Stéphane Marchand‐Maillet

Yuan Chen, Stéphane Marchand-Maillet

Deep hashing has shown to be a complexity-efficient solution for the Approximate Nearest Neighbor search problem in high dimensional space. Many methods usually build the loss function from pairwise or triplet data points to capture the local similarity structure. Other existing methods construct the similarity graph and consider all points simultaneously. Auto-encoding Twin-bottleneck Hashing is one such method that dynamically builds the graph. Specifically, each input data is encoded into a binary code and a continuous variable, or the so-called twin bottlenecks. The similarity graph is then computed from these binary codes, which get updated consistently during the training. In this work, we generalize the original model into a supervised deep hashing network by incorporating the label information. In addition, we examine the differences of codes structure between these two networks and consider the class imbalance problem especially in multi-labeled datasets. Experiments on three datasets yield statistically significant improvement against the original model. Results are also comparable and competitive to other supervised methods.

Nicola Messina, Giuseppe Amato, Andrea Esuli et al.

Despite the evolution of deep-learning-based visual-textual processing systems, precise multi-modal matching remains a challenging task. In this work, we tackle the task of cross-modal retrieval through image-sentence matching based on word-region alignments, using supervision only at the global image-sentence level. Specifically, we present a novel approach called Transformer Encoder Reasoning and Alignment Network (TERAN). TERAN enforces a fine-grained match between the underlying components of images and sentences, i.e., image regions and words, respectively, in order to preserve the informative richness of both modalities. TERAN obtains state-of-the-art results on the image retrieval task on both MS-COCO and Flickr30k datasets. Moreover, on MS-COCO, it also outperforms current approaches on the sentence retrieval task. Focusing on scalable cross-modal information retrieval, TERAN is designed to keep the visual and textual data pipelines well separated. Cross-attention links invalidate any chance to separately extract visual and textual features needed for the online search and the offline indexing steps in large-scale retrieval systems. In this respect, TERAN merges the information from the two domains only during the final alignment phase, immediately before the loss computation. We argue that the fine-grained alignments produced by TERAN pave the way towards the research for effective and efficient methods for large-scale cross-modal information retrieval. We compare the effectiveness of our approach against relevant state-of-the-art methods. On the MS-COCO 1K test set, we obtain an improvement of 5.7% and 3.5% respectively on the image and the sentence retrieval tasks on the Recall@1 metric. The code used for the experiments is publicly available on GitHub at https://github.com/mesnico/TERAN.

Mara Graziani, Sebastian Otalora, Stephane Marchand-Maillet et al.

Adopting Convolutional Neural Networks (CNNs) in the daily routine of primary diagnosis requires not only near-perfect precision, but also a sufficient degree of generalization to data acquisition shifts and transparency. Existing CNN models act as black boxes, not ensuring to the physicians that important diagnostic features are used by the model. Building on top of successfully existing techniques such as multi-task learning, domain adversarial training and concept-based interpretability, this paper addresses the challenge of introducing diagnostic factors in the training objectives. Here we show that our architecture, by learning end-to-end an uncertainty-based weighting combination of multi-task and adversarial losses, is encouraged to focus on pathology features such as density and pleomorphism of nuclei, e.g. variations in size and appearance, while discarding misleading features such as staining differences. Our results on breast lymph node tissue show significantly improved generalization in the detection of tumorous tissue, with best average AUC 0.89 (0.01) against the baseline AUC 0.86 (0.005). By applying the interpretability technique of linearly probing intermediate representations, we also demonstrate that interpretable pathology features such as nuclei density are learned by the proposed CNN architecture, confirming the increased transparency of this model. This result is a starting point towards building interpretable multi-task architectures that are robust to data heterogeneity. Our code is available at https://github.com/maragraziani/multitask_adversarial

