Publications

Generative Escher Meshes
Noam Aigerman
Thibault Groueix
2024-07-12
Special Interest Group on Computer Graphics and Interactive Techniques Conference Conference Papers '24 (published)
doi.org
Investigating Low-Rank Training in Transformer Language Models: Efficiency and Scaling Analysis
Xiuying Wei
Skander Moalla
Razvan Pascanu
Caglar Gulçehre
State-of-the-art LLMs often rely on scale with high computational costs, which has sparked a research agenda to reduce parameter counts and … (see more)costs without significantly impacting performance. Our study focuses on Transformer-based LLMs, specifically applying low-rank parametrization to the computationally intensive feedforward networks (FFNs), which are less studied than attention blocks. In contrast to previous works, (i) we explore low-rank parametrization at scale, up to 1.3B parameters; (ii) within Transformer language models rather than convolutional architectures; and (iii) starting from training from scratch. Experiments on the large RefinedWeb dataset show that low-rank parametrization is both efficient (e.g., 2.6
2024-07-12
ArXiv (preprint)
doi.org
arxiv.org
Mining Action Rules for Defect Reduction Planning
Khouloud Oueslati
Gabriel Laberge
Maxime Lamothe
Foutse Khomh
Defect reduction planning plays a vital role in enhancing software quality and minimizing software maintenance costs. By training a black bo… (see more)x machine learning model and "explaining" its predictions, explainable AI for software engineering aims to identify the code characteristics that impact maintenance risks. However, post-hoc explanations do not always faithfully reflect what the original model computes. In this paper, we introduce CounterACT, a Counterfactual ACTion rule mining approach that can generate defect reduction plans without black-box models. By leveraging action rules, CounterACT provides a course of action that can be considered as a counterfactual explanation for the class (e.g., buggy or not buggy) assigned to a piece of code. We compare the effectiveness of CounterACT with the original action rule mining algorithm and six established defect reduction approaches on 9 software projects. Our evaluation is based on (a) overlap scores between proposed code changes and actual developer modifications; (b) improvement scores in future releases; and (c) the precision, recall, and F1-score of the plans. Our results show that, compared to competing approaches, CounterACT's explainable plans achieve higher overlap scores at the release level (median 95%) and commit level (median 85.97%), and they offer better trade-off between precision and recall (median F1-score 88.12%). Finally, we venture beyond planning and explore leveraging Large Language models (LLM) for generating code edits from our generated plans. Our results show that suggested LLM code edits supported by our plans are actionable and are more likely to pass relevant test cases than vanilla LLM code recommendations.
2024-07-11
Proceedings of the ACM on Software Engineering (published)
doi.org
arxiv.org
A Comprehensive Analysis of Explainable AI for Malware Hunting
Mohd Saqib
Samaneh Mahdavifar
Benjamin C. M. Fung
Philippe Charland
In the past decade, the number of malware variants has increased rapidly. Many researchers have proposed to detect malware using intelligent… (see more) techniques, such as Machine Learning (ML) and Deep Learning (DL), which have high accuracy and precision. These methods, however, suffer from being opaque in the decision-making process. Therefore, we need Artificial Intelligence (AI)-based models to be explainable, interpretable, and transparent to be reliable and trustworthy. In this survey, we reviewed articles related to Explainable AI (XAI) and their application to the significant scope of malware detection. The article encompasses a comprehensive examination of various XAI algorithms employed in malware analysis. Moreover, we have addressed the characteristics, challenges, and requirements in malware analysis that cannot be accommodated by standard XAI methods. We discussed that even though Explainable Malware Detection (EMD) models provide explainability, they make an AI-based model more vulnerable to adversarial attacks. We also propose a framework that assigns a level of explainability to each XAI malware analysis model, based on the security features involved in each method. In summary, the proposed project focuses on combining XAI and malware analysis to apply XAI models for scrutinizing the opaque nature of AI systems and their applications to malware analysis.
2024-07-10
ACM Computing Surveys (published)
doi.org
DeepCodeProbe: Towards Understanding What Models Trained on Code Learn
Vahid Majdinasab
Amin Nikanjam
Foutse Khomh
2024-07-10
ArXiv (preprint)
doi.org
arxiv.org
Group Membership Bias
Ali Vardasbi
Maarten de Rijke
Fernando Diaz
Mostafa Dehghani
2024-07-10
Proceedings of the 47th International ACM SIGIR Conference on Research and Development in Information Retrieval (published)
doi.org
arxiv.org
Beyond A*: Better Planning with Transformers via Search Dynamics Bootstrapping
Lucas Lehnert
Sainbayar Sukhbaatar
DiJia Su
Paul McVay
Qinqing Zheng
Michael G. Rabbat
Yuandong Tian
While Transformers have enabled tremendous progress in various application settings, such architectures still lag behind traditional symboli… (see more)c planners for solving complex decision making tasks. In this work, we demonstrate how to train Transformers to solve complex planning tasks. This is accomplished by training an encoder-decoder Transformer model to predict the _search dynamics_ of the
2024-07-09
colmweb.org/COLM/2024/Conference (accepted)
doi.org
openreview.net
Chain of Targeted Verification Questions to Improve the Reliability of Code Generated by LLMs
Sylvain Kouemo Ngassom
Arghavan Moradi Dakhel
Florian Tambon
Foutse Khomh
LLM-based assistants, such as GitHub Copilot and ChatGPT, have the potential to generate code that fulfills a programming task described in … (see more)a natural language description, referred to as a prompt. The widespread accessibility of these assistants enables users with diverse backgrounds to generate code and integrate it into software projects. However, studies show that code generated by LLMs is prone to bugs and may miss various corner cases in task specifications. Presenting such buggy code to users can impact their reliability and trust in LLM-based assistants. Moreover, significant efforts are required by the user to detect and repair any bug present in the code, especially if no test cases are available. In this study, we propose a self-refinement method aimed at improving the reliability of code generated by LLMs by minimizing the number of bugs before execution, without human intervention, and in the absence of test cases. Our approach is based on targeted Verification Questions (VQs) to identify potential bugs within the initial code. These VQs target various nodes within the Abstract Syntax Tree (AST) of the initial code, which have the potential to trigger specific types of bug patterns commonly found in LLM-generated code. Finally, our method attempts to repair these potential bugs by re-prompting the LLM with the targeted VQs and the initial code. Our evaluation, based on programming tasks in the CoderEval dataset, demonstrates that our proposed method outperforms state-of-the-art methods by decreasing the number of targeted errors in the code between 21% to 62% and improving the number of executable code instances to 13%.
