Vincent Guigue
YOU?
Author Swipe
View article: AllSummedUp: un framework open-source pour comparer les metriques d'evaluation de resume
AllSummedUp: un framework open-source pour comparer les metriques d'evaluation de resume Open
This paper investigates reproducibility challenges in automatic text summarization evaluation. Based on experiments conducted across six representative metrics ranging from classical approaches like ROUGE to recent LLM-based methods (G-Eva…
View article: SEval-Ex: A Statement-Level Framework for Explainable Summarization Evaluation
SEval-Ex: A Statement-Level Framework for Explainable Summarization Evaluation Open
Evaluating text summarization quality remains a critical challenge in Natural Language Processing. Current approaches face a trade-off between performance and interpretability. We present SEval-Ex, a framework that bridges this gap by deco…
View article: Clarifying Ambiguities: on the Role of Ambiguity Types in Prompting Methods for Clarification Generation
Clarifying Ambiguities: on the Role of Ambiguity Types in Prompting Methods for Clarification Generation Open
In information retrieval (IR), providing appropriate clarifications to better understand users' information needs is crucial for building a proactive search-oriented dialogue system. Due to the strong in-context learning ability of large l…
View article: Drought forecasting using a hybrid neural architecture for integrating time series and static data
Drought forecasting using a hybrid neural architecture for integrating time series and static data Open
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 single-structured data. This paper presents a hybrid neural …
View article: Towards Lighter and Robust Evaluation for Retrieval Augmented Generation
Towards Lighter and Robust Evaluation for Retrieval Augmented Generation Open
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 framework. While there have been notable improvements for …
View article: SCOPE: A Self-supervised Framework for Improving Faithfulness in Conditional Text Generation
SCOPE: A Self-supervised Framework for Improving Faithfulness in Conditional Text Generation Open
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. This issue arises in typical conditional text generation task…
View article: Navigating Uncertainty: Optimizing API Dependency for Hallucination Reduction in Closed-Book Question Answering
Navigating Uncertainty: Optimizing API Dependency for Hallucination Reduction in Closed-Book Question Answering Open
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 rely on knowledge stored in parameters to guarantee truthf…
View article: LOCOST: State-Space Models for Long Document Abstractive Summarization
LOCOST: State-Space Models for Long Document Abstractive Summarization Open
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 architecture based on state-space models for conditional text gener…
View article: Interpretable time series neural representation for classification purposes
Interpretable time series neural representation for classification purposes Open
Deep learning has made significant advances in creating efficient representations of time series data by automatically identifying complex patterns. However, these approaches lack interpretability, as the time series is transformed into a …
View article: Improving generalization in large language models by learning prefix subspaces
Improving generalization in large language models by learning prefix subspaces Open
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 generalization capabilities of LLMs based on neural network sub…
View article: Interpretable time series neural representation for classification purposes
Interpretable time series neural representation for classification purposes Open
Deep learning has made significant advances in creating efficient representations of time series data by automatically identifying complex patterns. However, these approaches lack interpretability, as the time series is transformed into a …
View article: Of Spiky SVDs and Music Recommendation
Of Spiky SVDs and Music Recommendation Open
International audience
View article: Leveraging Multimodality for Biodiversity Data: Exploring joint representations of species descriptions and specimen images using CLIP
Leveraging Multimodality for Biodiversity Data: Exploring joint representations of species descriptions and specimen images using CLIP Open
In recent years, the field of biodiversity data analysis has witnessed significant advancements, with a number of models emerging to process and extract valuable insights from various data sources. One notable area of progress lies in the …
View article: Of Spiky SVDs and Music Recommendation
Of Spiky SVDs and Music Recommendation Open
The truncated singular value decomposition is a widely used methodology in music recommendation for direct similar-item retrieval or embedding musical items for downstream tasks. This paper investigates a curious effect that we show natura…
