Marco Lorenzi

Jean Ogier du Terrail, Samy-Safwan Ayed, Edwige Cyffers et al. · eth-zurich

Federated Learning (FL) is a novel approach enabling several clients holding sensitive data to collaboratively train machine learning models, without centralizing data. The cross-silo FL setting corresponds to the case of few ($2$--$50$) reliable clients, each holding medium to large datasets, and is typically found in applications such as healthcare, finance, or industry. While previous works have proposed representative datasets for cross-device FL, few realistic healthcare cross-silo FL datasets exist, thereby slowing algorithmic research in this critical application. In this work, we propose a novel cross-silo dataset suite focused on healthcare, FLamby (Federated Learning AMple Benchmark of Your cross-silo strategies), to bridge the gap between theory and practice of cross-silo FL. FLamby encompasses 7 healthcare datasets with natural splits, covering multiple tasks, modalities, and data volumes, each accompanied with baseline training code. As an illustration, we additionally benchmark standard FL algorithms on all datasets. Our flexible and modular suite allows researchers to easily download datasets, reproduce results and re-use the different components for their research. FLamby is available at~\url{www.github.com/owkin/flamby}.

Hasan Kassem, Orion Banks, Omar Benjelloun et al.

Decentralized AI systems, such as federated learning, can play a critical role in further unlocking AI asset marketplaces (e.g., healthcare data marketplaces) thanks to increased asset privacy protection. Unlocking this big potential necessitates governance mechanisms that are transparent, scalable, and verifiable. However current governance approaches rely on bespoke, infrastructure-specific policies that hinder asset interoperability and trust among systems. We are proposing a Technical Policy Blueprint that encodes governance requirements as policy-as-code objects and separates asset policy verification from asset policy enforcement. In this architecture the Policy Engine verifies evidence (e.g., identities, signatures, payments, trusted-hardware attestations) and issues capability packages. Asset Guardians (e.g. data guardians, model guardians, computation guardians, etc.) enforce access or execution solely based on these capability packages. This core concept of decoupling policy processing from capabilities enables governance to evolve without reconfiguring AI infrastructure, thus creating an approach that is transparent, auditable, and resilient to change.

Lucia Innocenti, Michela Antonelli, Francesco Cremonesi et al.

Healthcare data is often split into medium/small-sized collections across multiple hospitals and access to it is encumbered by privacy regulations. This brings difficulties to use them for the development of machine learning and deep learning models, which are known to be data-hungry. One way to overcome this limitation is to use collaborative learning (CL) methods, which allow hospitals to work collaboratively to solve a task, without the need to explicitly share local data. In this paper, we address a prostate segmentation problem from MRI in a collaborative scenario by comparing two different approaches: federated learning (FL) and consensus-based methods (CBM). To the best of our knowledge, this is the first work in which CBM, such as label fusion techniques, are used to solve a problem of collaborative learning. In this setting, CBM combine predictions from locally trained models to obtain a federated strong learner with ideally improved robustness and predictive variance properties. Our experiments show that, in the considered practical scenario, CBMs provide equal or better results than FL, while being highly cost-effective. Our results demonstrate that the consensus paradigm may represent a valid alternative to FL for typical training tasks in medical imaging.

Irene Balelli, Santiago Silva, Marco Lorenzi

We propose a novel federated learning paradigm to model data variability among heterogeneous clients in multi-centric studies. Our method is expressed through a hierarchical Bayesian latent variable model, where client-specific parameters are assumed to be realization from a global distribution at the master level, which is in turn estimated to account for data bias and variability across clients. We show that our framework can be effectively optimized through expectation maximization (EM) over latent master's distribution and clients' parameters. We also introduce formal differential privacy (DP) guarantees compatibly with our EM optimization scheme. We tested our method on the analysis of multi-modal medical imaging data and clinical scores from distributed clinical datasets of patients affected by Alzheimer's disease. We demonstrate that our method is robust when data is distributed either in iid and non-iid manners, even when local parameters perturbation is included to provide DP guarantees. Moreover, the variability of data, views and centers can be quantified in an interpretable manner, while guaranteeing high-quality data reconstruction as compared to state-of-the-art autoencoding models and federated learning schemes. The code is available at https://gitlab.inria.fr/epione/federated-multi-views-ppca.