Lorenzo Bini, Marco Sorbi, Stephane Marchand-Maillet

Graph Neural Networks (GNNs) have become increasingly popular for effectively modeling graph-structured data, and attention mechanisms have been pivotal in enabling these models to capture complex patterns. In our study, we reveal a critical yet underexplored consequence of integrating attention into edge-featured GNNs: the emergence of Massive Activations (MAs) within attention layers. By developing a novel method for detecting MAs on edge features, we show that these extreme activations are not only activation anomalies but encode domain-relevant signals. Our post-hoc interpretability analysis demonstrates that, in molecular graphs, MAs aggregate predominantly on common bond types (e.g., single and double bonds) while sparing more informative ones (e.g., triple bonds). Furthermore, our ablation studies confirm that MAs can serve as natural attribution indicators, reallocating to less informative edges. Our study assesses various edge-featured attention-based GNN models using benchmark datasets, including ZINC, TOX21, and PROTEINS. Key contributions include (1) establishing the direct link between attention mechanisms and MAs generation in edge-featured GNNs, (2) developing a robust definition and detection method for MAs enabling reliable post-hoc interpretability. Overall, our study reveals the complex interplay between attention mechanisms, edge-featured GNNs model, and MAs emergence, providing crucial insights for relating GNNs internals to domain knowledge.

Lorenzo Bini, Fatemeh Nassajian Mojarrad, Margarita Liarou et al.

This paper presents FlowCyt, the first comprehensive benchmark for multi-class single-cell classification in flow cytometry data. The dataset comprises bone marrow samples from 30 patients, with each cell characterized by twelve markers. Ground truth labels identify five hematological cell types: T lymphocytes, B lymphocytes, Monocytes, Mast cells, and Hematopoietic Stem/Progenitor Cells (HSPCs). Experiments utilize supervised inductive learning and semi-supervised transductive learning on up to 1 million cells per patient. Baseline methods include Gaussian Mixture Models, XGBoost, Random Forests, Deep Neural Networks, and Graph Neural Networks (GNNs). GNNs demonstrate superior performance by exploiting spatial relationships in graph-encoded data. The benchmark allows standardized evaluation of clinically relevant classification tasks, along with exploratory analyses to gain insights into hematological cell phenotypes. This represents the first public flow cytometry benchmark with a richly annotated, heterogeneous dataset. It will empower the development and rigorous assessment of novel methodologies for single-cell analysis.

Lorenzo Bini, Stephane Marchand-Maillet

We present LaplaceGNN, a novel self-supervised graph learning framework that bypasses the need for negative sampling by leveraging spectral bootstrapping techniques. Our method integrates Laplacian-based signals into the learning process, allowing the model to effectively capture rich structural representations without relying on contrastive objectives or handcrafted augmentations. By focusing on positive alignment, LaplaceGNN achieves linear scaling while offering a simpler, more efficient, self-supervised alternative for graph neural networks, applicable across diverse domains. Our contributions are twofold: we precompute spectral augmentations through max-min centrality-guided optimization, enabling rich structural supervision without relying on handcrafted augmentations, then we integrate an adversarial bootstrapped training scheme that further strengthens feature learning and robustness. Our extensive experiments on different benchmark datasets show that LaplaceGNN achieves superior performance compared to state-of-the-art self-supervised graph methods, offering a promising direction for efficiently learning expressive graph representations.

Lorenzo Bini, Stephane Marchand-Maillet

Generating high-fidelity and biologically plausible synthetic single-cell RNA sequencing (scRNA-seq) data, especially with conditional control, is challenging due to its high dimensionality, sparsity, and complex biological variations. Existing generative models often struggle to capture these unique characteristics and ensure robustness to structural noise in cellular networks. We introduce LapDDPM, a novel conditional Graph Diffusion Probabilistic Model for robust and high-fidelity scRNA-seq generation. LapDDPM uniquely integrates graph-based representations with a score-based diffusion model, enhanced by a novel spectral adversarial perturbation mechanism on graph edge weights. Our contributions are threefold: we leverage Laplacian Positional Encodings (LPEs) to enrich the latent space with crucial cellular relationship information; we develop a conditional score-based diffusion model for effective learning and generation from complex scRNA-seq distributions; and we employ a unique spectral adversarial training scheme on graph edge weights, boosting robustness against structural variations. Extensive experiments on diverse scRNA-seq datasets demonstrate LapDDPM's superior performance, achieving high fidelity and generating biologically-plausible, cell-type-specific samples. LapDDPM sets a new benchmark for conditional scRNA-seq data generation, offering a robust tool for various downstream biological applications.