View article: Time Series Continuous Modeling for Imputation and Forecasting with Implicit Neural Representations
Time Series Continuous Modeling for Imputation and Forecasting with Implicit Neural Representations Open
We introduce a novel modeling approach for time series imputation and forecasting, tailored to address the challenges often encountered in real-world data, such as irregular samples, missing data, or unaligned measurements from multiple se…
View article: Dynamic Named Entity Recognition
Dynamic Named Entity Recognition Open
Named Entity Recognition (NER) is a challenging and widely studied task that\ninvolves detecting and typing entities in text. So far,NER still approaches\nentity typing as a task of classification into universal classes (e.g. date,\nperson…
View article: Improving generalization in large langue model by learning prefix subspaces
Improving generalization in large langue model by learning prefix subspaces Open
This article focuses on large language models (LLMs) fine-tuning in the scarce data regime (also known as “few-shot learning setting”). We propose a method to increase the generalization capabilities of LLMs based on neural network subspac…
View article: Learning Unsupervised Hierarchies of Audio Concepts
Learning Unsupervised Hierarchies of Audio Concepts Open
Music signals are difficult to interpret from their low-level features, perhaps even more than images: e.g. highlighting part of a spectrogram or an image is often insufficient to convey high-level ideas that are genuinely relevant to huma…
View article: Learning Unsupervised Hierarchies of Audio Concepts
Learning Unsupervised Hierarchies of Audio Concepts Open
Dataset of our paper "Learning Unsupervised Hierarchies of Audio Concepts" published at the ISMIR 2022 conference. For usage, please refer to our code repository at github.com/deezer/concept_hierarchy. Paper abstract Music signals are diff…
View article: Learning Unsupervised Hierarchies of Audio Concepts
Learning Unsupervised Hierarchies of Audio Concepts Open
Dataset of our paper "Learning Unsupervised Hierarchies of Audio Concepts" published at the ISMIR 2022 conference. For usage, please refer to our code repository at github.com/deezer/concept_hierarchy. Paper abstract Music signals are diff…
View article: NEARSIDE: Structured kNowledge Extraction frAmework from SpecIes DEscriptions
NEARSIDE: Structured kNowledge Extraction frAmework from SpecIes DEscriptions Open
Species descriptions are stored in textual form in corpora such as in floras and faunas, but this large amount of information cannot be used directly by algorithms, nor can it be linked to other data sources. The production of knowledge ba…
View article: Separating Retention from Extraction in the Evaluation of End-to-end Relation Extraction
Separating Retention from Extraction in the Evaluation of End-to-end Relation Extraction Open
State-of-the-art NLP models can adopt shallow heuristics that limit their generalization capability (McCoy et al., 2019). Such heuristics include lexical overlap with the training set in Named-Entity Recognition (Taillé et al., 2020) and E…
View article: Coping with the Document Frequency Bias in Sentiment Classification
Coping with the Document Frequency Bias in Sentiment Classification Open
In this article, we study the polarity detection problem using linear supervised classifiers. We show the interest of penalizing the document frequencies in the regularization process to increase the accuracy. We propose a systematic compa…
View article: Towards Rigorous Interpretations: a Formalisation of Feature Attribution
Towards Rigorous Interpretations: a Formalisation of Feature Attribution Open
International audience
View article: Towards Rigorous Interpretations: a Formalisation of Feature Attribution
Towards Rigorous Interpretations: a Formalisation of Feature Attribution Open
Feature attribution is often loosely presented as the process of selecting a subset of relevant features as a rationale of a prediction. Task-dependent by nature, precise definitions of "relevance" encountered in the literature are however…
View article: Separating Retention from Extraction in the Evaluation of End-to-end Relation Extraction
Separating Retention from Extraction in the Evaluation of End-to-end Relation Extraction Open
State-of-the-art NLP models can adopt shallow heuristics that limit their\ngeneralization capability (McCoy et al., 2019). Such heuristics include lexical\noverlap with the training set in Named-Entity Recognition (Taill\\'e et al.,\n2020)…
View article: Let's Stop Incorrect Comparisons in End-to-end Relation Extraction!
Let's Stop Incorrect Comparisons in End-to-end Relation Extraction! Open
Despite efforts to distinguish three different evaluation setups (Bekoulis et al., 2018), numerous end-to-end Relation Extraction (RE) articles present unreliable performance comparison to previous work. In this paper, we first identify se…