Riccardo Taiello, Sergen Cansiz, Marc Vesin et al.

Deploying federated learning (FL) in real-world scenarios, particularly in healthcare, poses challenges in communication and security. In particular, with respect to the federated aggregation procedure, researchers have been focusing on the study of secure aggregation (SA) schemes to provide privacy guarantees over the model's parameters transmitted by the clients. Nevertheless, the practical availability of SA in currently available FL frameworks is currently limited, due to computational and communication bottlenecks. To fill this gap, this study explores the implementation of SA within the open-source Fed-BioMed framework. We implement and compare two SA protocols, Joye-Libert (JL) and Low Overhead Masking (LOM), by providing extensive benchmarks in a panel of healthcare data analysis problems. Our theoretical and experimental evaluations on four datasets demonstrate that SA protocols effectively protect privacy while maintaining task accuracy. Computational overhead during training is less than 1% on a CPU and less than 50% on a GPU for large models, with protection phases taking less than 10 seconds. Incorporating SA into Fed-BioMed impacts task accuracy by no more than 2% compared to non-SA scenarios. Overall this study demonstrates the feasibility of SA in real-world healthcare applications and contributes in reducing the gap towards the adoption of privacy-preserving technologies in sensitive applications.

Francesco Cremonesi, Marc Vesin, Sergen Cansiz et al.

The real-world implementation of federated learning is complex and requires research and development actions at the crossroad between different domains ranging from data science, to software programming, networking, and security. While today several FL libraries are proposed to data scientists and users, most of these frameworks are not designed to find seamless application in medical use-cases, due to the specific challenges and requirements of working with medical data and hospital infrastructures. Moreover, governance, design principles, and security assumptions of these frameworks are generally not clearly illustrated, thus preventing the adoption in sensitive applications. Motivated by the current technological landscape of FL in healthcare, in this document we present Fed-BioMed: a research and development initiative aiming at translating federated learning (FL) into real-world medical research applications. We describe our design space, targeted users, domain constraints, and how these factors affect our current and future software architecture.

Andre Altmann, Ana C Lawry Aguila, Neda Jahanshad et al.

A major challenge in imaging genetics and similar fields is to link high-dimensional data in one domain, e.g., genetic data, to high dimensional data in a second domain, e.g., brain imaging data. The standard approach in the area are mass univariate analyses across genetic factors and imaging phenotypes. That entails executing one genome-wide association study (GWAS) for each pre-defined imaging measure. Although this approach has been tremendously successful, one shortcoming is that phenotypes must be pre-defined. Consequently, effects that are not confined to pre-selected regions of interest or that reflect larger brain-wide patterns can easily be missed. In this work we introduce a Partial Least Squares (PLS)-based framework, which we term Cluster-Bootstrap PLS (CLUB-PLS), that can work with large input dimensions in both domains as well as with large sample sizes. One key factor of the framework is to use cluster bootstrap to provide robust statistics for single input features in both domains. We applied CLUB-PLS to investigating the genetic basis of surface area and cortical thickness in a sample of 33,000 subjects from the UK Biobank. We found 107 genome-wide significant locus-phenotype pairs that are linked to 386 different genes. We found that a vast majority of these loci could be technically validated at a high rate: using classic GWAS or Genome-Wide Inferred Statistics (GWIS) we found that 85 locus-phenotype pairs exceeded the genome-wide suggestive (P<1e-05) threshold.

Yann Fraboni, Richard Vidal, Laetitia Kameni et al.