Etienne Brangbour, Pierrick Bruneau, Thomas Tamisier et al.

We present novel active learning strategies dedicated to providing a solution to the cold start stage, i.e. initializing the classification of a large set of data with no attached labels. Moreover, proposed strategies are designed to handle an imbalanced context in which random selection is highly inefficient. Specifically, our active learning iterations address label scarcity and imbalance using element scores, combining information extracted from a clustering structure to a label propagation model. The strategy is illustrated by a case study on annotating Twitter content w.r.t. testimonies of a real flood event. We show that our method effectively copes with class imbalance, by boosting the recall of samples from the minority class.

Niccoló Marini, Manfredo Atzori, Sebastian Otálora et al.

Computational pathology is a domain that aims to develop algorithms to automatically analyze large digitized histopathology images, called whole slide images (WSI). WSIs are produced scanning thin tissue samples that are stained to make specific structures visible. They show stain colour heterogeneity due to different preparation and scanning settings applied across medical centers. Stain colour heterogeneity is a problem to train convolutional neural networks (CNN), the state-of-the-art algorithms for most computational pathology tasks, since CNNs usually underperform when tested on images including different stain variations than those within data used to train the CNN. Despite several methods that were developed, stain colour heterogeneity is still an unsolved challenge that limits the development of CNNs that can generalize on data from several medical centers. This paper aims to present a novel method to train CNNs that better generalize on data including several colour variations. The method, called H&E-adversarial CNN, exploits H&E matrix information to learn stain-invariant features during the training. The method is evaluated on the classification of colon and prostate histopathology images, involving eleven heterogeneous datasets, and compared with five other techniques used to handle stain colour heterogeneity. H&E-adversarial CNNs show an improvement in performance compared to the other algorithms, demonstrating that it can help to better deal with stain colour heterogeneous images.

Lionel Blondé, Alexandros Kalousis, Stéphane Marchand-Maillet

The performance of state-of-the-art offline RL methods varies widely over the spectrum of dataset qualities, ranging from far-from-optimal random data to close-to-optimal expert demonstrations. We re-implement these methods to test their reproducibility, and show that when a given method outperforms the others on one end of the spectrum, it never does on the other end. This prevents us from naming a victor across the board. We attribute the asymmetry to the amount of inductive bias injected into the agent to entice it to posit that the behavior underlying the offline dataset is optimal for the task. Our investigations confirm that careless injections of such optimality inductive biases make dominant agents subpar as soon as the offline policy is sub-optimal. To bridge this gap, we generalize importance-weighted regression methods that have proved the most versatile across the spectrum of dataset grades into a modular framework that allows for the design of methods that align with how much we know about the dataset. This modularity enables qualitatively different injections of optimality inductive biases. We show that certain orchestrations strike the right balance, improving the return on one end of the spectrum without harming it on the other end. While the formulation of guidelines for the design of an offline method reduces to aligning the amount of optimality bias to inject with what we know about the quality of the data, the design of an agnostic method for which we need not know the quality of the data beforehand is more nuanced. Only our framework allowed us to design a method that performed well across the spectrum while remaining modular if more information about the quality of the data ever becomes available.

Nicola Messina, Giuseppe Amato, Fabrizio Falchi et al.

Cross-modal retrieval is an important functionality in modern search engines, as it increases the user experience by allowing queries and retrieved objects to pertain to different modalities. In this paper, we focus on the image-sentence retrieval task, where the objective is to efficiently find relevant images for a given sentence (image-retrieval) or the relevant sentences for a given image (sentence-retrieval). Computer vision literature reports the best results on the image-sentence matching task using deep neural networks equipped with attention and self-attention mechanisms. They evaluate the matching performance on the retrieval task by performing sequential scans of the whole dataset. This method does not scale well with an increasing amount of images or captions. In this work, we explore different preprocessing techniques to produce sparsified deep multi-modal features extracting them from state-of-the-art deep-learning architectures for image-text matching. Our main objective is to lay down the paths for efficient indexing of complex multi-modal descriptions. We use the recently introduced TERN architecture as an image-sentence features extractor. It is designed for producing fixed-size 1024-d vectors describing whole images and sentences, as well as variable-length sets of 1024-d vectors describing the various building components of the two modalities (image regions and sentence words respectively). All these vectors are enforced by the TERN design to lie into the same common space. Our experiments show interesting preliminary results on the explored methods and suggest further experimentation in this important research direction.

