Dorina Thanou

Vasiliki Rizou, Pascal Frossard, Dorina Thanou

To leverage the full potential of multimodal data, we need representations that go beyond the state-of-the-art alignment and fusion approaches and exploit all cross-modal interactions without sacrificing modality-specific information. Learning disentangled representations is a principled way to identify these underlying shared and unique factors that are hidden in observational data. However, while multimodal disentanglement is a compelling paradigm, existing methods are largely confined to the two-modality regime due to its inherent scalability bottleneck. To address this, we propose RePercENT, a self-supervised framework designed to surpass these limitations and unlocks scalable pairwise disentanglement beyond two modalities. Through a multimodal `plug-and-play' architecture, our approach operates directly on pre-extracted embeddings, eliminating the need for extensive joint pre-training while making no assumptions regarding the underlying modalities or foundation model backbones. Moreover, we introduce a joint optimization objective for simultaneously deriving the shared and unique components, and provide formal theoretical guarantees that characterize the optimality of our solution. Across diverse modalities and tasks, RePercENT successfully recovers disentangled components while maintaining competitive performance and significantly reducing computational complexity.

Anand Choudhary, Xiaowu Sun, Thabo Mahendiran et al.

Invasive Coronary Angiography (ICA) is the clinical gold standard for the assessment of coronary artery disease. However, its interpretation remains subjective and prone to intra- and inter-operator variability. In this work, we introduce ODySSeI: an Open-source end-to-end framework for automated Detection, Segmentation, and Severity estimation of lesions in ICA images. ODySSeI integrates deep learning-based lesion detection and lesion segmentation models trained using a novel Pyramidal Augmentation Scheme (PAS) to enhance robustness and real-time performance across diverse patient cohorts (2149 patients from Europe, North America, and Asia). Furthermore, we propose a quantitative coronary angiography-free Lesion Severity Estimation (LSE) technique that directly computes the Minimum Lumen Diameter (MLD) and diameter stenosis from the predicted lesion geometry. Extensive evaluation on both in-distribution and out-of-distribution clinical datasets demonstrates ODySSeI's strong generalizability. Our PAS yields large performance gains in highly complex tasks as compared to relatively simpler ones, notably, a 2.5-fold increase in lesion detection performance versus a 1-3\% increase in lesion segmentation performance over their respective baselines. Our LSE technique achieves high accuracy, with predicted MLD values differing by only $\pm$ 2-3 pixels from the corresponding ground truths. On average, ODySSeI processes a raw ICA image within only a few seconds on a CPU and in a fraction of a second on a GPU and is available as a plug-and-play web interface at swisscardia.epfl.ch. Overall, this work establishes ODySSeI as a comprehensive and open-source framework which supports automated, reproducible, and scalable ICA analysis for real-time clinical decision-making.

Manuel Madeira, Dorina Thanou, Pascal Frossard

Graph-based representation approaches have been proven to be successful in the analysis of biomedical data, due to their capability of capturing intricate dependencies between biological entities, such as the spatial organization of different cell types in a tumor tissue. However, to further enhance our understanding of the underlying governing biological mechanisms, it is important to accurately capture the actual distributions of such complex data. Graph-based deep generative models are specifically tailored to accomplish that. In this work, we leverage state-of-the-art graph-based diffusion models to generate biologically meaningful cell-graphs. In particular, we show that the adopted graph diffusion model is able to accurately learn the distribution of cells in terms of their tertiary lymphoid structures (TLS) content, a well-established biomarker for evaluating the cancer progression in oncology research. Additionally, we further illustrate the utility of the learned generative models for data augmentation in a TLS classification task. To the best of our knowledge, this is the first work that leverages the power of graph diffusion models in generating meaningful biological cell structures.

