publication_date
stringlengths 20
20
| doc_url
stringlengths 33
34
| id
stringlengths 33
34
| authors
stringlengths 5
20.1k
| update_date
stringlengths 20
20
| category_all
stringlengths 5
138
| abstract
stringlengths 123
3.09k
| category_primary
stringclasses 78
values | title
stringlengths 10
211
|
|---|---|---|---|---|---|---|---|---|
2024-02-14T02:11:04Z | http://arxiv.org/pdf/2402.08898v1 | http://arxiv.org/abs/2402.08898v1 | Ruchao Fan, Natarajan Balaji Shanka, Abeer Alwan | 2024-02-14T02:11:04Z | eess.AS, cs.CL, cs.SD | Non-autoregressive automatic speech recognition (NASR) models have gained
attention due to their parallelism and fast inference. The encoder-based NASR,
e.g. connectionist temporal classification (CTC), can be initialized from the
speech foundation models (SFM) but does not account for any dependencies among
intermediate tokens. The encoder-decoder-based NASR, like CTC alignment-based
single-step non-autoregressive transformer (CASS-NAT), can mitigate the
dependency problem but is not able to efficiently integrate SFM. Inspired by
the success of recent work of speech-text joint pre-training with a shared
transformer encoder, we propose a new encoder-based NASR, UniEnc-CASSNAT, to
combine the advantages of CTC and CASS-NAT. UniEnc-CASSNAT consists of only an
encoder as the major module, which can be the SFM. The encoder plays the role
of both the CASS-NAT encoder and decoder by two forward passes. The first pass
of the encoder accepts the speech signal as input, while the concatenation of
the speech signal and the token-level acoustic embedding is used as the input
for the second pass. Examined on the Librispeech 100h, MyST, and Aishell1
datasets, the proposed UniEnc-CASSNAT achieves state-of-the-art NASR results
and is better or comparable to CASS-NAT with only an encoder and hence, fewer
model parameters. Our codes are publicly available. | eess.AS | UniEnc-CASSNAT: An Encoder-only Non-autoregressive ASR for Speech SSL Models |
2024-02-14T00:40:51Z | http://arxiv.org/pdf/2402.08874v1 | http://arxiv.org/abs/2402.08874v1 | Chenxi Lin, Jiayu Ren, Guoxiu He, Zhuoren Jiang, Haiyan Yu, Xiaomin Zhu | 2024-02-14T00:40:51Z | cs.CL | While large language models (LLMs) excel at understanding and generating
plain text, they are not specifically tailored to handle hierarchical text
structures. Extracting the task-desired property from their natural language
responses typically necessitates additional processing steps. In fact,
selectively comprehending the hierarchical structure of large-scale text is
pivotal to understanding its substance. Aligning LLMs more closely with the
classification or regression values of specific task through prompting also
remains challenging. To this end, we propose a novel framework called
Tree-Based Hard Attention with Self-Motivation for Large Language Models
(TEAROOM). TEAROOM incorporates a tree-based hard attention mechanism for LLMs
to process hierarchically structured text inputs. By leveraging prompting, it
enables a frozen LLM to selectively focus on relevant leaves in relation to the
root, generating a tailored symbolic representation of their relationship.
Moreover, TEAROOM comprises a self-motivation strategy for another LLM equipped
with a trainable adapter and a linear layer. The selected symbolic outcomes are
integrated into another prompt, along with the predictive value of the task. We
iteratively feed output values back into the prompt, enabling the trainable LLM
to progressively approximate the golden truth. TEAROOM outperforms existing
state-of-the-art methods in experimental evaluations across three benchmark
datasets, showing its effectiveness in estimating task-specific properties.
Through comprehensive experiments and analysis, we have validated the ability
of TEAROOM to gradually approach the underlying golden truth through multiple
inferences. | cs.CL | Tree-Based Hard Attention with Self-Motivation for Large Language Models |
2024-02-13T23:25:04Z | http://arxiv.org/pdf/2402.08846v1 | http://arxiv.org/abs/2402.08846v1 | Ziyang Ma, Guanrou Yang, Yifan Yang, Zhifu Gao, Jiaming Wang, Zhihao Du, Fan Yu, Qian Chen, Siqi Zheng, Shiliang Zhang, Xie Chen | 2024-02-13T23:25:04Z | cs.CL, cs.AI, cs.MM, cs.SD, eess.AS | In this paper, we focus on solving one of the most important tasks in the
field of speech processing, i.e., automatic speech recognition (ASR), with
speech foundation encoders and large language models (LLM). Recent works have
complex designs such as compressing the output temporally for the speech
encoder, tackling modal alignment for the projector, and utilizing
parameter-efficient fine-tuning for the LLM. We found that delicate designs are
not necessary, while an embarrassingly simple composition of off-the-shelf
speech encoder, LLM, and the only trainable linear projector is competent for
the ASR task. To be more specific, we benchmark and explore various
combinations of LLMs and speech encoders, leading to the optimal LLM-based ASR
system, which we call SLAM-ASR. The proposed SLAM-ASR provides a clean setup
and little task-specific design, where only the linear projector is trained. To
the best of our knowledge, SLAM-ASR achieves the best performance on the
Librispeech benchmark among LLM-based ASR models and even outperforms the
latest LLM-based audio-universal model trained on massive pair data. Finally,
we explore the capability emergence of LLM-based ASR in the process of modal
alignment. We hope that our study can facilitate the research on extending LLM
with cross-modality capacity and shed light on the LLM-based ASR community. | cs.CL | An Embarrassingly Simple Approach for LLM with Strong ASR Capacity |
2024-02-13T22:47:22Z | http://arxiv.org/pdf/2402.08837v1 | http://arxiv.org/abs/2402.08837v1 | Maneesh Bilalpur, Mert Inan, Dorsa Zeinali, Jeffrey F. Cohn, Malihe Alikhani | 2024-02-13T22:47:22Z | cs.CL | Addressing the critical shortage of mental health resources for effective
screening, diagnosis, and treatment remains a significant challenge. This
scarcity underscores the need for innovative solutions, particularly in
enhancing the accessibility and efficacy of therapeutic support. Embodied
agents with advanced interactive capabilities emerge as a promising and
cost-effective supplement to traditional caregiving methods. Crucial to these
agents' effectiveness is their ability to simulate non-verbal behaviors, like
backchannels, that are pivotal in establishing rapport and understanding in
therapeutic contexts but remain under-explored. To improve the rapport-building
capabilities of embodied agents we annotated backchannel smiles in videos of
intimate face-to-face conversations over topics such as mental health, illness,
and relationships. We hypothesized that both speaker and listener behaviors
affect the duration and intensity of backchannel smiles. Using cues from speech
prosody and language along with the demographics of the speaker and listener,
we found them to contain significant predictors of the intensity of backchannel
smiles. Based on our findings, we introduce backchannel smile production in
embodied agents as a generation problem. Our attention-based generative model
suggests that listener information offers performance improvements over the
baseline speaker-centric generation approach. Conditioned generation using the
significant predictors of smile intensity provides statistically significant
improvements in empirical measures of generation quality. Our user study by
transferring generated smiles to an embodied agent suggests that agent with
backchannel smiles is perceived to be more human-like and is an attractive
alternative for non-personal conversations over agent without backchannel
smiles. | cs.CL | Learning to Generate Context-Sensitive Backchannel Smiles for Embodied AI Agents with Applications in Mental Health Dialogues |
2024-02-13T22:26:24Z | http://arxiv.org/pdf/2402.08831v1 | http://arxiv.org/abs/2402.08831v1 | Bo Peng, Xinyi Ling, Ziru Chen, Huan Sun, Xia Ning | 2024-02-13T22:26:24Z | cs.CL, cs.AI, cs.IR | With tremendous efforts on developing effective e-commerce models,
conventional e-commerce models show limited success in generalist e-commerce
modeling, and suffer from unsatisfactory performance on new users and new
products - a typical out-of-domain generalization challenge. Meanwhile, large
language models (LLMs) demonstrate outstanding performance in generalist
modeling and out-of-domain generalizability in many fields. Toward fully
unleashing their power for e-commerce, in this paper, we construct ECInstruct,
the first open-sourced, large-scale, and high-quality benchmark instruction
dataset for e-commerce. Leveraging ECInstruct, we develop eCeLLM, a series of
e-commerce LLMs, by instruction-tuning general-purpose LLMs. Our comprehensive
experiments and evaluation demonstrate that eCeLLM models substantially
outperform baseline models, including the most advanced GPT-4, and the
state-of-the-art task-specific models in in-domain evaluation. Moreover, eCeLLM
exhibits excellent generalizability to out-of-domain settings, including unseen
products and unseen instructions, highlighting its superiority as a generalist
e-commerce model. Both the ECInstruct dataset and the eCeLLM models show great
potential in empowering versatile and effective LLMs for e-commerce. ECInstruct
and eCeLLM models are publicly accessible through
https://ninglab.github.io/eCeLLM. | cs.CL | eCeLLM: Generalizing Large Language Models for E-commerce from Large-scale, High-quality Instruction Data |
2024-02-13T22:22:51Z | http://arxiv.org/pdf/2402.08830v1 | http://arxiv.org/abs/2402.08830v1 | Sammy Khalife, Yann Ponty, Laurent Bulteau | 2024-02-13T22:22:51Z | cs.DS, cs.CC, cs.CL | Several popular language models represent local contexts in an input text as
bags of words. Such representations are naturally encoded by a sequence graph
whose vertices are the distinct words occurring in x, with edges representing
the (ordered) co-occurrence of two words within a sliding window of size w.
However, this compressed representation is not generally bijective, and may
introduce some degree of ambiguity. Some sequence graphs may admit several
realizations as a sequence, while others may not admit any realization. In this
paper, we study the realizability and ambiguity of sequence graphs from a
combinatorial and computational point of view. We consider the existence and
enumeration of realizations of a sequence graph under multiple settings: window
size w, presence/absence of graph orientation, and presence/absence of weights
(multiplicities). When w = 2, we provide polynomial time algorithms for
realizability and enumeration in all cases except the undirected/weighted
setting, where we show the #P-hardness of enumeration. For a window of size at
least 3, we prove hardness of all variants, even when w is considered as a
constant, with the notable exception of the undirected/unweighted case for
which we propose an XP algorithms for both (realizability and enumeration)
problems, tight due to a corresponding W[1]-hardness result. We conclude with
an integer program formulation to solve the realizability problem, and with
dynamic programming to solve the enumeration problem. This work leaves open the
membership to NP for both problems, a non-trivial question due to the existence
of minimum realizations having exponential size on the instance encoding. | cs.DS | Sequence graphs realizations and ambiguity in language models |
2024-02-13T20:54:24Z | http://arxiv.org/pdf/2402.08788v1 | http://arxiv.org/abs/2402.08788v1 | Timothy Wong, Claire Li, Sam Lam, Billy Chiu, Qin Lu, Minglei Li, Dan Xiong, Roy Shing Yu, Vincent T. Y. Ng | 2024-02-13T20:54:24Z | cs.CL, cs.SD, eess.AS, 94-06, I.2.7 | This paper reports our work on building up a Cantonese Speech-to-Text (STT)
system with a syllable based acoustic model. This is a part of an effort in
building a STT system to aid dyslexic students who have cognitive deficiency in
writing skills but have no problem expressing their ideas through speech. For
Cantonese speech recognition, the basic unit of acoustic models can either be
the conventional Initial-Final (IF) syllables, or the Onset-Nucleus-Coda (ONC)
syllables where finals are further split into nucleus and coda to reflect the
intra-syllable variations in Cantonese. By using the Kaldi toolkit, our system
is trained using the stochastic gradient descent optimization model with the
aid of GPUs for the hybrid Deep Neural Network and Hidden Markov Model
(DNN-HMM) with and without I-vector based speaker adaptive training technique.
The input features of the same Gaussian Mixture Model with speaker adaptive
training (GMM-SAT) to DNN are used in all cases. Experiments show that the
ONC-based syllable acoustic modeling with I-vector based DNN-HMM achieves the
best performance with the word error rate (WER) of 9.66% and the real time
factor (RTF) of 1.38812. | cs.CL | Syllable based DNN-HMM Cantonese Speech to Text System |
2024-02-13T20:51:58Z | http://arxiv.org/pdf/2402.08787v4 | http://arxiv.org/abs/2402.08787v4 | Sijia Liu, Yuanshun Yao, Jinghan Jia, Stephen Casper, Nathalie Baracaldo, Peter Hase, Yuguang Yao, Chris Yuhao Liu, Xiaojun Xu, Hang Li, Kush R. Varshney, Mohit Bansal, Sanmi Koyejo, Yang Liu | 2024-07-02T02:15:50Z | cs.LG, cs.CL | We explore machine unlearning (MU) in the domain of large language models
(LLMs), referred to as LLM unlearning. This initiative aims to eliminate
undesirable data influence (e.g., sensitive or illegal information) and the
associated model capabilities, while maintaining the integrity of essential
knowledge generation and not affecting causally unrelated information. We
envision LLM unlearning becoming a pivotal element in the life-cycle management
of LLMs, potentially standing as an essential foundation for developing
generative AI that is not only safe, secure, and trustworthy, but also
resource-efficient without the need of full retraining. We navigate the
unlearning landscape in LLMs from conceptual formulation, methodologies,
metrics, and applications. In particular, we highlight the often-overlooked
aspects of existing LLM unlearning research, e.g., unlearning scope, data-model
interaction, and multifaceted efficacy assessment. We also draw connections
between LLM unlearning and related areas such as model editing, influence
functions, model explanation, adversarial training, and reinforcement learning.
Furthermore, we outline an effective assessment framework for LLM unlearning
and explore its applications in copyright and privacy safeguards and
sociotechnical harm reduction. | cs.LG | Rethinking Machine Unlearning for Large Language Models |
2024-02-13T20:47:17Z | http://arxiv.org/pdf/2402.08785v1 | http://arxiv.org/abs/2402.08785v1 | Jianing Wang, Junda Wu, Yupeng Hou, Yao Liu, Ming Gao, Julian McAuley | 2024-02-13T20:47:17Z | cs.CL | Do current large language models (LLMs) better solve graph reasoning and
generation tasks with parameter updates? In this paper, we propose
InstructGraph, a framework that empowers LLMs with the abilities of graph
reasoning and generation by instruction tuning and preference alignment.
Specifically, we first propose a structured format verbalizer to unify all
graph data into a universal code-like format, which can simply represent the
graph without any external graph-specific encoders. Furthermore, a graph
instruction tuning stage is introduced to guide LLMs in solving graph reasoning
and generation tasks. Finally, we identify potential hallucination problems in
graph tasks and sample negative instances for preference alignment, the target
of which is to enhance the output's reliability of the model. Extensive
experiments across multiple graph-centric tasks exhibit that InstructGraph can
achieve the best performance and outperform GPT-4 and LLaMA2 by more than 13\%
and 38\%, respectively. | cs.CL | InstructGraph: Boosting Large Language Models via Graph-centric Instruction Tuning and Preference Alignment |
2024-02-13T20:21:29Z | http://arxiv.org/pdf/2402.08777v2 | http://arxiv.org/abs/2402.08777v2 | Zhihan Zhou, Weimin Wu, Harrison Ho, Jiayi Wang, Lizhen Shi, Ramana V Davuluri, Zhong Wang, Han Liu | 2024-02-15T04:55:23Z | q-bio.GN, cs.AI, cs.CE, cs.CL | Effective DNA embedding remains crucial in genomic analysis, particularly in
scenarios lacking labeled data for model fine-tuning, despite the significant
advancements in genome foundation models. A prime example is metagenomics
binning, a critical process in microbiome research that aims to group DNA
sequences by their species from a complex mixture of DNA sequences derived from
potentially thousands of distinct, often uncharacterized species. To fill the
lack of effective DNA embedding models, we introduce DNABERT-S, a genome
foundation model that specializes in creating species-aware DNA embeddings. To
encourage effective embeddings to error-prone long-read DNA sequences, we
introduce Manifold Instance Mixup (MI-Mix), a contrastive objective that mixes
the hidden representations of DNA sequences at randomly selected layers and
trains the model to recognize and differentiate these mixed proportions at the
output layer. We further enhance it with the proposed Curriculum Contrastive
Learning (C$^2$LR) strategy. Empirical results on 18 diverse datasets showed
DNABERT-S's remarkable performance. It outperforms the top baseline's
performance in 10-shot species classification with just a 2-shot training while
doubling the Adjusted Rand Index (ARI) in species clustering and substantially
increasing the number of correctly identified species in metagenomics binning.
The code, data, and pre-trained model are publicly available at
https://github.com/Zhihan1996/DNABERT_S. | q-bio.GN | DNABERT-S: Learning Species-Aware DNA Embedding with Genome Foundation Models |
2024-02-13T20:16:29Z | http://arxiv.org/pdf/2402.10963v2 | http://arxiv.org/abs/2402.10963v2 | Alex Havrilla, Sharath Raparthy, Christoforus Nalmpantis, Jane Dwivedi-Yu, Maksym Zhuravinskyi, Eric Hambro, Roberta Raileanu | 2024-06-25T03:14:10Z | cs.CL, cs.LG | State-of-the-art language models can exhibit impressive reasoning refinement
capabilities on math, science or coding tasks. However, recent work
demonstrates that even the best models struggle to identify \textit{when and
where to refine} without access to external feedback. Outcome-based Reward
Models (\textbf{ORMs}), trained to predict correctness of the final answer
indicating when to refine, offer one convenient solution for deciding when to
refine. Process Based Reward Models (\textbf{PRMs}), trained to predict
correctness of intermediate steps, can then be used to indicate where to
refine. But they are expensive to train, requiring extensive human annotations.
