Nicolas Dugué

Thibault Prouteau, Francis Lareau, Nicolas Dugué et al.

Topic models uncover latent thematic structures in text corpora, yet evaluating their quality remains challenging, particularly in specialized domains. Existing methods often rely on automated metrics like topic coherence and diversity, which may not fully align with human judgment. Human evaluation tasks, such as word intrusion, provide valuable insights but are costly and primarily validated on general-domain corpora. This paper introduces Topic Word Mixing (TWM), a novel human evaluation task assessing inter-topic distinctness by testing whether annotators can distinguish between word sets from single or mixed topics. TWM complements word intrusion's focus on intra-topic coherence and provides a human-grounded counterpart to diversity metrics. We evaluate six topic models - both statistical and embedding-based (LDA, NMF, Top2Vec, BERTopic, CFMF, CFMF-emb) - comparing automated metrics with human evaluation methods based on nearly 4,000 annotations from a domain-specific corpus of philosophy of science publications. Our findings reveal that word intrusion and coherence metrics do not always align, particularly in specialized domains, and that TWM captures human-perceived distinctness while appearing to align with diversity metrics. We release the annotated dataset and task generation code. This work highlights the need for evaluation frameworks bridging automated and human assessments, particularly for domain-specific corpora.

Thibault Prouteau, Nicolas Dugué, Simon Guillot

Modelling information from complex systems such as humans social interaction or words co-occurrences in our languages can help to understand how these systems are organized and function. Such systems can be modelled by networks, and network theory provides a useful set of methods to analyze them. Among these methods, graph embedding is a powerful tool to summarize the interactions and topology of a network in a vectorized feature space. When used in input of machine learning algorithms, embedding vectors help with common graph problems such as link prediction, graph matching, etc. Word embedding has the goal of representing the sense of words, extracting it from large text corpora. Despite differences in the structure of information in input of embedding algorithms, many graph embedding approaches are adapted and inspired from methods in NLP. Limits of these methods are observed in both domains. Most of these methods require long and resource greedy training. Another downside to most methods is that they are black-box, from which understanding how the information is structured is rather complex. Interpretability of a model allows understanding how the vector space is structured without the need for external information, and thus can be audited more easily. With both these limitations in mind, we propose a novel framework to efficiently embed network vertices in an interpretable vector space. Our Lower Dimension Bipartite Framework (LDBGF) leverages the bipartite projection of a network using cliques to reduce dimensionality. Along with LDBGF, we introduce two implementations of this framework that rely on communities instead of cliques: SINr-NR and SINr-MF. We show that SINr-MF can perform well on classical graphs and SINr-NR can produce high-quality graph and word embeddings that are interpretable and stable across runs.

Antoine Teissier, Marie Tahon, Nicolas Dugué et al.

Due to the rapid progress of speech synthesis, deepfake detection has become a major concern in the speech processing community. Because it is a critical task, systems must not only be efficient and robust, but also provide interpretable explanations. Among the different approaches for explainability, we focus on the interpretation of latent representations. In such paper, we focus on the last layer of embeddings of AASIST, a deepfake detection architecture. We use a TopK activation inspired by SAEs on this layer to obtain sparse representations which are used in the decision process. We demonstrate that sparse deepfake detection can improve detection performance, with an EER of 23.36% on ASVSpoof5 test set, with 95% of sparsity. We then show that these representations provide better disentanglement, using completeness and modularity metrics based on mutual information. Notably, some attacks are directly encoded in the latent space.

Théo Mariotte, Martin Lebourdais, Antonio Almudévar et al.

Audio pretrained models are widely employed to solve various tasks in speech processing, sound event detection, or music information retrieval. However, the representations learned by these models are unclear, and their analysis mainly restricts to linear probing of the hidden representations. In this work, we explore the use of Sparse Autoencoders (SAEs) to analyze the hidden representations of pretrained models, focusing on a case study in singing technique classification. We first demonstrate that SAEs retain both information about the original representations and class labels, enabling their internal structure to provide insights into self-supervised learning systems. Furthermore, we show that SAEs enhance the disentanglement of vocal attributes, establishing them as an effective tool for identifying the underlying factors encoded in the representations.

Nicolas Dugué, Victor Connes

Most of the time, the first step to learn word embeddings is to build a word co-occurrence matrix. As such matrices are equivalent to graphs, complex networks theory can naturally be used to deal with such data. In this paper, we consider applying community detection, a main tool of this field, to the co-occurrence matrix corresponding to a huge corpus. Community structure is used as a way to reduce the dimensionality of the initial space. Using this community structure, we propose a method to extract word embeddings that are comparable to the state-of-the-art approaches.

Jean-Valère Cossu, Vincent Labatut, Nicolas Dugué

Many works related to Twitter aim at characterizing its users in some way: role on the service (spammers, bots, organizations, etc.), nature of the user (socio-professional category, age, etc.), topics of interest , and others. However, for a given user classification problem, it is very difficult to select a set of appropriate features, because the many features described in the literature are very heterogeneous, with name overlaps and collisions, and numerous very close variants. In this article, we review a wide range of such features. In order to present a clear state-of-the-art description, we unify their names, definitions and relationships, and we propose a new, neutral, typology. We then illustrate the interest of our review by applying a selection of these features to the offline influence detection problem. This task consists in identifying users which are influential in real-life, based on their Twitter account and related data. We show that most features deemed efficient to predict online influence, such as the numbers of retweets and followers, are not relevant to this problem. However, We propose several content-based approaches to label Twitter users as Influencers or not. We also rank them according to a predicted influence level. Our proposals are evaluated over the CLEF RepLab 2014 dataset, and outmatch state-of-the-art methods.