In the last decade, we have witnessed the introduction of several novel deep neural network (DNN) architectures exhibiting ever-increasing performance across diverse tasks. Explaining the upward trend of their performance, however, remains difficult as different DNN architectures of comparable depth and width – common factors associated with their expressive power – may exhibit a drastically different performance even when trained on the same dataset. In this paper, we introduce the concept of the non-linearity signature of DNN, the first theoretically sound solution for approximately measuring the non-linearity of deep neural networks. Built upon a score derived from closed-form optimal transport mappings, this signature provides a better understanding of the inner workings of a wide range of DNN architectures and learning paradigms, with a particular emphasis on the computer vision task. We provide extensive experimental results that highlight the practical usefulness of the proposed non-linearity signature and its potential for long-reaching implications. The code for our work is available at https://github.com/qbouniot/AffScoreDeep.

Thibaud Leteno, Irina Proskurina, Antoine Gourru, Julien Velcin, Charlotte Laclau, Guillaume Metzler, Christophe Gravier. Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers). 2025.

Although graph-based learning has attracted a lot of attention, graph representation learning is still a challenging task whose resolution may impact key application fields such as chemistry or biology. To this end, we introduce GRALE, a novel graph autoencoder that encodes and decodes graphs of varying sizes into a shared embedding space. GRALE is trained using an Optimal Transport-inspired loss that compares the source and reconstructed graphs and leverages a differentiable matching module, which is trained jointly with the encoder and decoder. The proposed attention-based architecture relies on Evoformer, the core component of AlphaFold, which we extend to support both graph encoding and decoding. We show, in numerical experiments on simulated and molecular data, that GRALE enables a highly general form of pre-training, applicable to a wide range of downstream tasks, from classification and regression to more complex tasks such as graph interpolation, editing, matching, and prediction.

Abstract Temporal networks allow representing connections between objects while incorporating the temporal dimension. While static network models can capture unchanging topological regularities, they often fail to model the effects associated with the causal generative process of the network that occurs in time. Hence, exploiting the temporal aspect of networks has been the focus of many recent studies. In this context, we propose a new framework for generative models of continuous-time temporal networks. We assume that the activation of the edges in a temporal network is driven by a specified temporal point process. This approach allows to directly model the waiting time between events while incorporating time-varying history-based features as covariates in the predictions. Coupled with a thinning algorithm designed for the simulation of point processes, SimHawNet enables simulation of the evolution of temporal networks in continuous time. Finally, we introduce a comprehensive evaluation framework to assess the performance of such an approach, in which we demonstrate that SimHawNet successfully simulates the evolution of networks with very different generative processes and achieves performance comparable to the state of the art, while being significantly faster.

Group fairness is a central research topic in text classification, where reaching fair treatment between sensitive groups (e.g., women and men) remains an open challenge. We propose an approach that extends the use of the Wasserstein Dependency Measure for learning unbiased neural text classifiers. Given the challenge of distinguishing fair from unfair information in a text encoder, we draw inspiration from adversarial training by inducing independence between representations learned for the target label and those for a sensitive attribute. We further show that domain adaptation can be efficiently leveraged to remove the need for access to the sensitive attributes in the data set we cure. We provide both theoretical and empirical evidence that our approach is well-founded.

We propose Any2graph, a generic framework for end-to-end Supervised Graph Prediction (SGP) i.e. a deep learning model that predicts an entire graph for any kind of input. The framework is built on a novel Optimal Transport loss, the Partially-Masked Fused Gromov-Wasserstein, that exhibits all necessary properties (permutation invariance, differentiability and scalability) and is designed to handle any-sized graphs. Numerical experiments showcase the versatility of the approach that outperform existing competitors on a novel challenging synthetic dataset and a variety of real-world tasks such as map construction from satellite image (Sat2Graph) or molecule prediction from fingerprint (Fingerprint2Graph).

In recent years, there has been a significant surge in machine learning techniques, particularly in the domain of deep learning, tailored for handling attributed graphs. Nevertheless, to work, these methods assume that the attributes values are fully known, which is not realistic in numerous real-world applications. This paper explores the potential of Optimal Transport (OT) to impute missing attributes on graphs. To proceed, we design a novel multi-view OT loss function that can encompass both node feature data and the underlying topological structure of the graph by utilizing multiple graph representations. We then utilize this novel loss to train efficiently a Graph Convolutional Neural Network (GCN) architecture capable of imputing all missing values over the graph at once. We evaluate the interest of our approach with experiments both on synthetic data and real-world graphs, including different missingness mechanisms and a wide range of missing data. These experiments demonstrate that our method is competitive with the state-of-the-art in all cases and of particular interest on weakly homophilic graphs.

Group fairness is a central research topic in text classification, where reaching fair treatment between sensitive groups (e.g.women vs. men) remains an open challenge.This paper presents a novel method for mitigating biases in neural text classification, agnostic to the model architecture.Considering the difficulty to distinguish fair from unfair information in a text encoder, we take inspiration from adversarial training to induce Wasserstein independence between representations learned to predict our target label and the ones learned to predict some sensitive attribute.Our approach provides two significant advantages.Firstly, it does not require annotations of sensitive attributes in both testing and training data.This is more suitable for real-life scenarios compared to existing methods that require annotations of sensitive attributes at train time.Secondly, our approach exhibits a comparable or better fairness-accuracy trade-off compared to existing methods.Our implementation is available on Github 1 .