Etienne Brangbour, Pierrick Bruneau, Stéphane Marchand-Maillet et al.

In this paper, we investigate the conversion of a Twitter corpus into geo-referenced raster cells holding the probability of the associated geographical areas of being flooded. We describe a baseline approach that combines a density ratio function, aggregation using a spatio-temporal Gaussian kernel function, and TFIDF textual features. The features are transformed to probabilities using a logistic regression model. The described method is evaluated on a corpus collected after the floods that followed Hurricane Harvey in the Houston urban area in August-September 2017. The baseline reaches a F1 score of 68%. We highlight research directions likely to improve these initial results.

Xavier Ouvrard, Jean-Marie Le Goff, Stéphane Marchand-Maillet

Traditional verbatim browsers give back information in a linear way according to a ranking performed by a search engine that may not be optimal for the surfer. The latter may need to assess the pertinence of the information retrieved, particularly when s$\cdot$he wants to explore other facets of a multi-facetted information space. For instance, in a multimedia dataset different facets such as keywords, authors, publication category, organisations and figures can be of interest. The facet simultaneous visualisation can help to gain insights on the information retrieved and call for further searches. Facets are co-occurence networks, modeled by HyperBag-Graphs -- families of multisets -- and are in fact linked not only to the publication itself, but to any chosen reference. These references allow to navigate inside the dataset and perform visual queries. We explore here the case of scientific publications based on Arxiv searches.

Pablo Strasser, Stephane Armand, Stephane Marchand-Maillet et al.

Complex structures are typical in machine learning. Tailoring learning algorithms for every structure requires an effort that may be saved by defining a generic learning procedure adaptive to any complex structure. In this paper, we propose to map any complex structure onto a generic form, called serialization, over which we can apply any sequence-based density estimator. We then show how to transfer the learned density back onto the space of original structures. To expose the learning procedure to the structural particularities of the original structures, we take care that the serializations reflect accurately the structures' properties. Enumerating all serializations is infeasible. We propose an effective way to sample representative serializations from the complete set of serializations which preserves the statistics of the complete set. Our method is competitive or better than state of the art learning algorithms that have been specifically designed for given structures. In addition, since the serialization involves sampling from a combinatorial process it provides considerable protection from overfitting, which we clearly demonstrate on a number of experiments.

Etienne Brangbour, Pierrick Bruneau, Stéphane Marchand-Maillet et al.

In this paper, we discuss the collection of a corpus associated to tropical storm Harvey, as well as its analysis from both spatial and topical perspectives. From the spatial perspective, our goal here is to get a first estimation of the quality and precision of the geographical information featured in the collected corpus. From a topical perspective, we discuss the representation of Twitter posts, and strategies to process an initially unlabeled corpus of tweets.

Xavier Ouvrard, Jean-Marie Le Goff, Stephane Marchand-Maillet

Most networks tend to show complex and multiple relationships between entities. Networks are usually modeled by graphs or hypergraphs; nonetheless a given entity can occur many times in a relationship: this brings the need to deal with multisets instead of sets or simple edges. Diffusion processes are useful to highlight interesting parts of a network: they usually start with a stroke at one vertex and diffuse throughout the network to reach a uniform distribution. Several iterations of the process are required prior to reaching a stable solution. We propose an alternative solution to highlighting the main components of a network using a diffusion process based on exchanges: it is an iterative two-phase step exchange process. This process allows to evaluate the importance not only of the vertices but also of the regrouping level. To model the diffusion process, we extend the concept of hypergraphs that are families of sets to families of multisets, that we call hb-graphs. This version is an extended version of arXiv:1809.00190v1: the overlaps with the v1 are in black, the new content is in blue. The contributions of this extended version are: the proofs of conservation and convergence of the extracted sequences of the diffusion process, as well as the illustration of the speed of convergence and comparison to classical and modified random walks; the algorithms of the exchange-based diffusion and the modified random walk; the application to a use case based on Arxiv publications. All the figures except one have been either modified or added in this extended version to take into account the new developments.