Saba Nasiri, Selin Aviyente, Dorina Thanou

Learning representations from multiplex graphs, i.e., multi-layer networks where nodes interact through multiple relation types, is challenging due to the entanglement of shared (common) and layer-specific (private) information, which limits generalization and interpretability. In this work, we introduce a causal inference-based framework that disentangles common and private components in a self-supervised manner. CaDeM jointly (i) aligns shared embeddings across layers, (ii) enforces private embeddings to capture layer-specific signals, and (iii) applies backdoor adjustment to ensure that the common embeddings capture only global information while being separated from the private representations. Experiments on synthetic and real-world datasets demonstrate consistent improvements over existing baselines, highlighting the effectiveness of our approach for robust and interpretable multiplex graph representation learning.

Aref Einizade, Dorina Thanou, Fragkiskos D. Malliaros et al.

Simplicial complexes provide a powerful framework for modeling higher-order interactions in structured data, making them particularly suitable for applications such as trajectory prediction and mesh processing. However, existing simplicial neural networks (SNNs), whether convolutional or attention-based, rely primarily on discrete filtering techniques, which can be restrictive. In contrast, partial differential equations (PDEs) on simplicial complexes offer a principled approach to capture continuous dynamics in such structures. In this work, we introduce continuous simplicial neural network (COSIMO), a novel SNN architecture derived from PDEs on simplicial complexes. We provide theoretical and experimental justifications of COSIMO's stability under simplicial perturbations. Furthermore, we investigate the over-smoothing phenomenon, a common issue in geometric deep learning, demonstrating that COSIMO offers better control over this effect than discrete SNNs. Our experiments on real-world datasets demonstrate that COSIMO achieves competitive performance compared to state-of-the-art SNNs in complex and noisy environments. The implementation codes are available in https://github.com/ArefEinizade2/COSIMO.

Camille Challier, Xiaowu Sun, Thabo Mahendiran et al.

Accurate segmentation of coronary arteries remains a significant challenge in clinical practice, hindering the ability to effectively diagnose and manage coronary artery disease. The lack of large, annotated datasets for model training exacerbates this issue, limiting the development of automated tools that could assist radiologists. To address this, we introduce CM-UNet, which leverages self-supervised pre-training on unannotated datasets and transfer learning on limited annotated data, enabling accurate disease detection while minimizing the need for extensive manual annotations. Fine-tuning CM-UNet with only 18 annotated images instead of 500 resulted in a 15.2% decrease in Dice score, compared to a 46.5% drop in baseline models without pre-training. This demonstrates that self-supervised learning can enhance segmentation performance and reduce dependence on large datasets. This is one of the first studies to highlight the importance of self-supervised learning in improving coronary artery segmentation from X-ray angiography, with potential implications for advancing diagnostic accuracy in clinical practice. By enhancing segmentation accuracy in X-ray angiography images, the proposed approach aims to improve clinical workflows, reduce radiologists' workload, and accelerate disease detection, ultimately contributing to better patient outcomes. The source code is publicly available at https://github.com/CamilleChallier/Contrastive-Masked-UNet.

Boqi Chen, Cédric Vincent-Cuaz, Lydia A. Schoenpflug et al.

Vision foundation models (FMs) are accelerating the development of digital pathology algorithms and transforming biomedical research. These models learn, in a self-supervised manner, to represent histological features in highly heterogeneous tiles extracted from whole-slide images (WSIs) of real-world patient samples. The performance of these FMs is significantly influenced by the size, diversity, and balance of the pre-training data. However, data selection has been primarily guided by expert knowledge at the WSI level, focusing on factors such as disease classification and tissue types, while largely overlooking the granular details available at the tile level. In this paper, we investigate the potential of unsupervised automatic data curation at the tile-level, taking into account 350 million tiles. Specifically, we apply hierarchical clustering trees to pre-extracted tile embeddings, allowing us to sample balanced datasets uniformly across the embedding space of the pretrained FM. We further identify these datasets are subject to a trade-off between size and balance, potentially compromising the quality of representations learned by FMs, and propose tailored batch sampling strategies to mitigate this effect. We demonstrate the effectiveness of our method through improved performance on a diverse range of clinically relevant downstream tasks.