In this paper, we propose Stepwise ORMs (\textbf{SORMs}) which are trained,
only on synthetic data, to approximate the expected future reward of the
optimal policy or $V^{\star}$. More specifically, SORMs are trained to predict
the correctness of the final answer when sampling the current policy many times
(rather than only once as in the case of ORMs). Our experiments show that SORMs
can more accurately detect incorrect reasoning steps compared to ORMs, thus
improving downstream accuracy when doing refinements. We then train
\textit{global} refinement models, which take only the question and a draft
solution as input and predict a corrected solution, and \textit{local}
refinement models which also take as input a critique indicating the location
of the first reasoning error. We generate training data for both models
synthetically by reusing data used to train the SORM. We find combining global
and local refinements, using the ORM as a reranker, significantly outperforms
either one individually, as well as a best of three sample baseline. With this
strategy we can improve the accuracy of a LLaMA-2 13B model (already fine-tuned
with RL) on GSM8K from 53\% to 65\% when greedily sampled. | cs.CL | GLoRe: When, Where, and How to Improve LLM Reasoning via Global and Local Refinements |
2024-02-13T20:10:29Z | http://arxiv.org/pdf/2402.10962v3 | http://arxiv.org/abs/2402.10962v3 | Kenneth Li, Tianle Liu, Naomi Bashkansky, David Bau, Fernanda Viégas, Hanspeter Pfister, Martin Wattenberg | 2024-05-01T16:47:42Z | cs.CL, cs.AI, cs.LG | System-prompting is a standard tool for customizing language-model chatbots,
enabling them to follow a specific instruction. An implicit assumption in the
use of system prompts is that they will be stable, so the chatbot will continue
to generate text according to the stipulated instructions for the duration of a
conversation. We propose a quantitative benchmark to test this assumption,
evaluating instruction stability via self-chats between two instructed
chatbots. Testing popular models like LLaMA2-chat-70B and GPT-3.5, we reveal a
significant instruction drift within eight rounds of conversations. An
empirical and theoretical analysis of this phenomenon suggests the transformer
attention mechanism plays a role, due to attention decay over long exchanges.
To combat attention decay and instruction drift, we propose a lightweight
method called split-softmax, which compares favorably against two strong
baselines. | cs.CL | Measuring and Controlling Instruction (In)Stability in Language Model Dialogs |
2024-02-13T19:58:24Z | http://arxiv.org/pdf/2402.08764v1 | http://arxiv.org/abs/2402.08764v1 | Paul Engelmann, Peter Brunsgaard Trolle, Christian Hardmeier | 2024-02-13T19:58:24Z | cs.CL | Dehumanization is a mental process that enables the exclusion and ill
treatment of a group of people. In this paper, we present two data sets of
dehumanizing text, a large, automatically collected corpus and a smaller,
manually annotated data set. Both data sets include a combination of political
discourse and dialogue from movie subtitles. Our methods give us a broad and
varied amount of dehumanization data to work with, enabling further exploratory
analysis and automatic classification of dehumanization patterns. Both data
sets will be publicly released. | cs.CL | A Dataset for the Detection of Dehumanizing Language |
2024-02-13T19:54:29Z | http://arxiv.org/pdf/2402.08761v1 | http://arxiv.org/abs/2402.08761v1 | Jillian Fisher, Ximing Lu, Jaehun Jung, Liwei Jiang, Zaid Harchaoui, Yejin Choi | 2024-02-13T19:54:29Z | cs.CL, cs.AI | The permanence of online content combined with the enhanced authorship
identification techniques calls for stronger computational methods to protect
the identity and privacy of online authorship when needed, e.g., blind reviews
for scientific papers, anonymous online reviews, or anonymous interactions in
the mental health forums. In this paper, we propose an unsupervised
inference-time approach to authorship obfuscation to address the unique
challenges of authorship obfuscation: lack of supervision data for diverse
authorship and domains, and the need for a sufficient level of revision beyond
simple paraphrasing to obfuscate the authorship, all the while preserving the
original content and fluency.
We introduce JAMDEC, a user-controlled, inference-time algorithm for
authorship obfuscation that can be in principle applied to any text and
authorship. Our approach builds on small language models such as GPT2-XL in
order to help avoid disclosing the original content to proprietary LLM's APIs,
while also reducing the performance gap between small and large language models
via algorithmic enhancement. The key idea behind our approach is to boost the
creative power of smaller language models through constrained decoding, while
also allowing for user-specified controls and flexibility. Experimental results
demonstrate that our approach based on GPT2-XL outperforms previous
state-of-the-art methods based on comparably small models, while performing
competitively against GPT3.5 175B, a propriety model that is two orders of
magnitudes larger. | cs.CL | JAMDEC: Unsupervised Authorship Obfuscation using Constrained Decoding over Small Language Models |
2024-02-13T19:49:17Z | http://arxiv.org/pdf/2402.08756v1 | http://arxiv.org/abs/2402.08756v1 | Maurice Diesendruck, Jianzhe Lin, Shima Imani, Gayathri Mahalingam, Mingyang Xu, Jie Zhao | 2024-02-13T19:49:17Z | cs.CL, cs.CV | When LLMs perform zero-shot inference, they typically use a prompt with a
task specification, and generate a completion. However, there is no work to
explore the possibility of the reverse - going from completion to task
specification. In this paper, we employ both directions to perform
cycle-supervised learning entirely in-context. Our goal is to create a forward
map f : X -> Y (e.g. image -> generated caption), coupled with a backward map g
: Y -> X (e.g. caption -> generated image) to construct a cycle-consistency
"loss" (formulated as an update to the prompt) to enforce g(f(X)) ~= X. The
technique, called CyclePrompt, uses cycle-consistency as a free supervisory
signal to iteratively craft the prompt. Importantly, CyclePrompt reinforces
model performance without expensive fine-tuning, without training data, and
without the complexity of external environments (e.g. compilers, APIs). We
demonstrate CyclePrompt in two domains: code generation and image captioning.
Our results on the HumanEval coding benchmark put us in first place on the
leaderboard among models that do not rely on extra training data or usage of
external environments, and third overall. Compared to the GPT4 baseline, we
improve accuracy from 80.5% to 87.2%. In the vision-language space, we generate
detailed image captions which outperform baseline zero-shot GPT4V captions,
when tested against natural (VQAv2) and diagrammatic (FigureQA) visual
question-answering benchmarks. To the best of our knowledge, this is the first
use of self-supervised learning for prompting. | cs.CL | Learning How To Ask: Cycle-Consistency Refines Prompts in Multimodal Foundation Models |
2024-02-13T18:59:05Z | http://arxiv.org/pdf/2402.08680v1 | http://arxiv.org/abs/2402.08680v1 | Linxi Zhao, Yihe Deng, Weitong Zhang, Quanquan Gu | 2024-02-13T18:59:05Z | cs.LG, cs.AI, cs.CL, cs.CV | The advancement of Large Vision-Language Models (LVLMs) has increasingly
highlighted the critical issue of their tendency to hallucinate non-existing
objects in the images. To address this issue, previous works focused on using
specially curated datasets or powerful LLMs (e.g., GPT-3.5) to rectify the
outputs of LVLMs. However, these approaches require either expensive
training/fine-tuning or API access to advanced LLMs to correct the model's
output post-generation. In this paper, we tackle this challenge by introducing
a framework called Mitigating hallucinAtion via classifieR-Free guIdaNcE
(MARINE), which is both training-free and API-free, and can effectively and
efficiently reduce object hallucinations during the generation process.
Specifically, MARINE enriches the visual context of LVLMs by integrating
existing open-source vision models, and employs classifier-free guidance to
incorporate the additional object grounding features to improve the precision
of LVLMs' generations. Through comprehensive evaluations across $6$ popular
LVLMs with diverse evaluation metrics, we demonstrate the effectiveness of
MARINE, which even outperforms existing fine-tuning-based methods. Remarkably,
it not only reduces hallucinations but also improves the detailedness of LVLMs'
generations, as assessed by GPT-4V. | cs.LG | Mitigating Object Hallucination in Large Vision-Language Models via Classifier-Free Guidance |
2024-02-13T18:58:48Z | http://arxiv.org/pdf/2402.08679v2 | http://arxiv.org/abs/2402.08679v2 | Xingang Guo, Fangxu Yu, Huan Zhang, Lianhui Qin, Bin Hu | 2024-06-07T00:13:08Z | cs.LG, cs.AI, cs.CL | Jailbreaks on large language models (LLMs) have recently received increasing
attention. For a comprehensive assessment of LLM safety, it is essential to
consider jailbreaks with diverse attributes, such as contextual coherence and
sentiment/stylistic variations, and hence it is beneficial to study
controllable jailbreaking, i.e. how to enforce control on LLM attacks. In this
paper, we formally formulate the controllable attack generation problem, and
build a novel connection between this problem and controllable text generation,
a well-explored topic of natural language processing. Based on this connection,
we adapt the Energy-based Constrained Decoding with Langevin Dynamics (COLD), a
state-of-the-art, highly efficient algorithm in controllable text generation,
and introduce the COLD-Attack framework which unifies and automates the search
of adversarial LLM attacks under a variety of control requirements such as
fluency, stealthiness, sentiment, and left-right-coherence. The controllability
enabled by COLD-Attack leads to diverse new jailbreak scenarios which not only
cover the standard setting of generating fluent (suffix) attack with
continuation constraint, but also allow us to address new controllable attack
settings such as revising a user query adversarially with paraphrasing
constraint, and inserting stealthy attacks in context with position constraint.
Our extensive experiments on various LLMs (Llama-2, Mistral, Vicuna, Guanaco,
GPT-3.5, and GPT-4) show COLD-Attack's broad applicability, strong
controllability, high success rate, and attack transferability. Our code is
available at https://github.com/Yu-Fangxu/COLD-Attack. | cs.LG | COLD-Attack: Jailbreaking LLMs with Stealthiness and Controllability |
2024-02-13T18:48:23Z | http://arxiv.org/pdf/2402.08666v1 | http://arxiv.org/abs/2402.08666v1 | Irina Saparina, Mirella Lapata | 2024-02-13T18:48:23Z | cs.CL | Text-to-SQL semantic parsing has made significant progress in recent years,
with various models demonstrating impressive performance on the challenging
Spider benchmark. However, it has also been shown that these models often
struggle to generalize even when faced with small perturbations of previously
(accurately) parsed expressions. This is mainly due to the linguistic form of
questions in Spider which are overly specific, unnatural, and display limited
variation. In this work, we use data augmentation to enhance the robustness of
text-to-SQL parsers against natural language variations. Existing approaches
generate question reformulations either via models trained on Spider or only
introduce local changes. In contrast, we leverage the capabilities of large
language models to generate more realistic and diverse questions. Using only a
few prompts, we achieve a two-fold increase in the number of questions in
Spider. Training on this augmented dataset yields substantial improvements on a
range of evaluation sets, including robustness benchmarks and out-of-domain
data. | cs.CL | Improving Generalization in Semantic Parsing by Increasing Natural Language Variation |
2024-02-13T18:24:08Z | http://arxiv.org/pdf/2402.08644v3 | http://arxiv.org/abs/2402.08644v3 | Aishwarya P S, Pranav Ajit Nair, Yashas Samaga, Toby Boyd, Sanjiv Kumar, Prateek Jain, Praneeth Netrapalli | 2024-03-26T11:13:56Z | cs.AI, cs.CL | The autoregressive nature of conventional large language models (LLMs)
inherently limits inference speed, as tokens are generated sequentially. While
speculative and parallel decoding techniques attempt to mitigate this, they
face limitations: either relying on less accurate smaller models for generation
or failing to fully leverage the base LLM's representations.
We introduce a novel architecture, Tandem transformers, to address these
issues. This architecture uniquely combines (1) a small autoregressive model
and (2) a large model operating in block mode (processing multiple tokens
simultaneously). The small model's predictive accuracy is substantially
enhanced by granting it attention to the large model's richer representations.
On the PaLM2 pretraining dataset, a tandem of PaLM2-Bison and PaLM2-Gecko
demonstrates a 3.3% improvement in next-token prediction accuracy over a
standalone PaLM2-Gecko, offering a 1.16x speedup compared to a PaLM2-Otter
model with comparable downstream performance. We further incorporate the tandem
model within the speculative decoding (SPEED) framework where the large model
validates tokens from the small model. This ensures that the Tandem of
PaLM2-Bison and PaLM2-Gecko achieves substantial speedup (around 1.14x faster
than using vanilla PaLM2-Gecko in SPEED) while maintaining identical downstream
task accuracy. | cs.AI | Tandem Transformers for Inference Efficient LLMs |
2024-02-13T18:04:53Z | http://arxiv.org/pdf/2402.08638v5 | http://arxiv.org/abs/2402.08638v5 | Nedjma Ousidhoum, Shamsuddeen Hassan Muhammad, Mohamed Abdalla, Idris Abdulmumin, Ibrahim Said Ahmad, Sanchit Ahuja, Alham Fikri Aji, Vladimir Araujo, Abinew Ali Ayele, Pavan Baswani, Meriem Beloucif, Chris Biemann, Sofia Bourhim, Christine De Kock, Genet Shanko Dekebo, Oumaima Hourrane, Gopichand Kanumolu, Lokesh Madasu, Samuel Rutunda, Manish Shrivastava, Thamar Solorio, Nirmal Surange, Hailegnaw Getaneh Tilaye, Krishnapriya Vishnubhotla, Genta Winata, Seid Muhie Yimam, Saif M. Mohammad | 2024-05-31T15:57:58Z | cs.CL | Exploring and quantifying semantic relatedness is central to representing
language and holds significant implications across various NLP tasks. While
earlier NLP research primarily focused on semantic similarity, often within the
English language context, we instead investigate the broader phenomenon of
semantic relatedness. In this paper, we present \textit{SemRel}, a new semantic
relatedness dataset collection annotated by native speakers across 13
languages: \textit{Afrikaans, Algerian Arabic, Amharic, English, Hausa, Hindi,
Indonesian, Kinyarwanda, Marathi, Moroccan Arabic, Modern Standard Arabic,
Spanish,} and \textit{Telugu}. These languages originate from five distinct
language families and are predominantly spoken in Africa and Asia -- regions
characterised by a relatively limited availability of NLP resources. Each
instance in the SemRel datasets is a sentence pair associated with a score that
represents the degree of semantic textual relatedness between the two
sentences. The scores are obtained using a comparative annotation framework. We
describe the data collection and annotation processes, challenges when building
the datasets, baseline experiments, and their impact and utility in NLP. | cs.CL | SemRel2024: A Collection of Semantic Textual Relatedness Datasets for 13 Languages |
2024-02-13T17:59:34Z | http://arxiv.org/pdf/2402.08631v2 | http://arxiv.org/abs/2402.08631v2 | Xiaoshuai Song, Zhengyang Wang, Keqing He, Guanting Dong, Yutao Mou, Jinxu Zhao, Weiran Xu | 2024-02-17T16:06:10Z | cs.CL, cs.AI, cs.LG | Knowledge editing (KE) aims to efficiently and precisely modify the behavior
of large language models (LLMs) to update specific knowledge without negatively
influencing other knowledge. Current research primarily focuses on white-box
LLMs editing, overlooking an important scenario: black-box LLMs editing, where
LLMs are accessed through interfaces and only textual output is available. In
this paper, we first officially introduce KE on black-box LLMs and then propose
a comprehensive evaluation framework to overcome the limitations of existing
evaluations that are not applicable to black-box LLMs editing and lack
comprehensiveness. To tackle privacy leaks of editing data and style
over-editing in current methods, we introduce a novel postEdit framework,
resolving privacy concerns through downstream post-processing and maintaining
textual style consistency via fine-grained editing to original responses.
Experiments and analysis on two benchmarks demonstrate that postEdit
outperforms all baselines and achieves strong generalization, especially with
huge improvements on style retention (average $+20.82\%\uparrow$). | cs.CL | Knowledge Editing on Black-box Large Language Models |
2024-02-13T16:57:02Z | http://arxiv.org/pdf/2402.08594v1 | http://arxiv.org/abs/2402.08594v1 | Haeju Lee, Minchan Jeong, Se-Young Yun, Kee-Eung Kim | 2024-02-13T16:57:02Z | cs.CL | Prompt tuning, in which prompts are optimized to adapt large-scale
pre-trained language models to downstream tasks instead of fine-tuning the full
model parameters, has been shown to be particularly effective when the prompts
are trained in a multi-task transfer learning setting. These methods generally
involve individually training prompts for each source task and then aggregating
them to provide the initialization of the prompt for the target task. However,
this approach critically ignores the fact that some of the source tasks could
be negatively or positively interfering with each other. We argue that when we
extract knowledge from source tasks via training source prompts, we need to
consider this correlation among source tasks for better transfer to target
tasks. To this end, we propose a Bayesian approach where we work with the
posterior distribution of prompts across source tasks. We obtain representative
source prompts corresponding to the samples from the posterior utilizing Stein
Variational Gradient Descent, which are then aggregated to constitute the
initial target prompt. We show extensive experimental results on the standard
benchmark NLP tasks, where our Bayesian multi-task transfer learning approach
outperforms the state-of-the-art methods in many settings. Furthermore, our
approach requires no auxiliary models other than the prompt itself, achieving a
high degree of parameter efficiency. | cs.CL | Bayesian Multi-Task Transfer Learning for Soft Prompt Tuning |
2024-02-13T16:38:01Z | http://arxiv.org/pdf/2402.08702v3 | http://arxiv.org/abs/2402.08702v3 | Yongchao Chen, Jacob Arkin, Yilun Hao, Yang Zhang, Nicholas Roy, Chuchu Fan | 2024-06-16T18:01:06Z | cs.CL, cs.AI, cs.HC, cs.RO | Prompt optimization aims to find the best prompt to a large language model
(LLM) for a given task. LLMs have been successfully used to help find and
improve prompt candidates for single-step tasks. However, realistic tasks for
agents are multi-step and introduce new challenges: (1) Prompt content is
likely to be more extensive and complex, making it more difficult for LLMs to
analyze errors, (2) the impact of an individual step is difficult to evaluate,
and (3) different people may have varied preferences about task execution.
