11 papers
RunningTensor: Generalizing Linear Attention to Higher-Order Recurrent States
Luca Herranz-Celotti, Vincent Guigue
Linear attention and state-space models provide linear-time sequence modeling, but their recurrent memory remains a second-order tensor (a matrix), limiting the order of interactio…
Semantic Space Search Trajectory Networks
Julian Agudelo, Alberto Tonda, Gabriela Ochoa +3
Search Trajectory Networks (STNs) are a graph-based tool for visualizing and characterizing the behavior of optimization algorithms. STNs' reliance on discretization of the search…
RAC: Retrieval-Augmented Clarification for Faithful Conversational Search
Ahmed Rayane Kebir, Vincent Guigue, Lynda Said Lhadj +1
Clarification questions help conversational search systems resolve ambiguous or underspecified user queries. While prior work has focused on fluency and alignment with user intent,…
SEval-Ex: A Statement-Level Framework for Explainable Summarization Evaluation
Tanguy Herserant, Vincent Guigue
Evaluating text summarization quality remains a critical challenge in Natural Language Processing. Current approaches face a trade-off between performance and interpretability. We…
Clarifying Ambiguities: on the Role of Ambiguity Types in Prompting Methods for Clarification Generation
Anfu Tang, Laure Soulier, Vincent Guigue
In information retrieval (IR), providing appropriate clarifications to better understand users' information needs is crucial for building a proactive search-oriented dialogue syste…
Drought forecasting using a hybrid neural architecture for integrating time series and static data
Julian Agudelo, Vincent Guigue, Cristina Manfredotti +1
Reliable forecasting is critical for early warning systems and adaptive drought management. Most previous deep learning approaches focus solely on homogeneous regions and rely on s…
Towards Lighter and Robust Evaluation for Retrieval Augmented Generation
Alex-Razvan Ispas, Charles-Elie Simon, Fabien Caspani +1
Large Language Models are prompting us to view more NLP tasks from a generative perspective. At the same time, they offer a new way of accessing information, mainly through the RAG…
SCOPE: A Self-supervised Framework for Improving Faithfulness in Conditional Text Generation
Song Duong, Florian Le Bronnec, Alexandre Allauzen +4
Large Language Models (LLMs), when used for conditional text generation, often produce hallucinations, i.e., information that is unfaithful or not grounded in the input context. Th…
Navigating Uncertainty: Optimizing API Dependency for Hallucination Reduction in Closed-Book Question Answering
Pierre Erbacher, Louis Falissar, Vincent Guigue +1
While Large Language Models (LLM) are able to accumulate and restore knowledge, they are still prone to hallucination. Especially when faced with factual questions, LLM cannot only…
LOCOST: State-Space Models for Long Document Abstractive Summarization
Florian Le Bronnec, Song Duong, Mathieu Ravaut +6
State-space models are a low-complexity alternative to transformers for encoding long sequences and capturing long-term dependencies. We propose LOCOST: an encoder-decoder architec…
Improving generalization in large language models by learning prefix subspaces
Louis Falissard, Vincent Guigue, Laure Soulier
This article focuses on large language models (LLMs) fine-tuning in the scarce data regime (also known as the "few-shot" learning setting). We propose a method to increase the gene…