Xiaowu Sun, Thabo Mahendiran, Ortal Senouf et al.

Cardiovascular disease is the leading global cause of mortality, with coronary artery disease (CAD) as its most prevalent form, necessitating early risk prediction. While 3D coronary artery digital twins reconstructed from imaging offer detailed anatomy for personalized assessment, their analysis relies on computationally intensive computational fluid dynamics (CFD), limiting scalability. Data-driven approaches are hindered by scarce labeled data and lack of physiological priors. To address this, we present PINS-CAD, a physics-informed self-supervised learning framework. It pre-trains graph neural networks on 200,000 synthetic coronary digital twins to predict pressure and flow, guided by 1D Navier-Stokes equations and pressure-drop laws, eliminating the need for CFD or labeled data. When fine-tuned on clinical data from 635 patients in the multicenter FAME2 study, PINS-CAD predicts future cardiovascular events with an AUC of 0.73, outperforming clinical risk scores and data-driven baselines. This demonstrates that physics-informed pretraining boosts sample efficiency and yields physiologically meaningful representations. Furthermore, PINS-CAD generates spatially resolved pressure and fractional flow reserve curves, providing interpretable biomarkers. By embedding physical priors into geometric deep learning, PINS-CAD transforms routine angiography into a simulation-free, physiology-aware framework for scalable, preventive cardiology.

Alba Carballo-Castro, Manuel Madeira, Yiming Qin et al.

Directed graphs naturally model systems with asymmetric, ordered relationships, essential to applications in biology, transportation, social networks, and visual understanding. Generating such graphs enables tasks such as simulation, data augmentation and novel instance discovery; however, directed graph generation remains underexplored. We identify two key factors limiting progress in this direction: first, modeling edge directionality introduces a substantially larger dependency space, making the underlying distribution harder to learn; second, the absence of standardized benchmarks hinders rigorous evaluation. Addressing the former requires more expressive models that are sensitive to directional topologies. We propose Directo, the first generative model for directed graphs built upon the discrete flow matching framework. Our approach combines: (i) principled positional encodings tailored to asymmetric pairwise relations, (ii) a dual-attention mechanism capturing both incoming and outgoing dependencies, and (iii) a robust, discrete generative framework. To support evaluation, we introduce a benchmark suite covering synthetic and real-world datasets. It shows that our method performs strongly across diverse settings and even competes with specialized models for particular classes, such as directed acyclic graphs. Our results highlight the effectiveness and generality of our approach, establishing a solid foundation for future research in directed graph generation.

Manuel Madeira, Clement Vignac, Dorina Thanou et al.

Graph diffusion models have emerged as state-of-the-art techniques in graph generation; yet, integrating domain knowledge into these models remains challenging. Domain knowledge is particularly important in real-world scenarios, where invalid generated graphs hinder deployment in practical applications. Unconstrained and conditioned graph diffusion models fail to guarantee such domain-specific structural properties. We present ConStruct, a novel framework that enables graph diffusion models to incorporate hard constraints on specific properties, such as planarity or acyclicity. Our approach ensures that the sampled graphs remain within the domain of graphs that satisfy the specified property throughout the entire trajectory in both the forward and reverse processes. This is achieved by introducing an edge-absorbing noise model and a new projector operator. ConStruct demonstrates versatility across several structural and edge-deletion invariant constraints and achieves state-of-the-art performance for both synthetic benchmarks and attributed real-world datasets. For example, by incorporating planarity constraints in digital pathology graph datasets, the proposed method outperforms existing baselines, improving data validity by up to 71.1 percentage points.

Henry Kenlay, Dorina Thanou, Xiaowen Dong

Graph-structured data arise in a variety of real-world context ranging from sensor and transportation to biological and social networks. As a ubiquitous tool to process graph-structured data, spectral graph filters have been used to solve common tasks such as denoising and anomaly detection, as well as design deep learning architectures such as graph neural networks. Despite being an important tool, there is a lack of theoretical understanding of the stability properties of spectral graph filters, which are important for designing robust machine learning models. In this paper, we study filter stability and provide a novel and interpretable upper bound on the change of filter output, where the bound is expressed in terms of the endpoint degrees of the deleted and newly added edges, as well as the spatial proximity of those edges. This upper bound allows us to reason, in terms of structural properties of the graph, when a spectral graph filter will be stable. We further perform extensive experiments to verify intuition that can be gained from the bound.