We propose a novel framework to study asynchronous federated learning optimization with delays in gradient updates. Our theoretical framework extends the standard FedAvg aggregation scheme by introducing stochastic aggregation weights to represent the variability of the clients update time, due for example to heterogeneous hardware capabilities. Our formalism applies to the general federated setting where clients have heterogeneous datasets and perform at least one step of stochastic gradient descent (SGD). We demonstrate convergence for such a scheme and provide sufficient conditions for the related minimum to be the optimum of the federated problem. We show that our general framework applies to existing optimization schemes including centralized learning, FedAvg, asynchronous FedAvg, and FedBuff. The theory here provided allows drawing meaningful guidelines for designing a federated learning experiment in heterogeneous conditions. In particular, we develop in this work FedFix, a novel extension of FedAvg enabling efficient asynchronous federated training while preserving the convergence stability of synchronous aggregation. We empirically demonstrate our theory on a series of experiments showing that asynchronous FedAvg leads to fast convergence at the expense of stability, and we finally demonstrate the improvements of FedFix over synchronous and asynchronous FedAvg.

Yann Fraboni, Martin Van Waerebeke, Kevin Scaman et al.

Machine Unlearning (MU) is an increasingly important topic in machine learning safety, aiming at removing the contribution of a given data point from a training procedure. Federated Unlearning (FU) consists in extending MU to unlearn a given client's contribution from a federated training routine. While several FU methods have been proposed, we currently lack a general approach providing formal unlearning guarantees to the FedAvg routine, while ensuring scalability and generalization beyond the convex assumption on the clients' loss functions. We aim at filling this gap by proposing SIFU (Sequential Informed Federated Unlearning), a new FU method applying to both convex and non-convex optimization regimes. SIFU naturally applies to FedAvg without additional computational cost for the clients and provides formal guarantees on the quality of the unlearning task. We provide a theoretical analysis of the unlearning properties of SIFU, and practically demonstrate its effectiveness as compared to a panel of unlearning methods from the state-of-the-art.

Etrit Haxholli, Marco Lorenzi

In Diffusion Probabilistic Models (DPMs), the task of modeling the score evolution via a single time-dependent neural network necessitates extended training periods and may potentially impede modeling flexibility and capacity. To counteract these challenges, we propose leveraging the independence of learning tasks at different time points inherent to DPMs. More specifically, we partition the learning task by utilizing independent networks, each dedicated to learning the evolution of scores within a specific time sub-interval. Further, inspired by residual flows, we extend this strategy to its logical conclusion by employing separate networks to independently model the score at each individual time point. As empirically demonstrated on synthetic and image datasets, our approach not only significantly accelerates the training process by introducing an additional layer of parallelization atop data parallelization, but it also enhances density estimation performance when compared to the conventional training methodology for DPMs.

Irene Balelli, Aude Sportisse, Francesco Cremonesi et al.

Federated learning allows for the training of machine learning models on multiple decentralized local datasets without requiring explicit data exchange. However, data pre-processing, including strategies for handling missing data, remains a major bottleneck in real-world federated learning deployment, and is typically performed locally. This approach may be biased, since the subpopulations locally observed at each center may not be representative of the overall one. To address this issue, this paper first proposes a more consistent approach to data standardization through a federated model. Additionally, we propose Fed-MIWAE, a federated version of the state-of-the-art imputation method MIWAE, a deep latent variable model for missing data imputation based on variational autoencoders. MIWAE has the great advantage of being easily trainable with classical federated aggregators. Furthermore, it is able to deal with MAR (Missing At Random) data, a more challenging missing-data mechanism than MCAR (Missing Completely At Random), where the missingness of a variable can depend on the observed ones. We evaluate our method on multi-modal medical imaging data and clinical scores from a simulated federated scenario with the ADNI dataset. We compare Fed-MIWAE with respect to classical imputation methods, either performed locally or in a centralized fashion. Fed-MIWAE allows to achieve imputation accuracy comparable with the best centralized method, even when local data distributions are highly heterogeneous. In addition, thanks to the variational nature of Fed-MIWAE, our method is designed to perform multiple imputation, allowing for the quantification of the imputation uncertainty in the federated scenario.

Etrit Haxholli, Marco Lorenzi

While the neural ODE formulation of normalizing flows such as in FFJORD enables us to calculate the determinants of free form Jacobians in O(D) time, the flexibility of the transformation underlying neural ODEs has been shown to be suboptimal. In this paper, we present AFFJORD, a neural ODE-based normalizing flow which enhances the representation power of FFJORD by defining the neural ODE through special augmented transformation dynamics which preserve the topology of the space. Furthermore, we derive the Jacobian determinant of the general augmented form by generalizing the chain rule in the continuous sense into the Cable Rule, which expresses the forward sensitivity of ODEs with respect to their initial conditions. The cable rule gives an explicit expression for the Jacobian of a neural ODE transformation, and provides an elegant proof of the instantaneous change of variable. Our experimental results on density estimation in synthetic and high dimensional data, such as MNIST, CIFAR-10 and CelebA 32x32, show that AFFJORD outperforms the baseline FFJORD through the improved flexibility of the underlying vector field.