While humans struggle to optimize prompts, they are good at providing feedback
about LLM outputs; we therefore introduce a new LLM-driven discrete prompt
optimization framework PROMST that incorporates human-designed feedback rules
to automatically offer direct suggestions for improvement. We also use an extra
learned heuristic model that predicts prompt performance to efficiently sample
from prompt candidates. This approach significantly outperforms both
human-engineered prompts and several other prompt optimization methods across
11 representative multi-step tasks (an average 10.6\%-29.3\% improvement to
current best methods on five LLMs respectively). We believe our work can serve
as a benchmark for automatic prompt optimization for LLM-driven multi-step
tasks. Datasets and Codes are available at
https://github.com/yongchao98/PROMST. Project Page is available at
https://yongchao98.github.io/MIT-REALM-PROMST/. | cs.CL | PRompt Optimization in Multi-Step Tasks (PROMST): Integrating Human Feedback and Heuristic-based Sampling |
2024-02-13T16:35:48Z | http://arxiv.org/pdf/2402.08581v1 | http://arxiv.org/abs/2402.08581v1 | Yiyang Li, Lei Li, Dingxin Hu, Xueyi Hao, Marina Litvak, Natalia Vanetik, Yanquan Zhou | 2024-02-13T16:35:48Z | cs.CL | Improving factual consistency in abstractive summarization has been a focus
of current research. One promising approach is the post-editing method.
However, previous works have yet to make sufficient use of factual factors in
summaries and suffers from the negative effect of the training datasets. In
this paper, we first propose a novel factual error correction model FactCloze
based on a conditional-generation cloze task. FactCloze can construct the
causality among factual factors while being able to determine whether the blank
can be answered or not. Then, we propose a data distillation method to generate
a more faithful summarization dataset SummDSC via multiple-dimensional
evaluation. We experimentally validate the effectiveness of our approach, which
leads to an improvement in multiple factual consistency metrics compared to
baselines. | cs.CL | Improving Factual Error Correction for Abstractive Summarization via Data Distillation and Conditional-generation Cloze |
2024-02-13T16:28:28Z | http://arxiv.org/pdf/2402.08577v1 | http://arxiv.org/abs/2402.08577v1 | Dong Lu, Tianyu Pang, Chao Du, Qian Liu, Xianjun Yang, Min Lin | 2024-02-13T16:28:28Z | cs.CL, cs.CR, cs.CV, cs.LG, cs.MM | Backdoor attacks are commonly executed by contaminating training data, such
that a trigger can activate predetermined harmful effects during the test
phase. In this work, we present AnyDoor, a test-time backdoor attack against
multimodal large language models (MLLMs), which involves injecting the backdoor
into the textual modality using adversarial test images (sharing the same
universal perturbation), without requiring access to or modification of the
training data. AnyDoor employs similar techniques used in universal adversarial
attacks, but distinguishes itself by its ability to decouple the timing of
setup and activation of harmful effects. In our experiments, we validate the
effectiveness of AnyDoor against popular MLLMs such as LLaVA-1.5, MiniGPT-4,
InstructBLIP, and BLIP-2, as well as provide comprehensive ablation studies.
Notably, because the backdoor is injected by a universal perturbation, AnyDoor
can dynamically change its backdoor trigger prompts/harmful effects, exposing a
new challenge for defending against backdoor attacks. Our project page is
available at https://sail-sg.github.io/AnyDoor/. | cs.CL | Test-Time Backdoor Attacks on Multimodal Large Language Models |
2024-02-13T16:06:17Z | http://arxiv.org/pdf/2402.08567v2 | http://arxiv.org/abs/2402.08567v2 | Xiangming Gu, Xiaosen Zheng, Tianyu Pang, Chao Du, Qian Liu, Ye Wang, Jing Jiang, Min Lin | 2024-06-03T14:15:03Z | cs.CL, cs.CR, cs.CV, cs.LG, cs.MA | A multimodal large language model (MLLM) agent can receive instructions,
capture images, retrieve histories from memory, and decide which tools to use.
Nonetheless, red-teaming efforts have revealed that adversarial images/prompts
can jailbreak an MLLM and cause unaligned behaviors. In this work, we report an
even more severe safety issue in multi-agent environments, referred to as
infectious jailbreak. It entails the adversary simply jailbreaking a single
agent, and without any further intervention from the adversary, (almost) all
agents will become infected exponentially fast and exhibit harmful behaviors.
To validate the feasibility of infectious jailbreak, we simulate multi-agent
environments containing up to one million LLaVA-1.5 agents, and employ
randomized pair-wise chat as a proof-of-concept instantiation for multi-agent
interaction. Our results show that feeding an (infectious) adversarial image
into the memory of any randomly chosen agent is sufficient to achieve
infectious jailbreak. Finally, we derive a simple principle for determining
whether a defense mechanism can provably restrain the spread of infectious
jailbreak, but how to design a practical defense that meets this principle
remains an open question to investigate. Our project page is available at
https://sail-sg.github.io/Agent-Smith/. | cs.CL | Agent Smith: A Single Image Can Jailbreak One Million Multimodal LLM Agents Exponentially Fast |
2024-02-13T16:04:21Z | http://arxiv.org/pdf/2402.08562v1 | http://arxiv.org/abs/2402.08562v1 | Chongyang Gao, Kezhen Chen, Jinmeng Rao, Baochen Sun, Ruibo Liu, Daiyi Peng, Yawen Zhang, Xiaoyuan Guo, Jie Yang, VS Subrahmanian | 2024-02-13T16:04:21Z | cs.CL, cs.AI | Parameter-efficient tuning (PEFT) techniques like low-rank adaptation (LoRA)
offer training efficiency on Large Language Models, but their impact on model
performance remains limited. Recent efforts integrate LoRA and
Mixture-of-Experts (MoE) to improve the performance of PEFT methods. Despite
promising results, research on improving the efficiency of LoRA with MoE is
still in its early stages. Recent studies have shown that experts in the MoE
architecture have different strengths and also exhibit some redundancy. Does
this statement also apply to parameter-efficient MoE? In this paper, we
introduce a novel parameter-efficient MoE method,
\textit{\textbf{M}oE-L\textbf{o}RA with \textbf{L}ayer-wise Expert
\textbf{A}llocation (MoLA)} for Transformer-based models, where each model
layer has the flexibility to employ a varying number of LoRA experts. We
investigate several architectures with varying layer-wise expert
configurations. Experiments on six well-known NLP and commonsense QA benchmarks
demonstrate that MoLA achieves equal or superior performance compared to all
baselines. We find that allocating more LoRA experts to higher layers further
enhances the effectiveness of models with a certain number of experts in total.
With much fewer parameters, this allocation strategy outperforms the setting
with the same number of experts in every layer. This work can be widely used as
a plug-and-play parameter-efficient tuning approach for various applications.
The code is available at https://github.com/GCYZSL/MoLA. | cs.CL | Higher Layers Need More LoRA Experts |
2024-02-13T15:29:50Z | http://arxiv.org/pdf/2402.08526v3 | http://arxiv.org/abs/2402.08526v3 | Junhao Zheng, Shengjie Qiu, Qianli Ma | 2024-06-18T06:56:44Z | cs.LG, cs.CL | Large Language Models (LLMs) have achieved remarkable success across various
tasks, yet their ability to learn incrementally without forgetting remains
underexplored. Incremental learning (IL) is crucial as it enables models to
acquire new knowledge while retaining previously learned information, akin to
human learning. Existing benchmarks for IL are insufficient due to data leakage
issues and the overqualification of LLMs. To address these challenges, we
introduce Concept-1K, a novel dataset comprising 1,023 recently emerged
concepts across diverse domains. The concepts in Concept-1K are discrete,
interpretable units of knowledge that allow for fine-grained analysis of
learning and forgetting processes. Using Concept-1K as a testbed, we aim to
answer the question: ``Can LLMs learn new concepts incrementally without
forgetting like humans?'' Our investigation reveals that LLMs still suffer from
catastrophic forgetting and that LoRA, despite fine-tuning fewer parameters,
may lead to more forgetting on training data. Additionally, we explore the
roles of in-context learning, model scale, buffer size, and pretraining in IL
performance. These findings highlight the strengths and limitations of LLMs in
IL scenarios and provide a robust benchmark for future research. | cs.LG | Can LLMs Learn New Concepts Incrementally without Forgetting? |
2024-02-13T14:53:12Z | http://arxiv.org/pdf/2402.08498v4 | http://arxiv.org/abs/2402.08498v4 | Preetika Verma, Kokil Jaidka, Svetlana Churina | 2024-04-20T03:47:18Z | cs.CL | We audited large language models (LLMs) for their ability to create
evidence-based and stylistic counter-arguments to posts from the Reddit
ChangeMyView dataset. We benchmarked their rhetorical quality across a host of
qualitative and quantitative metrics and then ultimately evaluated them on
their persuasive abilities as compared to human counter-arguments. Our
evaluation is based on Counterfire: a new dataset of 32,000 counter-arguments
generated from large language models (LLMs): GPT-3.5 Turbo and Koala and their
fine-tuned variants, and PaLM 2, with varying prompts for evidence use and
argumentative style. GPT-3.5 Turbo ranked highest in argument quality with
strong paraphrasing and style adherence, particularly in `reciprocity' style
arguments. However, the stylistic counter-arguments still fall short of human
persuasive standards, where people also preferred reciprocal to evidence-based
rebuttals. The findings suggest that a balance between evidentiality and
stylistic elements is vital to a compelling counter-argument. We close with a
discussion of future research directions and implications for evaluating LLM
outputs. | cs.CL | Auditing Counterfire: Evaluating Advanced Counterargument Generation with Evidence and Style |
2024-02-13T14:51:45Z | http://arxiv.org/pdf/2402.08496v3 | http://arxiv.org/abs/2402.08496v3 | Chinonso Cynthia Osuji, Thiago Castro Ferreira, Brian Davis | 2024-02-27T00:05:28Z | cs.CL, cs.AI, cs.LG | This systematic review undertakes a comprehensive analysis of current
research on data-to-text generation, identifying gaps, challenges, and future
directions within the field. Relevant literature in this field on datasets,
evaluation metrics, application areas, multilingualism, language models, and
hallucination mitigation methods is reviewed. Various methods for producing
high-quality text are explored, addressing the challenge of hallucinations in
data-to-text generation. These methods include re-ranking, traditional and
neural pipeline architecture, planning architectures, data cleaning, controlled
generation, and modification of models and training techniques. Their
effectiveness and limitations are assessed, highlighting the need for
universally applicable strategies to mitigate hallucinations. The review also
examines the usage, popularity, and impact of datasets, alongside evaluation
metrics, with an emphasis on both automatic and human assessment. Additionally,
the evolution of data-to-text models, particularly the widespread adoption of
transformer models, is discussed. Despite advancements in text quality, the
review emphasizes the importance of research in low-resourced languages and the
engineering of datasets in these languages to promote inclusivity. Finally,
several application domains of data-to-text are highlighted, emphasizing their
relevance in such domains. Overall, this review serves as a guiding framework
for fostering innovation and advancing data-to-text generation. | cs.CL | A Systematic Review of Data-to-Text NLG |
2024-02-13T14:12:32Z | http://arxiv.org/pdf/2402.08479v4 | http://arxiv.org/abs/2402.08479v4 | Wei Jie Yeo, Ranjan Satapathy, Erik Cambria | 2024-02-25T16:25:45Z | cs.CL | The increasing use of complex and opaque black box models requires the
adoption of interpretable measures, one such option is extractive rationalizing
models, which serve as a more interpretable alternative. These models, also
known as Explain-Then-Predict models, employ an explainer model to extract
rationales and subsequently condition the predictor with the extracted
information. Their primary objective is to provide precise and faithful
explanations, represented by the extracted rationales. In this paper, we take a
semi-supervised approach to optimize for the plausibility of extracted
rationales. We adopt a pre-trained natural language inference (NLI) model and
further fine-tune it on a small set of supervised rationales ($10\%$). The NLI
predictor is leveraged as a source of supervisory signals to the explainer via
entailment alignment. We show that, by enforcing the alignment agreement
between the explanation and answer in a question-answering task, the
performance can be improved without access to ground truth labels. We evaluate
our approach on the ERASER dataset and show that our approach achieves
comparable results with supervised extractive models and outperforms
unsupervised approaches by $> 100\%$. | cs.CL | Plausible Extractive Rationalization through Semi-Supervised Entailment Signal |
2024-02-13T13:50:08Z | http://arxiv.org/pdf/2402.08467v1 | http://arxiv.org/abs/2402.08467v1 | Freddy Heppell, Mehmet E. Bakir, Kalina Bontcheva | 2024-02-13T13:50:08Z | cs.CL | As Large Language Models (LLMs) become more proficient, their misuse in
large-scale viral disinformation campaigns is a growing concern. This study
explores the capability of ChatGPT to generate unconditioned claims about the
war in Ukraine, an event beyond its knowledge cutoff, and evaluates whether
such claims can be differentiated by human readers and automated tools from
human-written ones. We compare war-related claims from ClaimReview, authored by
IFCN-registered fact-checkers, and similar short-form content generated by
ChatGPT. We demonstrate that ChatGPT can produce realistic, target-specific
disinformation cheaply, fast, and at scale, and that these claims cannot be
reliably distinguished by humans or existing automated tools. | cs.CL | Lying Blindly: Bypassing ChatGPT's Safeguards to Generate Hard-to-Detect Disinformation Claims at Scale |
2024-02-13T12:04:43Z | http://arxiv.org/pdf/2402.08403v4 | http://arxiv.org/abs/2402.08403v4 | Peter Wallis | 2024-05-08T10:07:05Z | cs.CL | Theory based AI research has had a hard time recently and the aim here is to
propose a model of what LLMs are actually doing when they impress us with their
language skills. The model integrates three established theories of human
decision-making from philosophy, sociology, and computer science. The paper
starts with the collective understanding of reasoning from the early days of AI
research - primarily because that model is how we humans think we think, and is
the most accessible. It then describes what is commonly thought of as "reactive
systems" which is the position taken by many philosophers and indeed many
contemporary AI researchers. The third component to the proposed model is from
sociology and, although not flattering to our modern ego, provides an
explanation to a puzzle that for many years has occupied those of us working on
conversational user interfaces. | cs.CL | LLMs and the Human Condition |
2024-02-13T11:37:30Z | http://arxiv.org/pdf/2402.08392v1 | http://arxiv.org/abs/2402.08392v1 | Chris Madge, Massimo Poesio | 2024-02-13T11:37:30Z | cs.CL | In this work we examine the use of Large Language Models (LLMs) in the
challenging setting of acting as a Minecraft agent. We apply and evaluate LLMs
in the builder and architect settings, introduce clarification questions and
examining the challenges and opportunities for improvement. In addition, we
present a platform for online interaction with the agents and an evaluation
against previous works. | cs.CL | Large Language Models as Minecraft Agents |
2024-02-13T11:22:52Z | http://arxiv.org/pdf/2402.08382v2 | http://arxiv.org/abs/2402.08382v2 | Junghyun Min, Minho Lee, Woochul Lee, Yeonsoo Lee | 2024-02-21T08:35:57Z | cs.CL | Unsupervised learning objectives like language modeling and de-noising
constitute a significant part in producing pre-trained models that perform
various downstream applications from natural language understanding to
conversational tasks. However, despite impressive generative capabilities of
recent large language models, their abilities to capture syntactic or semantic
structure within text lag behind. We hypothesize that the mismatch between
linguistic performance and competence in machines is attributable to
insufficient transfer of linguistic structure knowledge to computational
systems with currently popular pre-training objectives. We show that
punctuation restoration as a learning objective improves in- and
out-of-distribution performance on structure-related tasks like named entity
recognition, open information extraction, chunking, and part-of-speech tagging.
Punctuation restoration is an effective learning objective that can improve
structure understanding and yield a more robust structure-aware representations
of natural language. | cs.CL | Punctuation Restoration Improves Structure Understanding without Supervision |
2024-02-13T10:28:57Z | http://arxiv.org/pdf/2402.08349v2 | http://arxiv.org/abs/2402.08349v2 | Jonathan Fürst, Catherine Kosten, Farhard Nooralahzadeh, Yi Zhang, Kurt Stockinger | 2024-06-18T15:10:01Z | cs.DB, cs.AI, cs.CL | Text-to-SQL systems (also known as NL-to-SQL systems) have become an
increasingly popular solution for bridging the gap between user capabilities
and SQL-based data access. These systems translate user requests in natural
language to valid SQL statements for a specific database. Recent Text-to-SQL
systems have benefited from the rapid improvement of transformer-based language
models. However, while Text-to-SQL systems that incorporate such models
continuously reach new high scores on -- often synthetic -- benchmark datasets,
a systematic exploration of their robustness towards different data models in a
real-world, realistic scenario is notably missing. This paper provides the
first in-depth evaluation of the data model robustness of Text-to-SQL systems
in practice based on a multi-year international project focused on Text-to-SQL
interfaces. Our evaluation is based on a real-world deployment of FootballDB, a
system that was deployed over a 9 month period in the context of the FIFA World
Cup 2022, during which about 6K natural language questions were asked and
executed. All of our data is based on real user questions that were asked live
to the system. We manually labeled and translated a subset of these questions
for three different data models. For each data model, we explore the
performance of representative Text-to-SQL systems and language models. We
further quantify the impact of training data size, pre-, and post-processing
steps as well as language model inference time. Our comprehensive evaluation
sheds light on the design choices of real-world Text-to-SQL systems and their
impact on moving from research prototypes to real deployments. Last, we provide
a new benchmark dataset to the community, which is the first to enable the
evaluation of different data models for the same dataset and is substantially
more challenging than most previous datasets in terms of query complexity. | cs.DB | Evaluating the Data Model Robustness of Text-to-SQL Systems Based on Real User Queries |
2024-02-13T10:09:00Z | http://arxiv.org/pdf/2402.08341v2 | http://arxiv.org/abs/2402.08341v2 | Airlie Hilliard, Cristian Munoz, Zekun Wu, Adriano Soares Koshiyama | 2024-02-15T09:12:40Z | cs.CL, cs.AI | Large Language Models (LLMs) are increasingly being utilized by both
candidates and employers in the recruitment context. However, with this comes
numerous ethical concerns, particularly related to the lack of transparency in
these "black-box" models. Although previous studies have sought to increase the
transparency of these models by investigating the personality traits of LLMs,
many of the previous studies have provided them with personality assessments to
complete. On the other hand, this study seeks to obtain a better understanding
of such models by examining their output variations based on different input
prompts. Specifically, we use a novel elicitation approach using prompts
derived from common interview questions, as well as prompts designed to elicit
particular Big Five personality traits to examine whether the models were
susceptible to trait-activation like humans are, to measure their personality
based on the language used in their outputs. To do so, we repeatedly prompted
multiple LMs with different parameter sizes, including Llama-2, Falcon,
Mistral, Bloom, GPT, OPT, and XLNet (base and fine tuned versions) and examined
their personality using classifiers trained on the myPersonality dataset. Our
results reveal that, generally, all LLMs demonstrate high openness and low
extraversion. However, whereas LMs with fewer parameters exhibit similar
behaviour in personality traits, newer and LMs with more parameters exhibit a
broader range of personality traits, with increased agreeableness, emotional
stability, and openness. Furthermore, a greater number of parameters is
positively associated with openness and conscientiousness. Moreover, fine-tuned
models exhibit minor modulations in their personality traits, contingent on the
dataset. Implications and directions for future research are discussed. | cs.CL | Eliciting Personality Traits in Large Language Models |
2024-02-13T09:47:07Z | http://arxiv.org/pdf/2402.08327v2 | http://arxiv.org/abs/2402.08327v2 | Weizhe Lin, Jingbiao Mei, Jinghong Chen, Bill Byrne | 2024-06-05T11:46:23Z | cs.CL | Large Multimodal Models (LMMs) excel in natural language and visual
understanding but are challenged by exacting tasks such as Knowledge-based
Visual Question Answering (KB-VQA) which involve the retrieval of relevant
information from document collections to use in shaping answers to questions.