Henry Kenlay, Dorina Thanou, Xiaowen Dong

Graph neural networks are experiencing a surge of popularity within the machine learning community due to their ability to adapt to non-Euclidean domains and instil inductive biases. Despite this, their stability, i.e., their robustness to small perturbations in the input, is not yet well understood. Although there exists some results showing the stability of graph neural networks, most take the form of an upper bound on the magnitude of change due to a perturbation in the graph topology. However, the change in the graph topology captured in existing bounds tend not to be expressed in terms of structural properties, limiting our understanding of the model robustness properties. In this work, we develop an interpretable upper bound elucidating that graph neural networks are stable to rewiring between high degree nodes. This bound and further research in bounds of similar type provide further understanding of the stability properties of graph neural networks.

Xiaowen Dong, Dorina Thanou, Laura Toni et al.

The effective representation, processing, analysis, and visualization of large-scale structured data, especially those related to complex domains such as networks and graphs, are one of the key questions in modern machine learning. Graph signal processing (GSP), a vibrant branch of signal processing models and algorithms that aims at handling data supported on graphs, opens new paths of research to address this challenge. In this article, we review a few important contributions made by GSP concepts and tools, such as graph filters and transforms, to the development of novel machine learning algorithms. In particular, our discussion focuses on the following three aspects: exploiting data structure and relational priors, improving data and computational efficiency, and enhancing model interpretability. Furthermore, we provide new perspectives on future development of GSP techniques that may serve as a bridge between applied mathematics and signal processing on one side, and machine learning and network science on the other. Cross-fertilization across these different disciplines may help unlock the numerous challenges of complex data analysis in the modern age.

Effrosyni Simou, Dorina Thanou, Pascal Frossard

In order to perform network analysis tasks, representations that capture the most relevant information in the graph structure are needed. However, existing methods do not learn representations that can be interpreted in a straightforward way and that are robust to perturbations to the graph structure. In this work, we address these two limitations by proposing node2coords, a representation learning algorithm for graphs, which learns simultaneously a low-dimensional space and coordinates for the nodes in that space. The patterns that span the low dimensional space reveal the graph's most important structural information. The coordinates of the nodes reveal the proximity of their local structure to the graph structural patterns. In order to measure this proximity by taking into account the underlying graph, we propose to use Wasserstein distances. We introduce an autoencoder that employs a linear layer in the encoder and a novel Wasserstein barycentric layer at the decoder. Node connectivity descriptors, that capture the local structure of the nodes, are passed through the encoder to learn the small set of graph structural patterns. In the decoder, the node connectivity descriptors are reconstructed as Wasserstein barycenters of the graph structural patterns. The optimal weights for the barycenter representation of a node's connectivity descriptor correspond to the coordinates of that node in the low-dimensional space. Experimental results demonstrate that the representations learned with node2coords are interpretable, lead to node embeddings that are stable to perturbations of the graph structure and achieve competitive or superior results compared to state-of-the-art methods in node classification.

Eda Bayram, Dorina Thanou, Elif Vural et al.

Structure inference is an important task for network data processing and analysis in data science. In recent years, quite a few approaches have been developed to learn the graph structure underlying a set of observations captured in a data space. Although real-world data is often acquired in settings where relationships are influenced by a priori known rules, such domain knowledge is still not well exploited in structure inference problems. In this paper, we identify the structure of signals defined in a data space whose inner relationships are encoded by multi-layer graphs. We aim at properly exploiting the information originating from each layer to infer the global structure underlying the signals. We thus present a novel method for combining the multiple graphs into a global graph using mask matrices, which are estimated through an optimization problem that accommodates the multi-layer graph information and a signal representation model. The proposed mask combination method also estimates the contribution of each graph layer in the structure of signals. The experiments conducted both on synthetic and real-world data suggest that integrating the multi-layer graph representation of the data in the structure inference framework enhances the learning procedure considerably by adapting to the quality and the quantity of the input data.