Martin Van Waerebeke, Marco Lorenzi, Kevin Scaman et al.

In machine unlearning, $(\varepsilon,δ)-$unlearning is a popular framework that provides formal guarantees on the effectiveness of the removal of a subset of training data, the forget set, from a trained model. For strongly convex objectives, existing first-order methods achieve $(\varepsilon,δ)-$unlearning, but they only use the forget set to calibrate injected noise, never as a direct optimization signal. In contrast, efficient empirical heuristics often exploit the forget samples (e.g., via gradient ascent) but come with no formal unlearning guarantees. We bridge this gap by presenting the Variance-Reduced Unlearning (VRU) algorithm. To the best of our knowledge, VRU is the first first-order algorithm that directly includes forget set gradients in its update rule, while provably satisfying ($(\varepsilon,δ)-$unlearning. We establish the convergence of VRU and show that incorporating the forget set yields strictly improved rates, i.e. a better dependence on the achieved error compared to existing first-order $(\varepsilon,δ)-$unlearning methods. Moreover, we prove that, in a low-error regime, VRU asymptotically outperforms any first-order method that ignores the forget set.Experiments corroborate our theory, showing consistent gains over both state-of-the-art certified unlearning methods and over empirical baselines that explicitly leverage the forget set.

Etrit Haxholli, Marco Lorenzi

The study of loss function distributions is critical to characterize a model's behaviour on a given machine learning problem. For example, while the quality of a model is commonly determined by the average loss assessed on a testing set, this quantity does not reflect the existence of the true mean of the loss distribution. Indeed, the finiteness of the statistical moments of the loss distribution is related to the thickness of its tails, which are generally unknown. Since typical cross-validation schemes determine a family of testing loss distributions conditioned on the training samples, the total loss distribution must be recovered by marginalizing over the space of training sets. As we show in this work, the finiteness of the sampling procedure negatively affects the reliability and efficiency of classical tail estimation methods from the Extreme Value Theory, such as the Peaks-Over-Threshold approach. In this work we tackle this issue by developing a novel general theory for estimating the tails of marginal distributions, when there exists a large variability between locations of the individual conditional distributions underlying the marginal. To this end, we demonstrate that under some regularity conditions, the shape parameter of the marginal distribution is the maximum tail shape parameter of the family of conditional distributions. We term this estimation approach as Cross Tail Estimation (CTE). We test cross-tail estimation in a series of experiments on simulated and real data, showing the improved robustness and quality of tail estimation as compared to classical approaches, and providing evidence for the relationship between overfitting and loss distribution tail thickness.

Adrià Casamitjana, Marco Lorenzi, Sebastiano Ferraris et al.

Joint registration of a stack of 2D histological sections to recover 3D structure (``3D histology reconstruction'') finds application in areas such as atlas building and validation of \emph{in vivo} imaging. Straightforward pairwise registration of neighbouring sections yields smooth reconstructions but has well-known problems such as ``banana effect'' (straightening of curved structures) and ``z-shift'' (drift). While these problems can be alleviated with an external, linearly aligned reference (e.g., Magnetic Resonance (MR) images), registration is often inaccurate due to contrast differences and the strong nonlinear distortion of the tissue, including artefacts such as folds and tears. In this paper, we present a probabilistic model of spatial deformation that yields reconstructions for multiple histological stains that that are jointly smooth, robust to outliers, and follow the reference shape. The model relies on a spanning tree of latent transforms connecting all the sections and slices of the reference volume, and assumes that the registration between any pair of images can be see as a noisy version of the composition of (possibly inverted) latent transforms connecting the two images. Bayesian inference is used to compute the most likely latent transforms given a set of pairwise registrations between image pairs within and across modalities. The framework is used for accurate 3D reconstruction of two stains (Nissl and parvalbumin) from the Allen human brain atlas, showing its benefits on real data with severe distortions. Moreover, we also provide the registration of the reconstructed volume to MNI space, bridging the gaps between two of the most widely used atlases in histology and MRI. The 3D reconstructed volumes and atlas registration can be downloaded from https://openneuro.org/datasets/ds003590. The code is freely available at https://github.com/acasamitjana/3dhirest.