2024-07-09
Proceedings of the 1st ACM International Conference on AI-Powered Software (published)
doi.org
arxiv.org
Decompose and Compare Consistency: Measuring VLMs' Answer Reliability via Task-Decomposition Consistency Comparison
Qian Yang
Weixiang Yan
Aishwarya Agrawal
Despite tremendous advancements, current state-of-the-art Vision-Language Models (VLMs) are still far from perfect. They tend to hallucinate… (see more) and may generate biased responses. In such circumstances, having a way to assess the reliability of a given response generated by a VLM is quite useful. Existing methods, such as estimating uncertainty using answer likelihoods or prompt-based confidence generation, often suffer from overconfidence. Other methods use self-consistency comparison but are affected by confirmation biases. To alleviate these, we propose \textbf{De}compose and \textbf{C}ompare \textbf{C}onsistency (\texttt{DeCC}) for reliability measurement. By comparing the consistency between the direct answer generated using the VLM's internal reasoning process, and the indirect answers obtained by decomposing the question into sub-questions and reasoning over the sub-answers produced by the VLM, \texttt{DeCC} measures the reliability of VLM's direct answer. Experiments across six vision-language tasks with three VLMs show \texttt{DeCC}'s reliability estimation achieves better correlation with task accuracy compared to the existing methods.
2024-07-09
ArXiv (preprint)
doi.org
arxiv.org
Guiding Language Model Reasoning with Planning Tokens
Xinyi Wang
Lucas Caccia
Oleksiy Ostapenko
Xingdi Yuan
William Yang Wang
Alessandro Sordoni
Large language models (LLMs) have recently attracted considerable interest for their ability to perform complex reasoning tasks, such as cha… (see more)in-of-thought (CoT) reasoning. However, most of the existing approaches to enhance this ability rely heavily on data-driven methods, while neglecting the structural aspects of the model's reasoning capacity. To encourage a more structural generation of CoT steps, we propose a hierarchical generation scheme: we let the LM generate a planning token at the start of each reasoning step, intuitively serving as a high-level plan of the current step, and add their embeddings to the model parameters. Our approach requires a negligible increase in trainable parameters (0.001%) and can be applied through either full fine-tuning or a more parameter-efficient scheme. We demonstrate our method's effectiveness by applying it to three different LLMs, showing notable accuracy improvements across three math word problem datasets and one multihop QA dataset with respect to standard fine-tuning baselines.
2024-07-09
colmweb.org/COLM/2024/Conference (accepted)
doi.org
openreview.net
LLM2Vec: Large Language Models Are Secretly Powerful Text Encoders
Parishad BehnamGhader
Vaibhav Adlakha
Marius Mosbach
Dzmitry Bahdanau
Nicolas Chapados
Siva Reddy
Large decoder-only language models (LLMs) are the state-of-the-art models on most of today's NLP tasks and benchmarks. Yet, the community is… (see more) only slowly adopting these models for text embedding tasks, which require rich contextualized representations. In this work, we introduce LLM2Vec, a simple unsupervised approach that can transform any decoder-only LLM into a strong text encoder. LLM2Vec consists of three simple steps: 1) enabling bidirectional attention, 2) masked next token prediction, and 3) unsupervised contrastive learning. We demonstrate the effectiveness of LLM2Vec by applying it to 4 popular LLMs ranging from 1.3B to 8B parameters and evaluate the transformed models on English word- and sequence-level tasks. We outperform encoder-only models by a large margin on word-level tasks and reach a new unsupervised state-of-the-art performance on the Massive Text Embeddings Benchmark (MTEB). Moreover, when combining LLM2Vec with supervised contrastive learning, we achieve state-of-the-art performance on MTEB among models that train only on publicly available data (as of May 24, 2024). Our strong empirical results and extensive analysis demonstrate that LLMs can be effectively transformed into universal text encoders in a parameter-efficient manner without the need for expensive adaptation or synthetic GPT-4 generated data.
2024-07-09
colmweb.org/COLM/2024/Conference (accepted)
doi.org
openreview.net
Redesigning Information Markets in the Era of Language Models
Martin Weiss
Nasim Rahaman
Manuel Wuthrich
Yoshua Bengio
Li Erran Li
Bernhard Schölkopf
Christopher Pal
2024-07-09
colmweb.org/COLM/2024/Conference (accepted)
openreview.net
openreview.net