z | http://arxiv.org/pdf/2405.17361v1 | http://arxiv.org/abs/2405.17361v1 | Yuhao Zhang, Aws Albarghouthi, Loris D'Antoni | 2024-05-27T17:10:04Z | cs.CL | This paper reveals a key insight that a one-layer decoder-only Transformer is
equivalent to a two-layer Recurrent Neural Network (RNN). Building on this
insight, we propose ARC-Tran, a novel approach for verifying the robustness of
decoder-only Transformers against arbitrary perturbation spaces. Compared to
ARC-Tran, current robustness verification techniques are limited either to
specific and length-preserving perturbations like word substitutions or to
recursive models like LSTMs. ARC-Tran addresses these limitations by
meticulously managing position encoding to prevent mismatches and by utilizing
our key insight to achieve precise and scalable verification. Our evaluation
shows that ARC-Tran (1) trains models more robust to arbitrary perturbation
spaces than those produced by existing techniques and (2) shows high
certification accuracy of the resulting models. | cs.CL | A One-Layer Decoder-Only Transformer is a Two-Layer RNN: With an Application to Certified Robustness |
2024-05-27T17:02:27Z | http://arxiv.org/pdf/2405.17357v3 | http://arxiv.org/abs/2405.17357v3 | Yulong Mao, Kaiyu Huang, Changhao Guan, Ganglin Bao, Fengran Mo, Jinan Xu | 2024-06-26T14:41:33Z | cs.CL | Fine-tuning large-scale pre-trained models is inherently a resource-intensive
task. While it can enhance the capabilities of the model, it also incurs
substantial computational costs, posing challenges to the practical application
of downstream tasks. Existing parameter-efficient fine-tuning (PEFT) methods
such as Low-Rank Adaptation (LoRA) rely on a bypass framework that ignores the
differential parameter budget requirements across weight matrices, which may
lead to suboptimal fine-tuning outcomes. To address this issue, we introduce
the Dynamic Low-Rank Adaptation (DoRA) method. DoRA decomposes high-rank LoRA
layers into structured single-rank components, allowing for dynamic pruning of
parameter budget based on their importance to specific tasks during training,
which makes the most of the limited parameter budget. Experimental results
demonstrate that DoRA can achieve competitive performance compared with LoRA
and full model fine-tuning, and outperform various strong baselines with the
same storage parameter budget. Our code is available at
https://github.com/MIkumikumi0116/DoRA | cs.CL | DoRA: Enhancing Parameter-Efficient Fine-Tuning with Dynamic Rank Distribution |
2024-05-27T16:49:22Z | http://arxiv.org/pdf/2405.17345v2 | http://arxiv.org/abs/2405.17345v2 | Alejandro Tlaie | 2024-06-06T11:53:23Z | cs.AI, cs.CL | Large Language Models (LLMs) have become central to advancing automation and
decision-making across various sectors, raising significant ethical questions.
This study proposes a comprehensive comparative analysis of the most advanced
LLMs to assess their moral profiles. We subjected several state-of-the-art
models to a selection of ethical dilemmas and found that all the proprietary
ones are mostly utilitarian and all of the open-weights ones align mostly with
values-based ethics. Furthermore, when using the Moral Foundations
Questionnaire, all models we probed - except for Llama 2-7B - displayed a
strong liberal bias. Lastly, in order to causally intervene in one of the
studied models, we propose a novel similarity-specific activation steering
technique. Using this method, we were able to reliably steer the model's moral
compass to different ethical schools. All of these results showcase that there
is an ethical dimension in already deployed LLMs, an aspect that is generally
overlooked. | cs.AI | Exploring and steering the moral compass of Large Language Models |
2024-05-27T16:37:34Z | http://arxiv.org/pdf/2405.17337v1 | http://arxiv.org/abs/2405.17337v1 | Junnan Dong, Qinggang Zhang, Chuang Zhou, Hao Chen, Daochen Zha, Xiao Huang | 2024-05-27T16:37:34Z | cs.CL, cs.AI | Knowledge-based question answering (KBQA) is widely used in many scenarios
that necessitate domain knowledge. Large language models (LLMs) bring
opportunities to KBQA, while their costs are significantly higher and absence
of domain-specific knowledge during pre-training. We are motivated to combine
LLMs and prior small models on knowledge graphs (KGMs) for both inferential
accuracy and cost saving. However, it remains challenging since accuracy and