We present an extensive training and evaluation framework, M2KR, for KB-VQA.
M2KR contains a collection of vision and language tasks which we have
incorporated into a single suite of benchmark tasks for training and evaluating
general-purpose multi-modal retrievers. We use M2KR to develop PreFLMR, a
pre-trained version of the recently developed Fine-grained Late-interaction
Multi-modal Retriever (FLMR) approach to KB-VQA, and we report new
state-of-the-art results across a range of tasks. We also present
investigations into the scaling behaviors of PreFLMR intended to be useful in
future developments in general-purpose multi-modal retrievers. | cs.CL | PreFLMR: Scaling Up Fine-Grained Late-Interaction Multi-modal Retrievers |
2024-02-13T09:26:19Z | http://arxiv.org/pdf/2402.08318v3 | http://arxiv.org/abs/2402.08318v3 | Alba Morollon Diaz-Faes, Carla Sofia Ribeiro Murteira, Martin Ruskov | 2024-05-06T07:19:08Z | cs.CL, cs.CY, J.5; K.4.m | Looking at how social values are represented in fairy tales can give insights
about the variations in communication of values across cultures. We study how
values are communicated in fairy tales from Portugal, Italy and Germany using a
technique called word embedding with a compass to quantify vocabulary
differences and commonalities. We study how these three national traditions
differ in their explicit references to values. To do this, we specify a list of
value-charged tokens, consider their word stems and analyse the distance
between these in a bespoke pre-trained Word2Vec model. We triangulate and
critically discuss the validity of the resulting hypotheses emerging from this
quantitative model. Our claim is that this is a reusable and reproducible
method for the study of the values explicitly referenced in historical corpora.
Finally, our preliminary findings hint at a shared cultural understanding and
the expression of values such as Benevolence, Conformity, and Universalism
across the studied cultures, suggesting the potential existence of a
pan-European cultural memory. | cs.CL | Values That Are Explicitly Present in Fairy Tales: Comparing Samples from German, Italian and Portuguese Traditions |
2024-02-13T09:12:55Z | http://arxiv.org/pdf/2402.08309v2 | http://arxiv.org/abs/2402.08309v2 | Daniel Nahmias, Gal Engelberg, Dan Klein, Asaf Shabtai | 2024-02-14T08:10:38Z | cs.LG, cs.CL, cs.CR, I.2.7 | Spear-phishing attacks present a significant security challenge, with large
language models (LLMs) escalating the threat by generating convincing emails
and facilitating target reconnaissance. To address this, we propose a detection
approach based on a novel document vectorization method that utilizes an
ensemble of LLMs to create representation vectors. By prompting LLMs to reason
and respond to human-crafted questions, we quantify the presence of common
persuasion principles in the email's content, producing prompted contextual
document vectors for a downstream supervised machine learning model. We
evaluate our method using a unique dataset generated by a proprietary system
that automates target reconnaissance and spear-phishing email creation. Our
method achieves a 91% F1 score in identifying LLM-generated spear-phishing
emails, with the training set comprising only traditional phishing and benign
emails. Key contributions include an innovative document vectorization method
utilizing LLM reasoning, a publicly available dataset of high-quality
spear-phishing emails, and the demonstrated effectiveness of our method in
detecting such emails. This methodology can be utilized for various document
classification tasks, particularly in adversarial problem domains. | cs.LG | Prompted Contextual Vectors for Spear-Phishing Detection |
2024-02-13T09:06:14Z | http://arxiv.org/pdf/2402.08303v4 | http://arxiv.org/abs/2402.08303v4 | Yin Fang, Kangwei Liu, Ningyu Zhang, Xinle Deng, Penghui Yang, Zhuo Chen, Xiangru Tang, Mark Gerstein, Xiaohui Fan, Huajun Chen | 2024-02-20T02:26:39Z | cs.CL, cs.AI, cs.CE, cs.HC, cs.LG | As Large Language Models (LLMs) rapidly evolve, their influence in science is
becoming increasingly prominent. The emerging capabilities of LLMs in task
generalization and free-form dialogue can significantly advance fields like
chemistry and biology. However, the field of single-cell biology, which forms
the foundational building blocks of living organisms, still faces several
challenges. High knowledge barriers and limited scalability in current methods
restrict the full exploitation of LLMs in mastering single-cell data, impeding
direct accessibility and rapid iteration. To this end, we introduce ChatCell,
which signifies a paradigm shift by facilitating single-cell analysis with
natural language. Leveraging vocabulary adaptation and unified sequence
generation, ChatCell has acquired profound expertise in single-cell biology and
the capability to accommodate a diverse range of analysis tasks. Extensive
experiments further demonstrate ChatCell's robust performance and potential to
deepen single-cell insights, paving the way for more accessible and intuitive
exploration in this pivotal field. Our project homepage is available at
https://zjunlp.github.io/project/ChatCell. | cs.CL | ChatCell: Facilitating Single-Cell Analysis with Natural Language |
2024-02-13T08:12:48Z | http://arxiv.org/pdf/2402.08277v5 | http://arxiv.org/abs/2402.08277v5 | Tobias Schimanski, Jingwei Ni, Mathias Kraus, Elliott Ash, Markus Leippold | 2024-06-03T16:48:59Z | cs.CL, cs.LG | Advances towards more faithful and traceable answers of Large Language Models
(LLMs) are crucial for various research and practical endeavors. One avenue in
reaching this goal is basing the answers on reliable sources. However, this
Evidence-Based QA has proven to work insufficiently with LLMs in terms of
citing the correct sources (source quality) and truthfully representing the
information within sources (answer attributability). In this work, we
systematically investigate how to robustly fine-tune LLMs for better source
quality and answer attributability. Specifically, we introduce a data
generation pipeline with automated data quality filters, which can synthesize
diversified high-quality training and testing data at scale. We further
introduce four test sets to benchmark the robustness of fine-tuned specialist
models. Extensive evaluation shows that fine-tuning on synthetic data improves
performance on both in- and out-of-distribution. Furthermore, we show that data
quality, which can be drastically improved by proposed quality filters, matters
more than quantity in improving Evidence-Based QA. | cs.CL | Towards Faithful and Robust LLM Specialists for Evidence-Based Question-Answering |
2024-02-13T07:17:52Z | http://arxiv.org/pdf/2402.08259v1 | http://arxiv.org/abs/2402.08259v1 | Xuanliang Zhang, Dingzirui Wang, Longxu Dou, Qingfu Zhu, Wanxiang Che | 2024-02-13T07:17:52Z | cs.CL | Table reasoning, which aims to generate the corresponding answer to the
question following the user requirement according to the provided table, and
optionally a text description of the table, effectively improving the
efficiency of obtaining information. Recently, using Large Language Models
(LLMs) has become the mainstream method for table reasoning, because it not
only significantly reduces the annotation cost but also exceeds the performance
of previous methods. However, existing research still lacks a summary of
LLM-based table reasoning works. Due to the existing lack of research,
questions about which techniques can improve table reasoning performance in the
era of LLMs, why LLMs excel at table reasoning, and how to enhance table
reasoning abilities in the future, remain largely unexplored. This gap
significantly limits progress in research. To answer the above questions and
advance table reasoning research with LLMs, we present this survey to analyze
existing research, inspiring future work. In this paper, we analyze the
mainstream techniques used to improve table reasoning performance in the LLM
era, and the advantages of LLMs compared to pre-LLMs for solving table
reasoning. We provide research directions from both the improvement of existing
methods and the expansion of practical applications to inspire future research. | cs.CL | A Survey of Table Reasoning with Large Language Models |
2024-02-13T05:36:54Z | http://arxiv.org/pdf/2402.08227v1 | http://arxiv.org/abs/2402.08227v1 | Yixiang Yao, Fei Wang, Srivatsan Ravi, Muhao Chen | 2024-02-13T05:36:54Z | cs.CL | Language Models as a Service (LMaaS) offers convenient access for developers
and researchers to perform inference using pre-trained language models.
Nonetheless, the input data and the inference results containing private
information are exposed as plaintext during the service call, leading to
privacy issues. Recent studies have started tackling the privacy issue by
transforming input data into privacy-preserving representation from the
user-end with the techniques such as noise addition and content perturbation,
while the exploration of inference result protection, namely decision privacy,
is still a blank page. In order to maintain the black-box manner of LMaaS,
conducting data privacy protection, especially for the decision, is a
challenging task because the process has to be seamless to the models and
accompanied by limited communication and computation overhead. We thus propose
Instance-Obfuscated Inference (IOI) method, which focuses on addressing the
decision privacy issue of natural language understanding tasks in their
complete life-cycle. Besides, we conduct comprehensive experiments to evaluate
the performance as well as the privacy-protection strength of the proposed
method on various benchmarking tasks. | cs.CL | Privacy-Preserving Language Model Inference with Instance Obfuscation |
2024-02-13T05:15:46Z | http://arxiv.org/pdf/2402.08219v2 | http://arxiv.org/abs/2402.08219v2 | Haotian Sun, Yuchen Zhuang, Wei Wei, Chao Zhang, Bo Dai | 2024-05-28T17:04:45Z | cs.CL, cs.AI, cs.LG | Adapting state-of-the-art Large Language Models (LLMs) like GPT-4 and Gemini
for specific tasks is challenging. Due to the opacity in their parameters,
embeddings, and even output probabilities, existing fine-tuning adaptation
methods are inapplicable. Consequently, adapting these black-box LLMs is only
possible through their API services, raising concerns about transparency,
privacy, and cost. To address these challenges, we introduce BBox-Adapter, a
novel lightweight adapter for black-box LLMs. BBox-Adapter distinguishes target
and source domain data by treating target data as positive and source data as
negative. It employs a ranking-based Noise Contrastive Estimation (NCE) loss to
promote the likelihood of target domain data while penalizing that of the
source domain. Furthermore, it features an online adaptation mechanism, which
incorporates real-time positive data sampling from ground-truth, human, or AI
feedback, coupled with negative data from previous adaptations. Extensive
experiments demonstrate BBox-Adapter's effectiveness and cost efficiency. It
improves model performance by up to 6.77% across diverse tasks and domains,
while reducing training and inference costs by 31.30x and 1.84x, respectively. | cs.CL | BBox-Adapter: Lightweight Adapting for Black-Box Large Language Models |
2024-02-13T02:46:45Z | http://arxiv.org/pdf/2402.08183v1 | http://arxiv.org/abs/2402.08183v1 | Chenghao Xiao, Zhuoxu Huang, Danlu Chen, G Thomas Hudson, Yizhi Li, Haoran Duan, Chenghua Lin, Jie Fu, Jungong Han, Noura Al Moubayed | 2024-02-13T02:46:45Z | cs.CL, cs.CV | Pretrained language models are long known to be subpar in capturing sentence
and document-level semantics. Though heavily investigated, transferring
perturbation-based methods from unsupervised visual representation learning to
NLP remains an unsolved problem. This is largely due to the discreteness of
subword units brought by tokenization of language models, limiting small
perturbations of inputs to form semantics-preserved positive pairs. In this
work, we conceptualize the learning of sentence-level textual semantics as a
visual representation learning process. Drawing from cognitive and linguistic
sciences, we introduce an unsupervised visual sentence representation learning
framework, employing visually-grounded text perturbation methods like typos and
word order shuffling, resonating with human cognitive patterns, and enabling
perturbation to texts to be perceived as continuous. Our approach is further
bolstered by large-scale unsupervised topical alignment training and natural
language inference supervision, achieving comparable performance in semantic
textual similarity (STS) to existing state-of-the-art NLP methods.
Additionally, we unveil our method's inherent zero-shot cross-lingual
transferability and a unique leapfrogging pattern across languages during
iterative training. To our knowledge, this is the first representation learning
method devoid of traditional language models for understanding sentence and
document semantics, marking a stride closer to human-like textual
comprehension. Our code is available at
https://github.com/gowitheflow-1998/Pixel-Linguist | cs.CL | Pixel Sentence Representation Learning |
2024-02-13T01:31:08Z | http://arxiv.org/pdf/2402.08155v1 | http://arxiv.org/abs/2402.08155v1 | Lin Tian, Xiuzhen Zhang, Jey Han Lau | 2024-02-13T01:31:08Z | cs.CL, cs.AI | We apply causal mediation analysis to explain the decision-making process of
neural models for rumour detection on Twitter. Interventions at the input and
network level reveal the causal impacts of tweets and words in the model
output. We find that our approach CMA-R -- Causal Mediation Analysis for Rumour
detection -- identifies salient tweets that explain model predictions and show
strong agreement with human judgements for critical tweets determining the
truthfulness of stories. CMA-R can further highlight causally impactful words
in the salient tweets, providing another layer of interpretability and
transparency into these blackbox rumour detection systems. Code is available
at: https://github.com/ltian678/cma-r. | cs.CL | CMA-R:Causal Mediation Analysis for Explaining Rumour Detection |
2024-02-12T23:09:00Z | http://arxiv.org/pdf/2402.08114v2 | http://arxiv.org/abs/2402.08114v2 | William Muldrew, Peter Hayes, Mingtian Zhang, David Barber | 2024-06-28T08:22:01Z | cs.LG, cs.AI, cs.CL | As large language models (LLMs) become more capable, fine-tuning techniques
for aligning with human intent are increasingly important. A key consideration
for aligning these models is how to most effectively use human resources, or
model resources in the case where LLMs themselves are used as oracles.
Reinforcement learning from Human or AI preferences (RLHF/RLAIF) is the most
prominent example of such a technique, but is complex and often unstable.
Direct Preference Optimization (DPO) has recently been proposed as a simpler
and more stable alternative. In this work, we develop an active learning
strategy for DPO to make better use of preference labels. We propose a
practical acquisition function for prompt/completion pairs based on the
predictive entropy of the language model and a measure of certainty of the
implicit preference model optimized by DPO. We demonstrate how our approach
improves both the rate of learning and final performance of fine-tuning on
pairwise preference data. | cs.LG | Active Preference Learning for Large Language Models |
2024-02-12T23:08:37Z | http://arxiv.org/pdf/2402.08113v3 | http://arxiv.org/abs/2402.08113v3 | Samuel Schmidgall, Carl Harris, Ime Essien, Daniel Olshvang, Tawsifur Rahman, Ji Woong Kim, Rojin Ziaei, Jason Eshraghian, Peter Abadir, Rama Chellappa | 2024-02-20T23:45:43Z | cs.CL, cs.HC | There is increasing interest in the application large language models (LLMs)
to the medical field, in part because of their impressive performance on
medical exam questions. While promising, exam questions do not reflect the
complexity of real patient-doctor interactions. In reality, physicians'
decisions are shaped by many complex factors, such as patient compliance,
personal experience, ethical beliefs, and cognitive bias. Taking a step toward
understanding this, our hypothesis posits that when LLMs are confronted with
clinical questions containing cognitive biases, they will yield significantly
less accurate responses compared to the same questions presented without such
biases. In this study, we developed BiasMedQA, a benchmark for evaluating
cognitive biases in LLMs applied to medical tasks. Using BiasMedQA we evaluated
six LLMs, namely GPT-4, Mixtral-8x70B, GPT-3.5, PaLM-2, Llama 2 70B-chat, and
the medically specialized PMC Llama 13B. We tested these models on 1,273
questions from the US Medical Licensing Exam (USMLE) Steps 1, 2, and 3,
modified to replicate common clinically-relevant cognitive biases. Our analysis
revealed varying effects for biases on these LLMs, with GPT-4 standing out for
its resilience to bias, in contrast to Llama 2 70B-chat and PMC Llama 13B,
which were disproportionately affected by cognitive bias. Our findings
highlight the critical need for bias mitigation in the development of medical
LLMs, pointing towards safer and more reliable applications in healthcare. | cs.CL | Addressing cognitive bias in medical language models |
2024-02-12T22:47:57Z | http://arxiv.org/pdf/2402.10958v2 | http://arxiv.org/abs/2402.10958v2 | Yueqin Yin, Zhendong Wang, Yi Gu, Hai Huang, Weizhu Chen, Mingyuan Zhou | 2024-05-27T20:05:03Z | cs.CL, cs.AI, cs.LG | In the field of large language models (LLMs), aligning models with the
diverse preferences of users is a critical challenge. Direct Preference
Optimization (DPO) has played a key role in this area. It works by using pairs
of preferences derived from the same prompts, and it functions without needing
an additional reward model. However, DPO does not fully reflect the complex
nature of human learning, which often involves understanding contrasting
responses to not only identical but also similar questions. To overcome this
shortfall, we propose Relative Preference Optimization (RPO). RPO is designed
to discern between more and less preferred responses derived from both
identical and related prompts. It introduces a contrastive weighting mechanism,
enabling the tuning of LLMs using a broader range of preference data, including
both paired and unpaired sets. This approach expands the learning capabilities
of the model, allowing it to leverage insights from a more varied set of
prompts. Through empirical tests, including dialogue and summarization tasks,
and evaluations using the AlpacaEval2.0 leaderboard, RPO has demonstrated a
superior ability to align LLMs with user preferences and to improve their
adaptability during the training process. Our code can be viewed at
https://github.com/yinyueqin/relative-preference-optimization | cs.CL | Relative Preference Optimization: Enhancing LLM Alignment through Contrasting Responses across Identical and Diverse Prompts |
2024-02-12T22:35:40Z | http://arxiv.org/pdf/2402.08100v1 | http://arxiv.org/abs/2402.08100v1 | Federico Ranaldi, Elena Sofia Ruzzetti, Dario Onorati, Leonardo Ranaldi, Cristina Giannone, Andrea Favalli, Raniero Romagnoli, Fabio Massimo Zanzotto | 2024-02-12T22:35:40Z | cs.CL, cs.LG | Understanding textual description to generate code seems to be an achieved
capability of instruction-following Large Language Models (LLMs) in zero-shot
scenario. However, there is a severe possibility that this translation ability
may be influenced by having seen target textual descriptions and the related
code. This effect is known as Data Contamination.