Sarah Itani, Dorina Thanou

While the prevalence of Autism Spectrum Disorder (ASD) is increasing, research continues in an effort to identify common etiological and pathophysiological bases. In this regard, modern machine learning and network science pave the way for a better understanding of the neuropathology and the development of diagnosis aid systems. The present work addresses the classification of neurotypical and ASD subjects by combining knowledge about both the structure and the functional activity of the brain. In particular, we model the brain structure as a graph, and the resting-state functional MRI (rs-fMRI) signals as values that live on the nodes of that graph. We then borrow tools from the emerging field of Graph Signal Processing (GSP) to build features related to the frequency content of these signals. In order to make these features highly discriminative, we apply an extension of the Fukunaga-Koontz transform. Finally, we use these new markers to train a decision tree, an interpretable classification scheme, which results in a final diagnosis aid model. Interestingly, the resulting decision tree outperforms state-of-the-art methods on the publicly available Autism Brain Imaging Data Exchange (ABIDE) collection. Moreover, the analysis of the predictive markers reveals the influence of the frontal and temporal lobes in the diagnosis of the disorder, which is in line with previous findings in the literature of neuroscience. Our results indicate that exploiting jointly structural and functional information of the brain can reveal important information about the complexity of the neuropathology.

Xiaowen Dong, Dorina Thanou, Michael Rabbat et al.

The construction of a meaningful graph topology plays a crucial role in the effective representation, processing, analysis and visualization of structured data. When a natural choice of the graph is not readily available from the data sets, it is thus desirable to infer or learn a graph topology from the data. In this tutorial overview, we survey solutions to the problem of graph learning, including classical viewpoints from statistics and physics, and more recent approaches that adopt a graph signal processing (GSP) perspective. We further emphasize the conceptual similarities and differences between classical and GSP-based graph inference methods, and highlight the potential advantage of the latter in a number of theoretical and practical scenarios. We conclude with several open issues and challenges that are keys to the design of future signal processing and machine learning algorithms for learning graphs from data.

Hermina Petric Maretic, Dorina Thanou, Pascal Frossard

Graph signals offer a very generic and natural representation for data that lives on networks or irregular structures. The actual data structure is however often unknown a priori but can sometimes be estimated from the knowledge of the application domain. If this is not possible, the data structure has to be inferred from the mere signal observations. This is exactly the problem that we address in this paper, under the assumption that the graph signals can be represented as a sparse linear combination of a few atoms of a structured graph dictionary. The dictionary is constructed on polynomials of the graph Laplacian, which can sparsely represent a general class of graph signals composed of localized patterns on the graph. We formulate a graph learning problem, whose solution provides an ideal fit between the signal observations and the sparse graph signal model. As the problem is non-convex, we propose to solve it by alternating between a signal sparse coding and a graph update step. We provide experimental results that outline the good graph recovery performance of our method, which generally compares favourably to other recent network inference algorithms.

Dorina Thanou, Xiaowen Dong, Daniel Kressner et al.

Effective information analysis generally boils down to properly identifying the structure or geometry of the data, which is often represented by a graph. In some applications, this structure may be partly determined by design constraints or pre-determined sensing arrangements, like in road transportation networks for example. In general though, the data structure is not readily available and becomes pretty difficult to define. In particular, the global smoothness assumptions, that most of the existing works adopt, are often too general and unable to properly capture localized properties of data. In this paper, we go beyond this classical data model and rather propose to represent information as a sparse combination of localized functions that live on a data structure represented by a graph. Based on this model, we focus on the problem of inferring the connectivity that best explains the data samples at different vertices of a graph that is a priori unknown. We concentrate on the case where the observed data is actually the sum of heat diffusion processes, which is a quite common model for data on networks or other irregular structures. We cast a new graph learning problem and solve it with an efficient nonconvex optimization algorithm. Experiments on both synthetic and real world data finally illustrate the benefits of the proposed graph learning framework and confirm that the data structure can be efficiently learned from data observations only. We believe that our algorithm will help solving key questions in diverse application domains such as social and biological network analysis where it is crucial to unveil proper geometry for data understanding and inference.