Machine learning and data mining algorithms have been increasingly used recently to support decision-making systems in many areas of high societal importance such as healthcare, education, or security. While being very efficient in their predictive abilities, the deployed algorithms sometimes tend to learn an inductive model with a discriminative bias due to the presence of this latter in the learning sample. This problem gave rise to a new field of algorithmic fairness where the goal is to correct the discriminative bias introduced by a certain attribute in order to decorrelate it from the model’s output. In this paper, we study the problem of fairness for the task of edge prediction in graphs, a largely underinvestigated scenario compared to a more popular setting of fair classification. To this end, we formulate the problem of fair edge prediction, analyze it theoretically, and propose an embedding-agnostic repairing procedure for the adjacency matrix of an arbitrary graph with a trade-off between the group and individual fairness. We experimentally show the versatility of our approach and its capacity to provide explicit control over different notions of fairness and prediction accuracy.

The theoretical analysis of deep neural networks (DNN) is arguably among the most challenging research directions in machine learning (ML) right now, as it requires from scientists to lay novel statistical learning foundations to explain their behaviour in practice. While some success has been achieved recently in this endeavour, the question on whether DNNs can be analyzed using the tools from other scientific fields outside the ML community has not received the attention it may well have deserved. In this paper, we explore the interplay between DNNs and game theory (GT), and show how one can benefit from the classic readily available results from the latter when analyzing the former. In particular, we consider the widely studied class of congestion games, and illustrate their intrinsic relatedness to both linear and non-linear DNNs and to the properties of their loss surface. Beyond retrieving the state-of-the-art results from the literature, we argue that our work provides a very promising novel tool for analyzing the DNNs and support this claim by proposing concrete open problems that can advance significantly our understanding of DNNs when solved.

In this paper, we propose a theoretically supported sequential strategy for training a large-scale Recommender System (RS) over implicit feedback, mainly in the form of clicks. The proposed approach consists in minimizing pairwise ranking loss over blocks of consecutive items constituted by a sequence of non-clicked items followed by a clicked one for each user. We present two variants of this strategy where model parameters are updated using either the momentum method or a gradient-based approach. To prevent updating the parameters for an abnormally high number of clicks over some targeted items (mainly due to bots), we introduce an upper and a lower threshold on the number of updates for each user. These thresholds are estimated over the distribution of the number of blocks in the training set. They affect the decision of RS by shifting the distribution of items that are shown to the users. Furthermore, we provide a convergence analysis of both algorithms and demonstrate their practical efficiency over six large-scale collections with respect to various ranking measures and computational time.

In this paper, we present a novel and publicly available dataset for online recommendation provided by Purch1. The dataset records the clicks generated by users of one of Purch's high-tech website over the ads they have been shown for one month. In addition, the dataset contains contextual information about offers such as offer titles and keywords, as well as the anonymized content of the page on which offers were displayed. Then, besides a detailed description of the dataset, we evaluate the performance of six popular baselines and propose a simple yet effective strategy on how to overcome the existing challenges inherent to implicit feedback and popularity bias introduced while designing an efficient and scalable recommendation algorithm. More specifically, we propose to demonstrate the importance of introducing diversity based on an appropriate representation of items in Recommender Systems, when the available feedback is strongly biased.

In this paper, we present a novel method for co-clustering, an unsupervised learning approach that aims at discovering homogeneous groups of data instances and features by grouping them simultaneously. The proposed method uses the entropy regularized optimal transport between empirical measures defined on data instances and features in order to obtain an estimated joint probability density function represented by the optimal coupling matrix. This matrix is further factorized to obtain the induced row and columns partitions using multiscale representations approach. To justify our method theoretically, we show how the solution of the regularized optimal transport can be seen from the variational inference perspective thus motivating its use for co-clustering. The algorithm derived for the proposed method and its kernelized version based on the notion of Gromov-Wasserstein distance are fast, accurate and can determine automatically the number of both row and column clusters. These features are vividly demonstrated through extensive experimental evaluations.

In this paper, we describe a novel, publicly available collection for recommendation systems that records the behavior of customers of the European leader in eCommerce advertising, Kelkoo\footnote{\url{https://www.kelkoo.com/}}, during one month. This dataset gathers implicit feedback, in form of clicks, of users that have interacted with over 56 million offers displayed by Kelkoo, along with a rich set of contextual features regarding both customers and offers. In conjunction with a detailed description of the dataset, we show the performance of six state-of-the-art recommender models and raise some questions on how to encompass the existing contextual information in the system.

We propose two fuzzy co-clustering algorithms based on the double Kmeans algorithm. Fuzzy approaches are known to require more computation time than hard ones but the fuzziness principle allows a description of uncertainties that often appears in real world applications. The first algorithm proposed, fuzzy double Kmeans (FDK) is a fuzzy version of double Kmeans (DK). The second algorithm, weighted fuzzy double Kmeans (W-FDK), is an extension of FDK with automated variable weighting allowing co-clustering and feature selection simultaneously. We illustrate our contribution using Monte Carlo simulations on datasets with different parameters and real datasets commonly used in the co-clustering context.