In this study, we investigate the impact of Data Contamination on the
performance of GPT-3.5 in the Text-to-SQL code-generating tasks. Hence, we
introduce a novel method to detect Data Contamination in GPTs and examine
GPT-3.5's Text-to-SQL performances using the known Spider Dataset and our new
unfamiliar dataset Termite. Furthermore, we analyze GPT-3.5's efficacy on
databases with modified information via an adversarial table disconnection
(ATD) approach, complicating Text-to-SQL tasks by removing structural pieces of
information from the database. Our results indicate a significant performance
drop in GPT-3.5 on the unfamiliar Termite dataset, even with ATD modifications,
highlighting the effect of Data Contamination on LLMs in Text-to-SQL
translation tasks. | cs.CL | Investigating the Impact of Data Contamination of Large Language Models in Text-to-SQL Translation |
2024-02-12T22:21:30Z | http://arxiv.org/pdf/2402.08093v2 | http://arxiv.org/abs/2402.08093v2 | Mateusz Łajszczak, Guillermo Cámbara, Yang Li, Fatih Beyhan, Arent van Korlaar, Fan Yang, Arnaud Joly, Álvaro Martín-Cortinas, Ammar Abbas, Adam Michalski, Alexis Moinet, Sri Karlapati, Ewa Muszyńska, Haohan Guo, Bartosz Putrycz, Soledad López Gambino, Kayeon Yoo, Elena Sokolova, Thomas Drugman | 2024-02-15T18:57:26Z | cs.LG, cs.CL, eess.AS | We introduce a text-to-speech (TTS) model called BASE TTS, which stands for
$\textbf{B}$ig $\textbf{A}$daptive $\textbf{S}$treamable TTS with
$\textbf{E}$mergent abilities. BASE TTS is the largest TTS model to-date,
trained on 100K hours of public domain speech data, achieving a new
state-of-the-art in speech naturalness. It deploys a 1-billion-parameter
autoregressive Transformer that converts raw texts into discrete codes
("speechcodes") followed by a convolution-based decoder which converts these
speechcodes into waveforms in an incremental, streamable manner. Further, our
speechcodes are built using a novel speech tokenization technique that features
speaker ID disentanglement and compression with byte-pair encoding. Echoing the
widely-reported "emergent abilities" of large language models when trained on
increasing volume of data, we show that BASE TTS variants built with 10K+ hours
and 500M+ parameters begin to demonstrate natural prosody on textually complex
sentences. We design and share a specialized dataset to measure these emergent
abilities for text-to-speech. We showcase state-of-the-art naturalness of BASE
TTS by evaluating against baselines that include publicly available large-scale
text-to-speech systems: YourTTS, Bark and TortoiseTTS. Audio samples generated
by the model can be heard at https://amazon-ltts-paper.com/. | cs.LG | BASE TTS: Lessons from building a billion-parameter Text-to-Speech model on 100K hours of data |
2024-02-12T22:07:43Z | http://arxiv.org/pdf/2402.08086v2 | http://arxiv.org/abs/2402.08086v2 | Yun-Da Tsai, Ting-Yu Yen, Pei-Fu Guo, Zhe-Yan Li, Shou-De Lin | 2024-05-20T19:18:52Z | cs.LG, cs.CL, cs.CV | This research paper addresses the challenge of modality mismatch in
multimodal learning, where the modalities available during inference differ
from those available at training. We propose the Text-centric Alignment for
Multi-Modality Learning (TAMML) approach, an innovative method that utilizes
Large Language Models (LLMs) with in-context learning and foundation models to
enhance the generalizability of multimodal systems under these conditions. By
leveraging the unique properties of text as a unified semantic space, TAMML
demonstrates significant improvements in handling unseen, diverse, and
unpredictable modality combinations. TAMML not only adapts to varying
modalities but also maintains robust performance, showcasing the potential of
foundation models in overcoming the limitations of traditional fixed-modality
frameworks in embedding representations. This study contributes to the field by
offering a flexible, effective solution for real-world applications where
modality availability is dynamic and uncertain. | cs.LG | Text-centric Alignment for Multi-Modality Learning |
2024-02-12T21:44:32Z | http://arxiv.org/pdf/2402.08078v1 | http://arxiv.org/abs/2402.08078v1 | Yang Liu, Peng Sun, Hang Li | 2024-02-12T21:44:32Z | cs.CL, cs.LG | By formally defining the training processes of large language models (LLMs),
which usually encompasses pre-training, supervised fine-tuning, and
reinforcement learning with human feedback, within a single and unified machine
learning paradigm, we can glean pivotal insights for advancing LLM
technologies. This position paper delineates the parallels between the training
methods of LLMs and the strategies employed for the development of agents in
two-player games, as studied in game theory, reinforcement learning, and
multi-agent systems. We propose a re-conceptualization of LLM learning
processes in terms of agent learning in language-based games. This framework
unveils innovative perspectives on the successes and challenges in LLM
development, offering a fresh understanding of addressing alignment issues
among other strategic considerations. Furthermore, our two-player game approach
sheds light on novel data preparation and machine learning techniques for
training LLMs. | cs.CL | Large Language Models as Agents in Two-Player Games |
2024-02-12T21:14:45Z | http://arxiv.org/pdf/2402.08064v1 | http://arxiv.org/abs/2402.08064v1 | Jennifer Chu-Carroll, Andrew Beck, Greg Burnham, David OS Melville, David Nachman, A. Erdem Özcan, David Ferrucci | 2024-02-12T21:14:45Z | cs.AI, cs.CL | Since the advent of Large Language Models a few years ago, they have often
been considered the de facto solution for many AI problems. However, in
addition to the many deficiencies of LLMs that prevent them from broad industry
adoption, such as reliability, cost, and speed, there is a whole class of
common real world problems that Large Language Models perform poorly on,
namely, constraint satisfaction and optimization problems. These problems are
ubiquitous and current solutions are highly specialized and expensive to
implement. At Elemental Cognition, we developed our EC AI platform which takes
a neuro-symbolic approach to solving constraint satisfaction and optimization
problems. The platform employs, at its core, a precise and high performance
logical reasoning engine, and leverages LLMs for knowledge acquisition and user
interaction. This platform supports developers in specifying application logic
in natural and concise language while generating application user interfaces to
interact with users effectively. We evaluated LLMs against systems built on the
EC AI platform in three domains and found the EC AI systems to significantly
outperform LLMs on constructing valid and optimal solutions, on validating
proposed solutions, and on repairing invalid solutions. | cs.AI | Beyond LLMs: Advancing the Landscape of Complex Reasoning |
2024-02-12T19:35:37Z | http://arxiv.org/pdf/2402.08021v2 | http://arxiv.org/abs/2402.08021v2 | Allison Koenecke, Anna Seo Gyeong Choi, Katelyn X. Mei, Hilke Schellmann, Mona Sloane | 2024-05-03T02:18:20Z | cs.CL, cs.CY | Speech-to-text services aim to transcribe input audio as accurately as
possible. They increasingly play a role in everyday life, for example in
personal voice assistants or in customer-company interactions. We evaluate Open
AI's Whisper, a state-of-the-art automated speech recognition service
outperforming industry competitors, as of 2023. While many of Whisper's
transcriptions were highly accurate, we find that roughly 1\% of audio
transcriptions contained entire hallucinated phrases or sentences which did not
exist in any form in the underlying audio. We thematically analyze the
Whisper-hallucinated content, finding that 38\% of hallucinations include
explicit harms such as perpetuating violence, making up inaccurate
associations, or implying false authority. We then study why hallucinations
occur by observing the disparities in hallucination rates between speakers with
aphasia (who have a lowered ability to express themselves using speech and
voice) and a control group. We find that hallucinations disproportionately
occur for individuals who speak with longer shares of non-vocal durations -- a
common symptom of aphasia. We call on industry practitioners to ameliorate
these language-model-based hallucinations in Whisper, and to raise awareness of
potential biases amplified by hallucinations in downstream applications of
speech-to-text models. | cs.CL | Careless Whisper: Speech-to-Text Hallucination Harms |
2024-02-12T19:27:26Z | http://arxiv.org/pdf/2402.08017v2 | http://arxiv.org/abs/2402.08017v2 | Ashish Shenoy, Yichao Lu, Srihari Jayakumar, Debojeet Chatterjee, Mohsen Moslehpour, Pierce Chuang, Abhay Harpale, Vikas Bhardwaj, Di Xu, Shicong Zhao, Longfang Zhao, Ankit Ramchandani, Xin Luna Dong, Anuj Kumar | 2024-06-01T21:46:50Z | cs.CV, cs.CL, cs.LG | We introduce Lumos, the first end-to-end multimodal question-answering system
with text understanding capabilities. At the core of Lumos is a Scene Text
Recognition (STR) component that extracts text from first person point-of-view
images, the output of which is used to augment input to a Multimodal Large
Language Model (MM-LLM). While building Lumos, we encountered numerous
challenges related to STR quality, overall latency, and model inference. In
this paper, we delve into those challenges, and discuss the system
architecture, design choices, and modeling techniques employed to overcome
these obstacles. We also provide a comprehensive evaluation for each component,
showcasing high quality and efficiency. | cs.CV | Lumos : Empowering Multimodal LLMs with Scene Text Recognition |
2024-02-12T19:25:11Z | http://arxiv.org/pdf/2402.08015v5 | http://arxiv.org/abs/2402.08015v5 | Israel Abebe Azime, Atnafu Lambebo Tonja, Tadesse Destaw Belay, Mitiku Yohannes Fuge, Aman Kassahun Wassie, Eyasu Shiferaw Jada, Yonas Chanie, Walelign Tewabe Sewunetie, Seid Muhie Yimam | 2024-04-29T07:14:51Z | cs.CL | Large language models (LLMs) have received a lot of attention in natural
language processing (NLP) research because of their exceptional performance in
understanding and generating human languages. However, low-resource languages
are left behind due to the unavailability of resources. In this work, we focus
on enhancing the LLaMA-2-Amharic model by integrating task-specific and
generative datasets to improve language model performance for Amharic. We
compile an Amharic instruction fine-tuning dataset and fine-tuned
LLaMA-2-Amharic model. The fine-tuned model shows promising results in
different NLP tasks. We open-source our dataset creation pipeline, instruction
datasets, trained models, and evaluation outputs to promote language-specific
studies on these models. | cs.CL | Walia-LLM: Enhancing Amharic-LLaMA by Integrating Task-Specific and Generative Datasets |
2024-02-12T19:10:13Z | http://arxiv.org/pdf/2402.08005v1 | http://arxiv.org/abs/2402.08005v1 | Víctor Gallego | 2024-02-12T19:10:13Z | cs.CL, cs.LG | In this paper, we introduce \emph{refined Direct Preference Optimization}
(rDPO), a method for improving the behavioral alignment of Large Language
Models (LLMs) without the need for human-annotated data. The method involves
creating synthetic data using self-critique prompting by a teacher LLM and then
utilising a generalized DPO loss function to distil to a student LLM. The loss
function incorporates an additional external reward model to improve the
quality of synthetic data, making rDPO robust to potential noise in the
synthetic dataset. rDPO is shown to be effective in a diverse set of
behavioural alignment tasks, such as improved safety, robustness against
role-playing, and reduced sycophancy. Code to be released at
https://github.com/vicgalle/refined-dpo. | cs.CL | Refined Direct Preference Optimization with Synthetic Data for Behavioral Alignment of LLMs |
2024-02-12T19:10:11Z | http://arxiv.org/pdf/2402.16878v1 | http://arxiv.org/abs/2402.16878v1 | Johnathan Mercer | 2024-02-12T19:10:11Z | cs.AI, cs.CL, cs.LG, cs.NE | Formal mathematics is the discipline of translating mathematics into a
programming language in which any statement can be unequivocally checked by a
computer. Mathematicians and computer scientists have spent decades of
painstaking formalization efforts developing languages such as Coq, HOL, and
Lean. Machine learning research has converged on these formal math corpora and
given rise to an assortment of methodologies to aid in interactive and
automated theorem proving. However, these papers have primarily focused on one
method, for one proof task, in one language. This paper introduces EvoGPT-f: a
novel evolutionary framework for the first systematic quantitative analysis of
the differential machine learnability of five formal math corpora (Lean 3, Lean
4, Coq, HOL 4, HOL Light) using four tokenization methods (character,
word-level, Byte Pair Encoding and StarCoder tokenizer). This paper does not
put to rest the question of the "best" or "easiest" language to learn. Rather,
this framework and preliminary findings begin to illuminate the differential
machine learnability of these languages, offering a foundation to forge more
systematic quantitative and qualitative comparative research across
communities. | cs.AI | EvoGPT-f: An Evolutionary GPT Framework for Benchmarking Formal Math Languages |
2024-02-12T18:58:58Z | http://arxiv.org/pdf/2402.07899v2 | http://arxiv.org/abs/2402.07899v2 | Yulu Qin, Wentao Wang, Brenden M. Lake | 2024-05-10T18:54:59Z | cs.CL, cs.LG | Language models (LMs) have demonstrated remarkable proficiency in generating
linguistically coherent text, sparking discussions about their relevance to
understanding human language learnability. However, a significant gap exists
between the training data for these models and the linguistic input a child
receives. LMs are typically trained on data that is orders of magnitude larger
and fundamentally different from child-directed speech (Warstadt and Bowman,
2022; Warstadt et al., 2023; Frank, 2023a). Addressing this discrepancy, our
research focuses on training LMs on subsets of a single child's linguistic
input. Previously, Wang, Vong, Kim, and Lake (2023) found that LMs trained in
this setting can form syntactic and semantic word clusters and develop
sensitivity to certain linguistic phenomena, but they only considered LSTMs and
simpler neural networks trained from just one single-child dataset. Here, to
examine the robustness of learnability from single-child input, we
systematically train six different model architectures on five datasets (3
single-child and 2 baselines). We find that the models trained on single-child
datasets showed consistent results that matched with previous work,
underscoring the robustness of forming meaningful syntactic and semantic
representations from a subset of a child's linguistic input. | cs.CL | A systematic investigation of learnability from single child linguistic input |
2024-02-12T18:57:46Z | http://arxiv.org/pdf/2402.07896v2 | http://arxiv.org/abs/2402.07896v2 | Louis Castricato, Nathan Lile, Suraj Anand, Hailey Schoelkopf, Siddharth Verma, Stella Biderman | 2024-02-13T18:44:11Z | cs.CL | Existing methods for controlling language models, such as RLHF and
Constitutional AI, involve determining which LLM behaviors are desirable and
training them into a language model. However, in many cases, it is desirable
for LLMs to be controllable at inference time, so that they can be used in
multiple contexts with diverse needs. We illustrate this with the Pink Elephant
Problem: instructing an LLM to avoid discussing a certain entity (a ``Pink
Elephant''), and instead discuss a preferred entity (``Grey Elephant''). We
apply a novel simplification of Constitutional AI, Direct Principle Feedback,
which skips the ranking of responses and uses DPO directly on critiques and
revisions. Our results show that after DPF fine-tuning on our synthetic Pink
Elephants dataset, our 13B fine-tuned LLaMA 2 model significantly outperforms
Llama-2-13B-Chat and a prompted baseline, and performs as well as GPT-4 in on
our curated test set assessing the Pink Elephant Problem. | cs.CL | Suppressing Pink Elephants with Direct Principle Feedback |
2024-02-12T18:54:02Z | http://arxiv.org/pdf/2402.07891v3 | http://arxiv.org/abs/2402.07891v3 | Shir Ashury-Tahan, Ariel Gera, Benjamin Sznajder, Leshem Choshen, Liat Ein-Dor, Eyal Shnarch | 2024-06-06T11:07:17Z | cs.CL, cs.LG | Model selection for a given target task can be costly, as it may entail
extensive annotation of the quality of outputs of different models. We
introduce DiffUse, an efficient method to make an informed decision between
candidate text generation models based on preference annotations. DiffUse
reduces the required amount of annotations, thus saving valuable time and
resources in performing evaluation. DiffUse intelligently selects instances by
clustering embeddings that represent the semantic differences between model
outputs. Thus, it is able to identify a subset of examples that are more
informative for preference decisions. Our method is model-agnostic, and can be
applied to any text generation model for selecting between models, prompts and
configurations. Moreover, we propose a practical iterative approach for
dynamically determining how many instances to annotate. In a series of
experiments over hundreds of model pairs, we demonstrate that DiffUse can
dramatically reduce the required number of annotations -- by up to 75% -- while
maintaining high evaluation reliability. | cs.CL | Label-Efficient Model Selection for Text Generation |
2024-02-12T18:41:34Z | http://arxiv.org/pdf/2402.07876v4 | http://arxiv.org/abs/2402.07876v4 | Victor Zhong, Dipendra Misra, Xingdi Yuan, Marc-Alexandre Côté | 2024-04-18T20:35:32Z | cs.LG, cs.AI, cs.CL | We introduce Language Feedback Models (LFMs) that identify desirable
behaviour - actions that help achieve tasks specified in the instruction - for
imitation learning in instruction following. To train LFMs, we obtain feedback
from Large Language Models (LLMs) on visual trajectories verbalized to language
descriptions. First, by using LFMs to identify desirable behaviour to imitate,
we improve in task-completion rate over strong behavioural cloning baselines on
three distinct language grounding environments (Touchdown, ScienceWorld, and
ALFWorld). Second, LFMs outperform using LLMs as experts to directly predict
actions, when controlling for the number of LLM output tokens. Third, LFMs
generalize to unseen environments, improving task-completion rate by 3.5-12.0%
through one round of adaptation. Finally, LFM can be modified to provide
human-interpretable feedback without performance loss, allowing human
verification of desirable behaviour for imitation learning. | cs.LG | Policy Improvement using Language Feedback Models |
2024-02-12T18:33:47Z | http://arxiv.org/pdf/2402.07871v1 | http://arxiv.org/abs/2402.07871v1 | Jakub Krajewski, Jan Ludziejewski, Kamil Adamczewski, Maciej Pióro, Michał Krutul, Szymon Antoniak, Kamil Ciebiera, Krystian Król, Tomasz Odrzygóźdź, Piotr Sankowski, Marek Cygan, Sebastian Jaszczur | 2024-02-12T18:33:47Z | cs.LG, cs.AI, cs.CL | Mixture of Experts (MoE) models have emerged as a primary solution for
reducing the computational cost of Large Language Models. In this work, we
analyze their scaling properties, incorporating an expanded range of variables.