Dorina Thanou, Philip A. Chou, Pascal Frossard

This paper addresses the problem of compression of 3D point cloud sequences that are characterized by moving 3D positions and color attributes. As temporally successive point cloud frames are similar, motion estimation is key to effective compression of these sequences. It however remains a challenging problem as the point cloud frames have varying numbers of points without explicit correspondence information. We represent the time-varying geometry of these sequences with a set of graphs, and consider 3D positions and color attributes of the points clouds as signals on the vertices of the graphs. We then cast motion estimation as a feature matching problem between successive graphs. The motion is estimated on a sparse set of representative vertices using new spectral graph wavelet descriptors. A dense motion field is eventually interpolated by solving a graph-based regularization problem. The estimated motion is finally used for removing the temporal redundancy in the predictive coding of the 3D positions and the color characteristics of the point cloud sequences. Experimental results demonstrate that our method is able to accurately estimate the motion between consecutive frames. Moreover, motion estimation is shown to bring significant improvement in terms of the overall compression performance of the sequence. To the best of our knowledge, this is the first paper that exploits both the spatial correlation inside each frame (through the graph) and the temporal correlation between the frames (through the motion estimation) to compress the color and the geometry of 3D point cloud sequences in an efficient way.

Xiaowen Dong, Dorina Thanou, Pascal Frossard et al.

The construction of a meaningful graph plays a crucial role in the success of many graph-based representations and algorithms for handling structured data, especially in the emerging field of graph signal processing. However, a meaningful graph is not always readily available from the data, nor easy to define depending on the application domain. In particular, it is often desirable in graph signal processing applications that a graph is chosen such that the data admit certain regularity or smoothness on the graph. In this paper, we address the problem of learning graph Laplacians, which is equivalent to learning graph topologies, such that the input data form graph signals with smooth variations on the resulting topology. To this end, we adopt a factor analysis model for the graph signals and impose a Gaussian probabilistic prior on the latent variables that control these signals. We show that the Gaussian prior leads to an efficient representation that favors the smoothness property of the graph signals. We then propose an algorithm for learning graphs that enforces such property and is based on minimizing the variations of the signals on the learned graph. Experiments on both synthetic and real world data demonstrate that the proposed graph learning framework can efficiently infer meaningful graph topologies from signal observations under the smoothness prior.

Dorina Thanou, David I Shuman, Pascal Frossard

In sparse signal representation, the choice of a dictionary often involves a tradeoff between two desirable properties -- the ability to adapt to specific signal data and a fast implementation of the dictionary. To sparsely represent signals residing on weighted graphs, an additional design challenge is to incorporate the intrinsic geometric structure of the irregular data domain into the atoms of the dictionary. In this work, we propose a parametric dictionary learning algorithm to design data-adapted, structured dictionaries that sparsely represent graph signals. In particular, we model graph signals as combinations of overlapping local patterns. We impose the constraint that each dictionary is a concatenation of subdictionaries, with each subdictionary being a polynomial of the graph Laplacian matrix, representing a single pattern translated to different areas of the graph. The learning algorithm adapts the patterns to a training set of graph signals. Experimental results on both synthetic and real datasets demonstrate that the dictionaries learned by the proposed algorithm are competitive with and often better than unstructured dictionaries learned by state-of-the-art numerical learning algorithms in terms of sparse approximation of graph signals. In contrast to the unstructured dictionaries, however, the dictionaries learned by the proposed algorithm feature localized atoms and can be implemented in a computationally efficient manner in signal processing tasks such as compression, denoising, and classification.