Magda Gregorová, Alexandros Kalousis, Stéphane Marchand-Maillet

We investigate structured sparsity methods for variable selection in regression problems where the target depends nonlinearly on the inputs. We focus on general nonlinear functions not limiting a priori the function space to additive models. We propose two new regularizers based on partial derivatives as nonlinear equivalents of group lasso and elastic net. We formulate the problem within the framework of learning in reproducing kernel Hilbert spaces and show how the variational problem can be reformulated into a more practical finite dimensional equivalent. We develop a new algorithm derived from the ADMM principles that relies solely on closed forms of the proximal operators. We explore the empirical properties of our new algorithm for Nonlinear Variable Selection based on Derivatives (NVSD) on a set of experiments and confirm favourable properties of our structured-sparsity models and the algorithm in terms of both prediction and variable selection accuracy.

Magda Gregorová, Jason Ramapuram, Alexandros Kalousis et al.

We propose a new method for input variable selection in nonlinear regression. The method is embedded into a kernel regression machine that can model general nonlinear functions, not being a priori limited to additive models. This is the first kernel-based variable selection method applicable to large datasets. It sidesteps the typical poor scaling properties of kernel methods by mapping the inputs into a relatively low-dimensional space of random features. The algorithm discovers the variables relevant for the regression task together with learning the prediction model through learning the appropriate nonlinear random feature maps. We demonstrate the outstanding performance of our method on a set of large-scale synthetic and real datasets.

Magda Gregorova, Alexandros Kalousis, Stephane Marchand-Maillet

We present a new method for forecasting systems of multiple interrelated time series. The method learns the forecast models together with discovering leading indicators from within the system that serve as good predictors improving the forecast accuracy and a cluster structure of the predictive tasks around these. The method is based on the classical linear vector autoregressive model (VAR) and links the discovery of the leading indicators to inferring sparse graphs of Granger causality. We formulate a new constrained optimisation problem to promote the desired sparse structures across the models and the sharing of information amongst the learning tasks in a multi-task manner. We propose an algorithm for solving the problem and document on a battery of synthetic and real-data experiments the advantages of our new method over baseline VAR models as well as the state-of-the-art sparse VAR learning methods.

Magda Gregorová, Alexandros Kalousis, Stéphane Marchand-Maillet

Traditional linear methods for forecasting multivariate time series are not able to satisfactorily model the non-linear dependencies that may exist in non-Gaussian series. We build on the theory of learning vector-valued functions in the reproducing kernel Hilbert space and develop a method for learning prediction functions that accommodate such non-linearities. The method not only learns the predictive function but also the matrix-valued kernel underlying the function search space directly from the data. Our approach is based on learning multiple matrix-valued kernels, each of those composed of a set of input kernels and a set of output kernels learned in the cone of positive semi-definite matrices. In addition to superior predictive performance in the presence of strong non-linearities, our method also recovers the hidden dynamic relationships between the series and thus is a new alternative to existing graphical Granger techniques.

Magda Gregorova, Alexandros Kalousis, Stéphane Marchand-Maillet

We consider the problem of learning models for forecasting multiple time-series systems together with discovering the leading indicators that serve as good predictors for the system. We model the systems by linear vector autoregressive models (VAR) and link the discovery of leading indicators to inferring sparse graphs of Granger-causality. We propose new problem formulations and develop two new methods to learn such models, gradually increasing the complexity of assumptions and approaches. While the first method assumes common structures across the whole system, our second method uncovers model clusters based on the Granger-causality and leading indicators together with learning the model parameters. We study the performance of our methods on a comprehensive set of experiments and confirm their efficacy and their advantages over state-of-the-art sparse VAR and graphical Granger learning methods.