Specifically, we introduce a new hyperparameter, granularity, whose adjustment
enables precise control over the size of the experts. Building on this, we
establish scaling laws for fine-grained MoE, taking into account the number of
training tokens, model size, and granularity. Leveraging these laws, we derive
the optimal training configuration for a given computational budget. Our
findings not only show that MoE models consistently outperform dense
Transformers but also highlight that the efficiency gap between dense and MoE
models widens as we scale up the model size and training budget. Furthermore,
we demonstrate that the common practice of setting the size of experts in MoE
to mirror the feed-forward layer is not optimal at almost any computational
budget. | cs.LG | Scaling Laws for Fine-Grained Mixture of Experts |
2024-02-12T18:33:47Z | http://arxiv.org/pdf/2402.07872v1 | http://arxiv.org/abs/2402.07872v1 | Soroush Nasiriany, Fei Xia, Wenhao Yu, Ted Xiao, Jacky Liang, Ishita Dasgupta, Annie Xie, Danny Driess, Ayzaan Wahid, Zhuo Xu, Quan Vuong, Tingnan Zhang, Tsang-Wei Edward Lee, Kuang-Huei Lee, Peng Xu, Sean Kirmani, Yuke Zhu, Andy Zeng, Karol Hausman, Nicolas Heess, Chelsea Finn, Sergey Levine, Brian Ichter | 2024-02-12T18:33:47Z | cs.RO, cs.CL, cs.CV, cs.LG | Vision language models (VLMs) have shown impressive capabilities across a
variety of tasks, from logical reasoning to visual understanding. This opens
the door to richer interaction with the world, for example robotic control.
However, VLMs produce only textual outputs, while robotic control and other
spatial tasks require outputting continuous coordinates, actions, or
trajectories. How can we enable VLMs to handle such settings without
fine-tuning on task-specific data?
In this paper, we propose a novel visual prompting approach for VLMs that we
call Prompting with Iterative Visual Optimization (PIVOT), which casts tasks as
iterative visual question answering. In each iteration, the image is annotated
with a visual representation of proposals that the VLM can refer to (e.g.,
candidate robot actions, localizations, or trajectories). The VLM then selects
the best ones for the task. These proposals are iteratively refined, allowing
the VLM to eventually zero in on the best available answer. We investigate
PIVOT on real-world robotic navigation, real-world manipulation from images,
instruction following in simulation, and additional spatial inference tasks
such as localization. We find, perhaps surprisingly, that our approach enables
zero-shot control of robotic systems without any robot training data,
navigation in a variety of environments, and other capabilities. Although
current performance is far from perfect, our work highlights potentials and
limitations of this new regime and shows a promising approach for
Internet-Scale VLMs in robotic and spatial reasoning domains. Website:
pivot-prompt.github.io and HuggingFace:
https://huggingface.co/spaces/pivot-prompt/pivot-prompt-demo. | cs.RO | PIVOT: Iterative Visual Prompting Elicits Actionable Knowledge for VLMs |
2024-02-12T18:21:14Z | http://arxiv.org/pdf/2402.07865v2 | http://arxiv.org/abs/2402.07865v2 | Siddharth Karamcheti, Suraj Nair, Ashwin Balakrishna, Percy Liang, Thomas Kollar, Dorsa Sadigh | 2024-05-30T13:08:48Z | cs.CV, cs.AI, cs.CL, cs.LG | Visually-conditioned language models (VLMs) have seen growing adoption in
applications such as visual dialogue, scene understanding, and robotic task
planning; adoption that has fueled a wealth of new models such as LLaVa,
InstructBLIP, and PaLI-3. Despite the volume of new releases, key design
decisions around image preprocessing, architecture, and optimization are
under-explored, making it challenging to understand what factors account for
model performance $-$ a challenge further complicated by the lack of objective,
consistent evaluations. To address these gaps, we first compile a suite of
standardized evaluations spanning visual question answering, object
localization, and challenge sets that probe properties such as hallucination;
evaluations that provide fine-grained insight VLM capabilities. Second, we
rigorously investigate VLMs along key design axes, including pretrained visual
representations and training from base vs. instruct-tuned language models,
amongst others. We couple our analysis with three resource contributions: (1) a
unified framework for evaluating VLMs, (2) optimized, flexible training code,
and (3) checkpoints for all models, including a family of VLMs at the 7-13B
scale that strictly outperform InstructBLIP and LLaVa v1.5, the
state-of-the-art in open VLMs. | cs.CV | Prismatic VLMs: Investigating the Design Space of Visually-Conditioned Language Models |
2024-02-12T18:14:43Z | http://arxiv.org/pdf/2402.07862v1 | http://arxiv.org/abs/2402.07862v1 | Philipp Schoenegger, Peter S. Park, Ezra Karger, Philip E. Tetlock | 2024-02-12T18:14:43Z | cs.CY, cs.AI, cs.CL, cs.LG | Large language models (LLMs) show impressive capabilities, matching and
sometimes exceeding human performance in many domains. This study explores the
potential of LLMs to augment judgement in forecasting tasks. We evaluated the
impact on forecasting accuracy of two GPT-4-Turbo assistants: one designed to
provide high-quality advice ('superforecasting'), and the other designed to be
overconfident and base-rate-neglecting. Participants (N = 991) had the option
to consult their assigned LLM assistant throughout the study, in contrast to a
control group that used a less advanced model (DaVinci-003) without direct
forecasting support. Our preregistered analyses reveal that LLM augmentation
significantly enhances forecasting accuracy by 23% across both types of
assistants, compared to the control group. This improvement occurs despite the
superforecasting assistant's higher accuracy in predictions, indicating the
augmentation's benefit is not solely due to model prediction accuracy.
Exploratory analyses showed a pronounced effect in one forecasting item,
without which we find that the superforecasting assistant increased accuracy by
43%, compared with 28% for the biased assistant. We further examine whether LLM
augmentation disproportionately benefits less skilled forecasters, degrades the
wisdom-of-the-crowd by reducing prediction diversity, or varies in
effectiveness with question difficulty. Our findings do not consistently
support these hypotheses. Our results suggest that access to an LLM assistant,
even a biased one, can be a helpful decision aid in cognitively demanding tasks
where the answer is not known at the time of interaction. | cs.CY | AI-Augmented Predictions: LLM Assistants Improve Human Forecasting Accuracy |
2024-02-12T18:10:17Z | http://arxiv.org/pdf/2402.07859v2 | http://arxiv.org/abs/2402.07859v2 | Mirelle Bueno, Roberto Lotufo, Rodrigo Nogueira | 2024-02-20T15:12:13Z | cs.CL, cs.AI | Language models are now capable of solving tasks that require dealing with
long sequences consisting of hundreds of thousands of tokens. However, they
often fail on tasks that require repetitive use of simple rules, even on
sequences that are much shorter than those seen during training. For example,
state-of-the-art LLMs can find common items in two lists with up to 20 items
but fail when lists have 80 items. In this paper, we introduce Lissard, a
benchmark comprising seven tasks whose goal is to assess the ability of models
to process and generate wide-range sequence lengths, requiring repetitive
procedural execution. Our evaluation of open-source (Mistral-7B and
Mixtral-8x7B) and proprietary models (GPT-3.5 and GPT-4) show a consistent
decline in performance across all models as the complexity of the sequence
increases. The datasets and code are available at
https://github.com/unicamp-dl/Lissard | cs.CL | Lissard: Long and Simple Sequential Reasoning Datasets |
2024-02-12T17:53:22Z | http://arxiv.org/pdf/2402.07844v4 | http://arxiv.org/abs/2402.07844v4 | Mingzhe Du, Anh Tuan Luu, Bin Ji, Qian Liu, See-Kiong Ng | 2024-06-11T17:44:56Z | cs.SE, cs.CL | Amidst the recent strides in evaluating Large Language Models for Code (Code
LLMs), existing benchmarks have mainly focused on the functional correctness of
generated code, neglecting the importance of their computational efficiency. To
fill the gap, we present Mercury, the first code efficiency benchmark for Code
LLMs. It comprises 1,889 Python tasks, each accompanied by adequate solutions
that serve as real-world efficiency baselines, enabling a comprehensive
analysis of the runtime distribution. Based on the distribution, we introduce a
new metric Beyond, which computes a runtime-percentile-weighted Pass score to
reflect functional correctness and code efficiency simultaneously. On Mercury,
leading Code LLMs can achieve 65% on Pass, while less than 50% on Beyond. Given
that an ideal Beyond score would be aligned with the Pass score, it indicates
that while Code LLMs exhibit impressive capabilities in generating functionally
correct code, there remains a notable gap in their efficiency. Finally, our
empirical experiments reveal that Direct Preference Optimization (DPO) serves
as a robust baseline for enhancing code efficiency compared with Supervised
Fine Tuning (SFT), which paves a promising avenue for future exploration of
efficient code generation. Our code and data are available on GitHub:
https://github.com/Elfsong/Mercury. | cs.SE | Mercury: A Code Efficiency Benchmark for Code Large Language Models |
2024-02-12T17:52:05Z | http://arxiv.org/pdf/2402.07841v1 | http://arxiv.org/abs/2402.07841v1 | Michael Duan, Anshuman Suri, Niloofar Mireshghallah, Sewon Min, Weijia Shi, Luke Zettlemoyer, Yulia Tsvetkov, Yejin Choi, David Evans, Hannaneh Hajishirzi | 2024-02-12T17:52:05Z | cs.CL | Membership inference attacks (MIAs) attempt to predict whether a particular
datapoint is a member of a target model's training data. Despite extensive
research on traditional machine learning models, there has been limited work
studying MIA on the pre-training data of large language models (LLMs). We
perform a large-scale evaluation of MIAs over a suite of language models (LMs)
trained on the Pile, ranging from 160M to 12B parameters. We find that MIAs
barely outperform random guessing for most settings across varying LLM sizes
and domains. Our further analyses reveal that this poor performance can be
attributed to (1) the combination of a large dataset and few training
iterations, and (2) an inherently fuzzy boundary between members and
non-members. We identify specific settings where LLMs have been shown to be
vulnerable to membership inference and show that the apparent success in such
settings can be attributed to a distribution shift, such as when members and
non-members are drawn from the seemingly identical domain but with different
temporal ranges. We release our code and data as a unified benchmark package
that includes all existing MIAs, supporting future work. | cs.CL | Do Membership Inference Attacks Work on Large Language Models? |
2024-02-12T17:34:13Z | http://arxiv.org/pdf/2402.07827v1 | http://arxiv.org/abs/2402.07827v1 | Ahmet Üstün, Viraat Aryabumi, Zheng-Xin Yong, Wei-Yin Ko, Daniel D'souza, Gbemileke Onilude, Neel Bhandari, Shivalika Singh, Hui-Lee Ooi, Amr Kayid, Freddie Vargus, Phil Blunsom, Shayne Longpre, Niklas Muennighoff, Marzieh Fadaee, Julia Kreutzer, Sara Hooker | 2024-02-12T17:34:13Z | cs.CL | Recent breakthroughs in large language models (LLMs) have centered around a
handful of data-rich languages. What does it take to broaden access to
breakthroughs beyond first-class citizen languages? Our work introduces Aya, a
massively multilingual generative language model that follows instructions in
101 languages of which over 50% are considered as lower-resourced. Aya
outperforms mT0 and BLOOMZ on the majority of tasks while covering double the
number of languages. We introduce extensive new evaluation suites that broaden
the state-of-art for multilingual eval across 99 languages -- including
discriminative and generative tasks, human evaluation, and simulated win rates
that cover both held-out tasks and in-distribution performance. Furthermore, we
conduct detailed investigations on the optimal finetuning mixture composition,
data pruning, as well as the toxicity, bias, and safety of our models. We
open-source our instruction datasets and our model at
https://hf.co/CohereForAI/aya-101 | cs.CL | Aya Model: An Instruction Finetuned Open-Access Multilingual Language Model |
2024-02-12T17:24:15Z | http://arxiv.org/pdf/2402.07818v4 | http://arxiv.org/abs/2402.07818v4 | Z Liu, J Lou, W Bao, Y Hu, B Li, Z Qin, K Ren | 2024-05-09T09:41:23Z | cs.LG, cs.AI, cs.CL | Fine-tuning on task-specific datasets is a widely-embraced paradigm of
harnessing the powerful capability of pretrained LLMs for various downstream
tasks. Due to the popularity of LLMs fine-tuning and its accompanying privacy
concerns, differentially private (DP) fine-tuning of pretrained LLMs has been
widely used to safeguarding the privacy of task-specific datasets. Lying at the
design core of DP LLM fine-tuning methods is the satisfactory tradeoff among
privacy, utility, and scalability. Most existing methods build upon the seminal
work of DP-SGD. Despite pushing the scalability of DP-SGD to its limit,
DP-SGD-based fine-tuning methods are unfortunately limited by the inherent
inefficiency of SGD.
In this paper, we investigate the potential of DP zeroth-order methods for
LLM pretraining, which avoids the scalability bottleneck of SGD by
approximating the gradient with the more efficient zeroth-order gradient.
Rather than treating the zeroth-order method as a drop-in replacement for SGD,
this paper presents a comprehensive study both theoretically and empirically.
First, we propose the stagewise DP zeroth-order method (DP-ZOSO) that
dynamically schedules key hyperparameters. This design is grounded on the
synergy between DP random perturbation and the gradient approximation error of
the zeroth-order method, and its effect on fine-tuning trajectory.
We provide theoretical analysis for both proposed methods. We conduct
extensive empirical analysis on both encoder-only masked language model and
decoder-only autoregressive language model, achieving impressive results in
terms of scalability and utility (compared with DPZero, DP-ZOPO improves 4.5%
on SST-5, 5.5% on MNLI with RoBERTa-Large and 9.2% on CB, 3.9% on BoolQ with
OPT-2.7B when $\epsilon=4$). | cs.LG | Differentially Private Zeroth-Order Methods for Scalable Large Language Model Finetuning |
2024-02-12T17:22:42Z | http://arxiv.org/pdf/2402.07817v1 | http://arxiv.org/abs/2402.07817v1 | Anna Mosolova, Marie Candito, Carlos Ramisch | 2024-02-12T17:22:42Z | cs.CL | While static word embeddings are blind to context, for lexical semantics
tasks context is rather too present in contextual word embeddings, vectors of
same-meaning occurrences being too different (Ethayarajh, 2019). Fine-tuning
pre-trained language models (PLMs) using contrastive learning was proposed,
leveraging automatically self-augmented examples (Liu et al., 2021b). In this
paper, we investigate how to inject a lexicon as an alternative source of
supervision, using the English Wiktionary. We also test how dimensionality
reduction impacts the resulting contextual word embeddings. We evaluate our
approach on the Word-In-Context (WiC) task, in the unsupervised setting (not
using the training set). We achieve new SoTA result on the original WiC test
set. We also propose two new WiC test sets for which we show that our
fine-tuning method achieves substantial improvements. We also observe
improvements, although modest, for the semantic frame induction task. Although
we experimented on English to allow comparison with related work, our method is
adaptable to the many languages for which large Wiktionaries exist. | cs.CL | Injecting Wiktionary to improve token-level contextual representations using contrastive learning |
2024-02-12T17:17:50Z | http://arxiv.org/pdf/2402.07812v1 | http://arxiv.org/abs/2402.07812v1 | Thomas Pouplin, Hao Sun, Samuel Holt, Mihaela van der Schaar | 2024-02-12T17:17:50Z | cs.CL, cs.AI, cs.IR, cs.LG, H.3.3; I.2.6; I.2.7; I.2.8 | Large Language Models (LLMs) have demonstrated their strong ability to assist
people and show "sparks of intelligence". However, several open challenges
hinder their wider application: such as concerns over privacy, tendencies to
produce hallucinations, and difficulties in handling long contexts. In this
work, we address those challenges by introducing the Retrieval-Augmented
Thought Process (RATP). Given access to external knowledge, RATP formulates the
thought generation of LLMs as a multiple-step decision process. To optimize
such a thought process, RATP leverages Monte-Carlo Tree Search, and learns a
Q-value estimator that permits cost-efficient inference. In addressing the task
of question-answering with private data, where ethical and security concerns
limit LLM training methods, RATP achieves a 50% improvement over existing
in-context retrieval-augmented language models. | cs.CL | Retrieval-Augmented Thought Process as Sequential Decision Making |
2024-02-12T16:54:22Z | http://arxiv.org/pdf/2402.07788v1 | http://arxiv.org/abs/2402.07788v1 | Mingzhe Li, Xiuying Chen, Jing Xiang, Qishen Zhang, Changsheng Ma, Chenchen Dai, Jinxiong Chang, Zhongyi Liu, Guannan Zhang | 2024-02-12T16:54:22Z | cs.CL | Text matching systems have become a fundamental service in most searching
platforms. For instance, they are responsible for matching user queries to
relevant candidate items, or rewriting the user-input query to a pre-selected
high-performing one for a better search experience. In practice, both the
queries and items often contain multiple attributes, such as the category of
the item and the location mentioned in the query, which represent condensed key
information that is helpful for matching. However, most of the existing works
downplay the effectiveness of attributes by integrating them into text
representations as supplementary information. Hence, in this work, we focus on
exploring the relationship between the attributes from two sides. Since
attributes from two ends are often not aligned in terms of number and type, we
propose to exploit the benefit of attributes by multiple-intent modeling. The
intents extracted from attributes summarize the diverse needs of queries and
provide rich content of items, which are more refined and abstract, and can be
aligned for paired inputs. Concretely, we propose a multi-intent
attribute-aware matching model (MIM), which consists of three main components:
attribute-aware encoder, multi-intent modeling, and intent-aware matching. In
the attribute-aware encoder, the text and attributes are weighted and processed
through a scaled attention mechanism with regard to the attributes' importance.