cost are not readily combined in the optimization as two distinct metrics. It
is also laborious for model selection since different models excel in diverse
knowledge. To this end, we propose Coke, a novel cost-efficient strategy for
KBQA with LLMs, modeled as a tailored multi-armed bandit problem to minimize
calls to LLMs within limited budgets. We first formulate the accuracy
expectation with a cluster-level Thompson Sampling for either KGMs or LLMs. A
context-aware policy is optimized to further distinguish the expert model
subject to the question semantics. The overall decision is bounded by the cost
regret according to historical expenditure on failures. Extensive experiments
showcase the superior performance of Coke, which moves the Pareto frontier with
up to 20.89% saving of GPT-4 fees while achieving a 2.74% higher accuracy on
the benchmark datasets. | cs.CL | Cost-efficient Knowledge-based Question Answering with Large Language Models |
2024-05-27T16:37:17Z | http://arxiv.org/pdf/2405.17336v1 | http://arxiv.org/abs/2405.17336v1 | Xianfu Cheng, Hang Zhang, Jian Yang, Xiang Li, Weixiao Zhou, Kui Wu, Fei Liu, Wei Zhang, Tao Sun, Tongliang Li, Zhoujun Li | 2024-05-27T16:37:17Z | cs.CL | In the domain of document AI, semi-structured form parsing plays a crucial
role. This task leverages techniques from key information extraction (KIE),
dealing with inputs that range from plain text to intricate modal data
comprising images and structural layouts. The advent of pre-trained multimodal
models has driven the extraction of key information from form documents in
different formats such as PDFs and images. Nonetheless, the endeavor of form
parsing is still encumbered by notable challenges like subpar capabilities in
multi-lingual parsing and diminished recall in contexts rich in text and
visuals. In this work, we introduce a simple but effective \textbf{M}ultimodal
and \textbf{M}ultilingual semi-structured \textbf{FORM} \textbf{PARSER}
(\textbf{XFormParser}), which is anchored on a comprehensive pre-trained
language model and innovatively amalgamates semantic entity recognition (SER)
and relation extraction (RE) into a unified framework, enhanced by a novel
staged warm-up training approach that employs soft labels to significantly
refine form parsing accuracy without amplifying inference overhead.
Furthermore, we have developed a groundbreaking benchmark dataset, named
InDFormBench, catering specifically to the parsing requirements of multilingual
forms in various industrial contexts. Through rigorous testing on established
multilingual benchmarks and InDFormBench, XFormParser has demonstrated its
unparalleled efficacy, notably surpassing the state-of-the-art (SOTA) models in
RE tasks within language-specific setups by achieving an F1 score improvement
of up to 1.79\%. Our framework exhibits exceptionally improved performance
across tasks in both multi-language and zero-shot contexts when compared to
existing SOTA benchmarks. The code is publicly available at
https://github.com/zhbuaa0/layoutlmft. | cs.CL | XFormParser: A Simple and Effective Multimodal Multilingual Semi-structured Form Parser |
2024-05-27T16:32:40Z | http://arxiv.org/pdf/2406.01608v1 | http://arxiv.org/abs/2406.01608v1 | Arya Ramteke, Sankalp Tembhurne, Gunesh Sonawane, Ratnmala N. Bhimanpallewar | 2024-05-27T16:32:40Z | cs.IR, cs.AI, cs.CL, cs.HC | Dark patterns are deceptive user interfaces employed by e-commerce websites
to manipulate user's behavior in a way that benefits the website, often
unethically. This study investigates the detection of such dark patterns.
Existing solutions include UIGuard, which uses computer vision and natural
language processing, and approaches that categorize dark patterns based on
detectability or utilize machine learning models trained on datasets. We
propose combining web scraping techniques with fine-tuned BERT language models
and generative capabilities to identify dark patterns, including outliers. The
approach scrapes textual content, feeds it into the BERT model for detection,
and leverages BERT's bidirectional analysis and generation abilities. The study
builds upon research on automatically detecting and explaining dark patterns,
aiming to raise awareness and protect consumers. | cs.IR | Detecting Deceptive Dark Patterns in E-commerce Platforms |
Subsets and Splits
No saved queries yet
Save your SQL queries to embed, download, and access them later. Queries will appear here once saved.