Afterward, the multi-intent modeling extracts intents from two ends and aligns
them. Herein, we come up with a distribution loss to ensure the learned intents
are diverse but concentrated, and a kullback-leibler divergence loss that
aligns the learned intents. Finally, in the intent-aware matching, the intents
are evaluated by a self-supervised masking task, and then incorporated to
output the final matching result. | cs.CL | Multi-Intent Attribute-Aware Text Matching in Searching |
2024-02-12T16:52:26Z | http://arxiv.org/pdf/2402.07787v3 | http://arxiv.org/abs/2402.07787v3 | Xiaowei Zhao, Yong Zhou, Xiujuan Xu, Yu Liu | 2024-03-04T08:42:32Z | cs.AI, cs.CL | Aspect-based Sentiment Analysis (ABSA) evaluates sentiment expressions within
a text to comprehend sentiment information. Previous studies integrated
external knowledge, such as knowledge graphs, to enhance the semantic features
in ABSA models. Recent research has examined the use of Graph Neural Networks
(GNNs) on dependency and constituent trees for syntactic analysis. With the
ongoing development of ABSA, more innovative linguistic and structural features
are being incorporated (e.g. latent graph), but this also introduces complexity
and confusion. As of now, a scalable framework for integrating diverse
linguistic and structural features into ABSA does not exist. This paper
presents the Extensible Multi-Granularity Fusion (EMGF) network, which
integrates information from dependency and constituent syntactic, attention
semantic , and external knowledge graphs. EMGF, equipped with multi-anchor
triplet learning and orthogonal projection, efficiently harnesses the combined
potential of each granularity feature and their synergistic interactions,
resulting in a cumulative effect without additional computational expenses.
Experimental findings on SemEval 2014 and Twitter datasets confirm EMGF's
superiority over existing ABSA methods. | cs.AI | Extensible Multi-Granularity Fusion Network for Aspect-based Sentiment Analysis |
2024-02-12T16:41:54Z | http://arxiv.org/pdf/2402.07776v2 | http://arxiv.org/abs/2402.07776v2 | Hui Liu, Wenya Wang, Haoru Li, Haoliang Li | 2024-05-28T06:14:34Z | cs.CL | The proliferation of fake news has emerged as a severe societal problem,
raising significant interest from industry and academia. While existing
deep-learning based methods have made progress in detecting fake news
accurately, their reliability may be compromised caused by the non-transparent
reasoning processes, poor generalization abilities and inherent risks of
integration with large language models (LLMs). To address this challenge, we
propose {\methodname}, a novel framework for trustworthy fake news detection
that prioritizes explainability, generalizability and controllability of
models. This is achieved via a dual-system framework that integrates cognition
and decision systems, adhering to the principles above. The cognition system
harnesses human expertise to generate logical predicates, which guide LLMs in
generating human-readable logic atoms. Meanwhile, the decision system deduces
generalizable logic rules to aggregate these atoms, enabling the identification
of the truthfulness of the input news across diverse domains and enhancing
transparency in the decision-making process. Finally, we present comprehensive
evaluation results on four datasets, demonstrating the feasibility and
trustworthiness of our proposed framework. Our implementation is available at
\url{https://github.com/less-and-less-bugs/Trust_TELLER}. | cs.CL | TELLER: A Trustworthy Framework for Explainable, Generalizable and Controllable Fake News Detection |
2024-02-12T16:32:37Z | http://arxiv.org/pdf/2402.07770v1 | http://arxiv.org/abs/2402.07770v1 | David Selby, Kai Spriestersbach, Yuichiro Iwashita, Dennis Bappert, Archana Warrier, Sumantrak Mukherjee, Muhammad Nabeel Asim, Koichi Kise, Sebastian Vollmer | 2024-02-12T16:32:37Z | cs.IR, cs.CL, stat.AP | Large language models (LLMs) have been extensively studied for their
abilities to generate convincing natural language sequences, however their
utility for quantitative information retrieval is less well understood. In this
paper we explore the feasibility of LLMs as a mechanism for quantitative
knowledge retrieval to aid data analysis tasks such as elicitation of prior
distributions for Bayesian models and imputation of missing data. We present a
prompt engineering framework, treating an LLM as an interface to a latent space
of scientific literature, comparing responses in different contexts and domains
against more established approaches. Implications and challenges of using LLMs
as 'experts' are discussed. | cs.IR | Quantitative knowledge retrieval from large language models |
2024-02-12T16:30:41Z | http://arxiv.org/pdf/2402.07767v2 | http://arxiv.org/abs/2402.07767v2 | Sourabrata Mukherjee, Akanksha Bansal, Atul Kr. Ojha, John P. McCrae, Ondřej Dušek | 2024-06-09T18:48:06Z | cs.CL | This paper focuses on text detoxification, i.e., automatically converting
toxic text into non-toxic text. This task contributes to safer and more
respectful online communication and can be considered a Text Style Transfer
(TST) task, where the text style changes while its content is preserved. We
present three approaches: knowledge transfer from a similar task, multi-task
learning approach, combining sequence-to-sequence modeling with various
toxicity classification tasks, and delete and reconstruct approach. To support
our research, we utilize a dataset provided by Dementieva et al.(2021), which
contains multiple versions of detoxified texts corresponding to toxic texts. In
our experiments, we selected the best variants through expert human annotators,
creating a dataset where each toxic sentence is paired with a single,
appropriate detoxified version. Additionally, we introduced a small Hindi
parallel dataset, aligning with a part of the English dataset, suitable for
evaluation purposes. Our results demonstrate that our approach effectively
balances text detoxication while preserving the actual content and maintaining
fluency. | cs.CL | Text Detoxification as Style Transfer in English and Hindi |
2024-02-12T16:23:28Z | http://arxiv.org/pdf/2402.07754v1 | http://arxiv.org/abs/2402.07754v1 | Jiacheng Ye, Shansan Gong, Liheng Chen, Lin Zheng, Jiahui Gao, Han Shi, Chuan Wu, Zhenguo Li, Wei Bi, Lingpeng Kong | 2024-02-12T16:23:28Z | cs.CL, cs.AI, cs.LG | Diffusion models have gained attention in text processing, offering many
potential advantages over traditional autoregressive models. This work explores
the integration of diffusion models and Chain-of-Thought (CoT), a
well-established technique to improve the reasoning ability in autoregressive
language models. We propose Diffusion-of-Thought (DoT), allowing reasoning
steps to diffuse over time through the diffusion process. In contrast to
traditional autoregressive language models that make decisions in a
left-to-right, token-by-token manner, DoT offers more flexibility in the
trade-off between computation and reasoning performance. Our experimental
results demonstrate the effectiveness of DoT in multi-digit multiplication and
grade school math problems. Additionally, DoT showcases promising
self-correction abilities and benefits from existing reasoning-enhancing
techniques like self-consistency decoding. Our findings contribute to the
understanding and development of reasoning capabilities in diffusion language
models. | cs.CL | Diffusion of Thoughts: Chain-of-Thought Reasoning in Diffusion Language Models |
2024-02-12T16:14:22Z | http://arxiv.org/pdf/2402.07744v2 | http://arxiv.org/abs/2402.07744v2 | Zonghan Yang, An Liu, Zijun Liu, Kaiming Liu, Fangzhou Xiong, Yile Wang, Zeyuan Yang, Qingyuan Hu, Xinrui Chen, Zhenhe Zhang, Fuwen Luo, Zhicheng Guo, Peng Li, Yang Liu | 2024-02-14T18:43:54Z | cs.AI, cs.CL, cs.LG | The rapid progress of foundation models has led to the prosperity of
autonomous agents, which leverage the universal capabilities of foundation
models to conduct reasoning, decision-making, and environmental interaction.
However, the efficacy of agents remains limited when operating in intricate,
realistic environments. In this work, we introduce the principles of
$\mathbf{U}$nified $\mathbf{A}$lignment for $\mathbf{A}$gents
($\mathbf{UA}^2$), which advocate for the simultaneous alignment of agents with
human intentions, environmental dynamics, and self-constraints such as the
limitation of monetary budgets. From the perspective of $\mathbf{UA}^2$, we
review the current agent research and highlight the neglected factors in
existing agent benchmarks and method candidates. We also conduct
proof-of-concept studies by introducing realistic features to WebShop,
including user profiles to demonstrate intentions, personalized reranking for
complex environmental dynamics, and runtime cost statistics to reflect
self-constraints. We then follow the principles of $\mathbf{UA}^2$ to propose
an initial design of our agent, and benchmark its performance with several
candidate baselines in the retrofitted WebShop. The extensive experimental
results further prove the importance of the principles of $\mathbf{UA}^2$. Our
research sheds light on the next steps of autonomous agent research with
improved general problem-solving abilities. | cs.AI | Towards Unified Alignment Between Agents, Humans, and Environment |
2024-02-12T16:04:01Z | http://arxiv.org/pdf/2402.07742v1 | http://arxiv.org/abs/2402.07742v1 | Yifei Yuan, Clemencia Siro, Mohammad Aliannejadi, Maarten de Rijke, Wai Lam | 2024-02-12T16:04:01Z | cs.CL, cs.CV | In mixed-initiative conversational search systems, clarifying questions are
used to help users who struggle to express their intentions in a single query.
These questions aim to uncover user's information needs and resolve query
ambiguities. We hypothesize that in scenarios where multimodal information is
pertinent, the clarification process can be improved by using non-textual
information. Therefore, we propose to add images to clarifying questions and
formulate the novel task of asking multimodal clarifying questions in
open-domain, mixed-initiative conversational search systems. To facilitate
research into this task, we collect a dataset named Melon that contains over 4k
multimodal clarifying questions, enriched with over 14k images. We also propose
a multimodal query clarification model named Marto and adopt a prompt-based,
generative fine-tuning strategy to perform the training of different stages
with different prompts. Several analyses are conducted to understand the
importance of multimodal contents during the query clarification phase.
Experimental results indicate that the addition of images leads to significant
improvements of up to 90% in retrieval performance when selecting the relevant
images. Extensive analyses are also performed to show the superiority of Marto
compared with discriminative baselines in terms of effectiveness and
efficiency. | cs.CL | Asking Multimodal Clarifying Questions in Mixed-Initiative Conversational Search |
2024-02-12T15:41:22Z | http://arxiv.org/pdf/2402.07729v1 | http://arxiv.org/abs/2402.07729v1 | Qian Yang, Jin Xu, Wenrui Liu, Yunfei Chu, Ziyue Jiang, Xiaohuan Zhou, Yichong Leng, Yuanjun Lv, Zhou Zhao, Chang Zhou, Jingren Zhou | 2024-02-12T15:41:22Z | eess.AS, cs.CL, cs.LG, cs.SD | Recently, instruction-following audio-language models have received broad
attention for human-audio interaction. However, the absence of benchmarks
capable of evaluating audio-centric interaction capabilities has impeded
advancements in this field. Previous models primarily focus on assessing
different fundamental tasks, such as Automatic Speech Recognition (ASR), and
lack an assessment of the open-ended generative capabilities centered around
audio. Thus, it is challenging to track the progression in the Large
Audio-Language Models (LALMs) domain and to provide guidance for future
improvement. In this paper, we introduce AIR-Bench (\textbf{A}udio
\textbf{I}nst\textbf{R}uction \textbf{Bench}mark), the first benchmark designed
to evaluate the ability of LALMs to understand various types of audio signals
(including human speech, natural sounds, and music), and furthermore, to
interact with humans in the textual format. AIR-Bench encompasses two
dimensions: \textit{foundation} and \textit{chat} benchmarks. The former
consists of 19 tasks with approximately 19k single-choice questions, intending
to inspect the basic single-task ability of LALMs. The latter one contains 2k
instances of open-ended question-and-answer data, directly assessing the
comprehension of the model on complex audio and its capacity to follow
instructions. Both benchmarks require the model to generate hypotheses
directly. We design a unified framework that leverages advanced language
models, such as GPT-4, to evaluate the scores of generated hypotheses given the
meta-information of the audio. Experimental results demonstrate a high level of
consistency between GPT-4-based evaluation and human evaluation. By revealing
the limitations of existing LALMs through evaluation results, AIR-Bench can
provide insights into the direction of future research. | eess.AS | AIR-Bench: Benchmarking Large Audio-Language Models via Generative Comprehension |
2024-02-12T15:39:05Z | http://arxiv.org/pdf/2402.07726v1 | http://arxiv.org/abs/2402.07726v1 | Zhengsheng Guo, Zhiwei He, Wenxiang Jiao, Xing Wang, Rui Wang, Kehai Chen, Zhaopeng Tu, Yong Xu, Min Zhang | 2024-02-12T15:39:05Z | cs.CL | Motivated by the success of unsupervised neural machine translation (UNMT),
we introduce an unsupervised sign language translation and generation network
(USLNet), which learns from abundant single-modality (text and video) data
without parallel sign language data. USLNet comprises two main components:
single-modality reconstruction modules (text and video) that rebuild the input
from its noisy version in the same modality and cross-modality back-translation
modules (text-video-text and video-text-video) that reconstruct the input from
its noisy version in the different modality using back-translation
procedure.Unlike the single-modality back-translation procedure in text-based
UNMT, USLNet faces the cross-modality discrepancy in feature representation, in
which the length and the feature dimension mismatch between text and video
sequences. We propose a sliding window method to address the issues of aligning
variable-length text with video sequences. To our knowledge, USLNet is the
first unsupervised sign language translation and generation model capable of
generating both natural language text and sign language video in a unified
manner. Experimental results on the BBC-Oxford Sign Language dataset (BOBSL)
and Open-Domain American Sign Language dataset (OpenASL) reveal that USLNet
achieves competitive results compared to supervised baseline models, indicating
its effectiveness in sign language translation and generation. | cs.CL | Unsupervised Sign Language Translation and Generation |
2024-02-12T15:34:56Z | http://arxiv.org/pdf/2402.07721v2 | http://arxiv.org/abs/2402.07721v2 | Hongyun Zhou, Xiangyu Lu, Wang Xu, Conghui Zhu, Tiejun Zhao, Muyun Yang | 2024-06-18T15:13:12Z | cs.LG, cs.CL | Low-Rank Adaptation (LoRA) is currently the most commonly used
Parameter-efficient fine-tuning (PEFT) method, it introduces auxiliary
parameters for each layer to fine-tune the pre-trained model under limited
computing resources. However, it still faces resource consumption challenges
during training when scaling up to larger models. Most previous studies have
tackled this issue by using pruning techniques, which involve removing LoRA
parameters deemed unimportant. Nonetheless, these efforts only analyze LoRA
parameter features to evaluate their importance, such as parameter count, size,
and gradient. In fact, the output of LoRA (product of LoRA parameter and hidden
state), directly impacts the final results. Preliminary experiments indicate
that a fraction of LoRA elements possesses significantly high output values,
substantially influencing the layer output. Motivated by the observation, we
propose LoRA-drop. Concretely, LoRA-drop evaluates the importance of LoRA based
on the LoRA output. Then we retain LoRA for important layers and the other
layers share the same LoRA. We conduct abundant experiments with models of
different scales on NLU and NLG tasks. Results demonstrate that LoRA-drop can
achieve performance comparable to full fine-tuning and LoRA, while retaining
50\% of the LoRA parameters on average. | cs.LG | LoRA-drop: Efficient LoRA Parameter Pruning based on Output Evaluation |
2024-02-12T14:53:37Z | http://arxiv.org/pdf/2402.07689v2 | http://arxiv.org/abs/2402.07689v2 | Irina Alekseevskaia, Konstantin Arkhipenko | 2024-04-06T21:41:10Z | cs.CL, cs.AI | The use of third-party datasets and pre-trained machine learning models poses
a threat to NLP systems due to possibility of hidden backdoor attacks. Existing
attacks involve poisoning the data samples such as insertion of tokens or
sentence paraphrasing, which either alter the semantics of the original texts
or can be detected. Our main difference from the previous work is that we use
the reposition of a two words in a sentence as a trigger. By designing and
applying specific part-of-speech (POS) based rules for selecting these tokens,
we maintain high attack success rate on SST-2 and AG classification datasets
while outperforming existing attacks in terms of perplexity and semantic
similarity to the clean samples. In addition, we show the robustness of our
attack to the ONION defense method. All the code and data for the paper can be
obtained at https://github.com/alekseevskaia/OrderBkd. | cs.CL | OrderBkd: Textual backdoor attack through repositioning |
2024-02-12T14:42:33Z | http://arxiv.org/pdf/2402.07682v1 | http://arxiv.org/abs/2402.07682v1 | Marie Candito | 2024-02-12T14:42:33Z | cs.CL | The biaffine parser of Dozat and Manning (2017) was successfully extended to
semantic dependency parsing (SDP) (Dozat and Manning, 2018). Its performance on
graphs is surprisingly high given that, without the constraint of producing a
tree, all arcs for a given sentence are predicted independently from each other
(modulo a shared representation of tokens). To circumvent such an independence
of decision, while retaining the O(n^2) complexity and highly parallelizable
architecture, we propose to use simple auxiliary tasks that introduce some form
of interdependence between arcs. Experiments on the three English acyclic
datasets of SemEval 2015 task 18 (Oepen et al., 2015), and on French deep
syntactic cyclic graphs (Ribeyre et al., 2014) show modest but systematic
performance gains on a near state-of-the-art baseline using transformer-based
contextualized representations. This provides a simple and robust method to
boost SDP performance. | cs.CL | Auxiliary Tasks to Boost Biaffine Semantic Dependency Parsing |
2024-02-12T14:40:54Z | http://arxiv.org/pdf/2402.07681v1 | http://arxiv.org/abs/2402.07681v1 | Vicent Briva-Iglesias, Joao Lucas Cavalheiro Camargo, Gokhan Dogru | 2024-02-12T14:40:54Z | cs.CL, cs.AI | This study evaluates the machine translation (MT) quality of two
state-of-the-art large language models (LLMs) against a tradition-al neural
machine translation (NMT) system across four language pairs in the legal
domain. It combines automatic evaluation met-rics (AEMs) and human evaluation
(HE) by professional transla-tors to assess translation ranking, fluency and
adequacy. The re-sults indicate that while Google Translate generally
outperforms LLMs in AEMs, human evaluators rate LLMs, especially GPT-4,
comparably or slightly better in terms of producing contextually adequate and
fluent translations. This discrepancy suggests LLMs' potential in handling
specialized legal terminology and context, highlighting the importance of human
evaluation methods in assessing MT quality. The study underscores the evolving
capabil-ities of LLMs in specialized domains and calls for reevaluation of
traditional AEMs to better capture the nuances of LLM-generated translations. | cs.CL | Large Language Models "Ad Referendum": How Good Are They at Machine Translation in the Legal Domain? |
2024-02-12T14:01:12Z | http://arxiv.org/pdf/2402.07658v1 | http://arxiv.org/abs/2402.07658v1 | Ayo Adedeji, Sarita Joshi, Brendan Doohan | 2024-02-12T14:01:12Z | cs.CL, cs.SD, eess.AS | In the rapidly evolving landscape of medical documentation, transcribing
clinical dialogues accurately is increasingly paramount. This study explores
the potential of Large Language Models (LLMs) to enhance the accuracy of
Automatic Speech Recognition (ASR) systems in medical transcription. Utilizing
the PriMock57 dataset, which encompasses a diverse range of primary care
consultations, we apply advanced LLMs to refine ASR-generated transcripts. Our
research is multifaceted, focusing on improvements in general Word Error Rate
(WER), Medical Concept WER (MC-WER) for the accurate transcription of essential
medical terms, and speaker diarization accuracy. Additionally, we assess the
role of LLM post-processing in improving semantic textual similarity, thereby
preserving the contextual integrity of clinical dialogues. Through a series of
experiments, we compare the efficacy of zero-shot and Chain-of-Thought (CoT)
prompting techniques in enhancing diarization and correction accuracy. Our
findings demonstrate that LLMs, particularly through CoT prompting, not only
improve the diarization accuracy of existing ASR systems but also achieve
state-of-the-art performance in this domain. This improvement extends to more
accurately capturing medical concepts and enhancing the overall semantic
coherence of the transcribed dialogues. These findings illustrate the dual role
of LLMs in augmenting ASR outputs and independently excelling in transcription
tasks, holding significant promise for transforming medical ASR systems and
leading to more accurate and reliable patient records in healthcare settings. | cs.CL | The Sound of Healthcare: Improving Medical Transcription ASR Accuracy with Large Language Models |
2024-02-12T13:34:33Z | http://arxiv.org/pdf/2402.07645v1 | http://arxiv.org/abs/2402.07645v1 | Isabelle Lorge, Dan W. Joyce, Niall Taylor, Alejo Nevado-Holgado, Andrea Cipriani, Andrey Kormilitzin | 2024-02-12T13:34:33Z | cs.CL | Difficult-to-treat depression (DTD) has been proposed as a broader and more
clinically comprehensive perspective on a person's depressive disorder where
despite treatment, they continue to experience significant burden. We sought to
develop a Large Language Model (LLM)-based tool capable of interrogating
routinely-collected, narrative (free-text) electronic health record (EHR) data
to locate published prognostic factors that capture the clinical syndrome of
DTD. In this work, we use LLM-generated synthetic data (GPT3.5) and a
Non-Maximum Suppression (NMS) algorithm to train a BERT-based span extraction
model. The resulting model is then able to extract and label spans related to a
variety of relevant positive and negative factors in real clinical data (i.e.
spans of text that increase or decrease the likelihood of a patient matching
the DTD syndrome). We show it is possible to obtain good overall performance
(0.70 F1 across polarity) on real clinical data on a set of as many as 20
different factors, and high performance (0.85 F1 with 0.95 precision) on a
subset of important DTD factors such as history of abuse, family history of
affective disorder, illness severity and suicidality by training the model
exclusively on synthetic data. Our results show promise for future healthcare
applications especially in applications where traditionally, highly
confidential medical data and human-expert annotation would normally be
required. | cs.CL | Detecting the Clinical Features of Difficult-to-Treat Depression using Synthetic Data from Large Language Models |
2024-02-12T13:09:21Z | http://arxiv.org/pdf/2402.07625v2 | http://arxiv.org/abs/2402.07625v2 | Yifan Zhang, Yifan Luo, Yang Yuan, Andrew Chi-Chih Yao | 2024-04-02T04:17:30Z | cs.CL, cs.AI, cs.LG | To improve language models' proficiency in mathematical reasoning via
continual pretraining, we introduce a novel strategy that leverages base
language models for autonomous data selection. Departing from conventional
supervised fine-tuning or trained classifiers with human-annotated data, our
approach Autonomous Data Selection (AutoDS) utilizes meta-prompted language
models as zero-shot verifiers to evaluate and select high-quality mathematical
content autonomously. To demonstrate the efficacy of our method, we
continuously pretrained a 7B-parameter language model on our curated dataset,
achieving substantial improvements in downstream performance on the MATH,
GSM8K, and BIG-Bench Hard (BBH) tasks with a token amount reduced by orders of
magnitude compared to previous continual pretraining works. Our method
showcases a 2 times increase in pretraining token efficiency compared to
state-of-the-art baselines, underscoring the potential of our approach in
enhancing models' mathematical reasoning capabilities. The AutoMathText dataset
is available at https://huggingface.co/datasets/math-ai/AutoMathText. The code
is available at https://github.com/yifanzhang-pro/AutoMathText. | cs.CL | Autonomous Data Selection with Language Models for Mathematical Texts |
2024-02-12T12:48:02Z | http://arxiv.org/pdf/2402.07616v3 | http://arxiv.org/abs/2402.07616v3 | Jianhui Pang, Fanghua Ye, Derek Fai Wong, Xin He, Wanshun Chen, Longyue Wang | 2024-06-01T04:52:17Z | cs.CL, cs.AI | Large language models (LLMs) predominantly employ decoder-only transformer
architectures, necessitating the retention of keys/values information for
historical tokens to provide contextual information and avoid redundant
computation. However, the substantial size and parameter volume of these LLMs
require massive GPU memory. This memory demand increases with the length of the
input text, leading to an urgent need for more efficient methods of information
storage and processing. This study introduces Anchor-based LLMs (AnLLMs), which
utilize an innovative anchor-based self-attention network (AnSAN) and also an
anchor-based inference strategy. This approach enables LLMs to compress
sequence information into an anchor token, reducing the keys/values cache and
enhancing inference efficiency. Experiments on question-answering benchmarks
reveal that AnLLMs maintain similar accuracy levels while achieving up to 99%
keys/values cache reduction and up to 3.5 times faster inference. Despite a
minor compromise in accuracy, the substantial enhancements of AnLLMs employing
the AnSAN technique in resource utilization and computational efficiency
underscore their potential for practical LLM applications. | cs.CL | Anchor-based Large Language Models |
2024-02-12T12:30:42Z | http://arxiv.org/pdf/2402.07610v3 | http://arxiv.org/abs/2402.07610v3 | Haoyu Wang, Guozheng Ma, Ziqiao Meng, Zeyu Qin, Li Shen, Zhong Zhang, Bingzhe Wu, Liu Liu, Yatao Bian, Tingyang Xu, Xueqian Wang, Peilin Zhao | 2024-06-27T16:38:35Z | cs.CL, cs.AI | Self-alignment is an effective way to reduce the cost of human annotation
while ensuring promising model capability. However, most current methods
complete the data collection and training steps in a single round, which may
overlook the continuously improving ability of self-aligned models. This gives
rise to a key query: What if we do multi-time bootstrapping self-alignment?
Does this strategy enhance model performance or lead to rapid degradation? In
this paper, our pioneering exploration delves into the impact of bootstrapping
self-alignment on large language models. Our findings reveal that bootstrapping
self-alignment markedly surpasses the single-round approach, by guaranteeing
data diversity from in-context learning. To further exploit the capabilities of
bootstrapping, we investigate and adjust the training order of data, which
yields improved performance of the model. Drawing on these findings, we propose
Step-On-Feet Tuning (SOFT) which leverages model's continuously enhanced
few-shot ability to boost zero or one-shot performance. Based on easy-to-hard
training recipe, we propose SOFT+ which further boost self-alignment's
performance. Our experiments demonstrate the efficiency of SOFT (SOFT+) across
various classification and generation tasks, highlighting the potential of
bootstrapping self-alignment on continually enhancing model alignment
performance. | cs.CL | Step-On-Feet Tuning: Scaling Self-Alignment of LLMs via Bootstrapping |
2024-02-12T11:18:32Z | http://arxiv.org/pdf/2402.07577v2 | http://arxiv.org/abs/2402.07577v2 | Thong Nguyen, Xiaobao Wu, Xinshuai Dong, Cong-Duy T Nguyen, See-Kiong Ng, Anh Tuan Luu | 2024-03-09T05:35:21Z | cs.CL | Recent representation learning approaches enhance neural topic models by
optimizing the weighted linear combination of the evidence lower bound (ELBO)
of the log-likelihood and the contrastive learning objective that contrasts
pairs of input documents. However, document-level contrastive learning might
capture low-level mutual information, such as word ratio, which disturbs topic
modeling. Moreover, there is a potential conflict between the ELBO loss that
memorizes input details for better reconstruction quality, and the contrastive
loss which attempts to learn topic representations that generalize among input
documents. To address these issues, we first introduce a novel contrastive
learning method oriented towards sets of topic vectors to capture useful
semantics that are shared among a set of input documents. Secondly, we
explicitly cast contrastive topic modeling as a gradient-based multi-objective
optimization problem, with the goal of achieving a Pareto stationary solution
that balances the trade-off between the ELBO and the contrastive objective.
Extensive experiments demonstrate that our framework consistently produces
higher-performing neural topic models in terms of topic coherence, topic
diversity, and downstream performance. | cs.CL | Topic Modeling as Multi-Objective Contrastive Optimization |
2024-02-12T10:11:50Z | http://arxiv.org/pdf/2402.07543v1 | http://arxiv.org/abs/2402.07543v1 | Mohamad Ballout, Ulf Krumnack, Gunther Heidemann, Kai-Uwe Kuehnberger | 2024-02-12T10:11:50Z | cs.CL, cs.AI, cs.LG | Our research demonstrates the significant benefits of using fine-tuning with
explanations to enhance the performance of language models. Unlike prompting,
which maintains the model's parameters, fine-tuning allows the model to learn
and update its parameters during a training phase. In this study, we applied
fine-tuning to various sized language models using data that contained
explanations of the output rather than merely presenting the answers. We found
that even smaller language models with as few as 60 million parameters
benefited substantially from this approach. Interestingly, our results
indicated that the detailed explanations were more beneficial to smaller models
than larger ones, with the latter gaining nearly the same advantage from any
form of explanation, irrespective of its length. Additionally, we demonstrate
that the inclusion of explanations enables the models to solve tasks that they
were not able to solve without explanations. Lastly, we argue that despite the
challenging nature of adding explanations, samples that contain explanations
not only reduce the volume of data required for training but also promote a
more effective generalization by the model. In essence, our findings suggest
that fine-tuning with explanations significantly bolsters the performance of
large language models. | cs.CL | Show Me How It's Done: The Role of Explanations in Fine-Tuning Language Models |
2024-02-12T10:09:16Z | http://arxiv.org/pdf/2402.07540v1 | http://arxiv.org/abs/2402.07540v1 | Nolwenn Bernard, Ivica Kostric, Weronika Łajewska, Krisztian Balog, Petra Galuščáková, Vinay Setty, Martin G. Skjæveland | 2024-02-12T10:09:16Z | cs.HC, cs.AI, cs.CL | Personal knowledge graphs (PKGs) offer individuals a way to store and
consolidate their fragmented personal data in a central place, improving
service personalization while maintaining full user control. Despite their
potential, practical PKG implementations with user-friendly interfaces remain
scarce. This work addresses this gap by proposing a complete solution to
represent, manage, and interface with PKGs. Our approach includes (1) a
user-facing PKG Client, enabling end-users to administer their personal data
easily via natural language statements, and (2) a service-oriented PKG API. To
tackle the complexity of representing these statements within a PKG, we present
an RDF-based PKG vocabulary that supports this, along with properties for
access rights and provenance. | cs.HC | PKG API: A Tool for Personal Knowledge Graph Management |
2024-02-12T10:04:07Z | http://arxiv.org/pdf/2402.07536v1 | http://arxiv.org/abs/2402.07536v1 | Kang Zhang, Osamu Yoshie, Weiran Huang | 2024-02-12T10:04:07Z | cs.AI, cs.CL | Trading range breakout (TRB) is a key method in the technical analysis of
financial trading, widely employed by traders in financial markets such as
stocks, futures, and foreign exchange. However, distinguishing between true and
false breakout and providing the correct rationale cause significant challenges
to investors. Recently, large language models have achieved success in various
downstream applications, but their effectiveness in the domain of financial
breakout detection has been subpar. The reason is that the unique data and
specific knowledge are required in breakout detection. To address these issues,
we introduce BreakGPT, the first large language model for financial breakout
detection. Furthermore, we have developed a novel framework for large language
models, namely multi-stage structure, effectively reducing mistakes in
downstream applications. Experimental results indicate that compared to
GPT-3.5, BreakGPT improves the accuracy of answers and rational by 44%, with
the multi-stage structure contributing 17.6% to the improvement. Additionally,
it outperforms ChatGPT-4 by 42.07%. Our Code is publicly available:
https://github.com/Neviim96/BreakGPT | cs.AI | BreakGPT: A Large Language Model with Multi-stage Structure for Financial Breakout Detection |
2024-02-12T09:41:00Z | http://arxiv.org/pdf/2402.07519v1 | http://arxiv.org/abs/2402.07519v1 | Prachi Jain, Ashutosh Sathe, Varun Gumma, Kabir Ahuja, Sunayana Sitaram | 2024-02-12T09:41:00Z | cs.CL, cs.CY | Pretrained Language Models (PLMs) are widely used in NLP for various tasks.
Recent studies have identified various biases that such models exhibit and have
proposed methods to correct these biases. However, most of the works address a
limited set of bias dimensions independently such as gender, race, or religion.
Moreover, the methods typically involve finetuning the full model to maintain
the performance on the downstream task. In this work, we aim to modularly
debias a pretrained language model across multiple dimensions. Previous works
extensively explored debiasing PLMs using limited US-centric counterfactual
data augmentation (CDA). We use structured knowledge and a large generative
model to build a diverse CDA across multiple bias dimensions in a
semi-automated way. We highlight how existing debiasing methods do not consider
interactions between multiple societal biases and propose a debiasing model
that exploits the synergy amongst various societal biases and enables
multi-bias debiasing simultaneously. An extensive evaluation on multiple tasks
and languages demonstrates the efficacy of our approach. | cs.CL | MAFIA: Multi-Adapter Fused Inclusive LanguAge Models |
2024-02-12T09:35:13Z | http://arxiv.org/pdf/2402.07513v1 | http://arxiv.org/abs/2402.07513v1 | Ajinkya Kulkarni, Anna Tokareva, Rameez Qureshi, Miguel Couceiro | 2024-02-12T09:35:13Z | cs.CL, cs.AI, cs.CY | In the field of spoken language understanding, systems like Whisper and
Multilingual Massive Speech (MMS) have shown state-of-the-art performances.
This study is dedicated to a comprehensive exploration of the Whisper and MMS
systems, with a focus on assessing biases in automatic speech recognition (ASR)
inherent to casual conversation speech specific to the Portuguese language. Our
investigation encompasses various categories, including gender, age, skin tone
color, and geo-location. Alongside traditional ASR evaluation metrics such as
Word Error Rate (WER), we have incorporated p-value statistical significance
for gender bias analysis. Furthermore, we extensively examine the impact of
data distribution and empirically show that oversampling techniques alleviate
such stereotypical biases. This research represents a pioneering effort in
quantifying biases in the Portuguese language context through the application
of MMS and Whisper, contributing to a better understanding of ASR systems'
performance in multilingual settings. | cs.CL | The Balancing Act: Unmasking and Alleviating ASR Biases in Portuguese |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.