Razvan V. Marinescu, Arman Eshaghi, Marco Lorenzi et al.

Here we present DIVE: Data-driven Inference of Vertexwise Evolution. DIVE is an image-based disease progression model with single-vertex resolution, designed to reconstruct long-term patterns of brain pathology from short-term longitudinal data sets. DIVE clusters vertex-wise biomarker measurements on the cortical surface that have similar temporal dynamics across a patient population, and concurrently estimates an average trajectory of vertex measurements in each cluster. DIVE uniquely outputs a parcellation of the cortex into areas with common progression patterns, leading to a new signature for individual diseases. DIVE further estimates the disease stage and progression speed for every visit of every subject, potentially enhancing stratification for clinical trials or management. On simulated data, DIVE can recover ground truth clusters and their underlying trajectory, provided the average trajectories are sufficiently different between clusters. We demonstrate DIVE on data from two cohorts: the Alzheimer's Disease Neuroimaging Initiative (ADNI) and the Dementia Research Centre (DRC), UK, containing patients with Posterior Cortical Atrophy (PCA) as well as typical Alzheimer's disease (tAD). DIVE finds similar spatial patterns of atrophy for tAD subjects in the two independent datasets (ADNI and DRC), and further reveals distinct patterns of pathology in different diseases (tAD vs PCA) and for distinct types of biomarker data: cortical thickness from Magnetic Resonance Imaging (MRI) vs amyloid load from Positron Emission Tomography (PET). Finally, DIVE can be used to estimate a fine-grained spatial distribution of pathology in the brain using any kind of voxelwise or vertexwise measures including Jacobian compression maps, fractional anisotropy (FA) maps from diffusion imaging or other PET measures. DIVE source code is available online: https://github.com/mrazvan22/dive

Razvan V. Marinescu, Marco Lorenzi, Stefano B. Blumberg et al.

We introduce Disease Knowledge Transfer (DKT), a novel technique for transferring biomarker information between related neurodegenerative diseases. DKT infers robust multimodal biomarker trajectories in rare neurodegenerative diseases even when only limited, unimodal data is available, by transferring information from larger multimodal datasets from common neurodegenerative diseases. DKT is a joint-disease generative model of biomarker progressions, which exploits biomarker relationships that are shared across diseases. Our proposed method allows, for the first time, the estimation of plausible, multimodal biomarker trajectories in Posterior Cortical Atrophy (PCA), a rare neurodegenerative disease where only unimodal MRI data is available. For this we train DKT on a combined dataset containing subjects with two distinct diseases and sizes of data available: 1) a larger, multimodal typical AD (tAD) dataset from the TADPOLE Challenge, and 2) a smaller unimodal Posterior Cortical Atrophy (PCA) dataset from the Dementia Research Centre (DRC), for which only a limited number of Magnetic Resonance Imaging (MRI) scans are available. Although validation is challenging due to lack of data in PCA, we validate DKT on synthetic data and two patient datasets (TADPOLE and PCA cohorts), showing it can estimate the ground truth parameters in the simulation and predict unseen biomarkers on the two patient datasets. While we demonstrated DKT on Alzheimer's variants, we note DKT is generalisable to other forms of related neurodegenerative diseases. Source code for DKT is available online: https://github.com/mrazvan22/dkt.

Giuseppe A. Orlando, Paolo Papotti, Maria A. Zuluaga et al.

Vision-language models (VLMs) have shown potential for automated radiology report generation, yet existing approaches rely on global embedding compression of volumetric data, often leading to hallucinated findings and limited anatomical grounding in 3D CT imaging. We introduce MedScribe, a hypothesis-driven framework that reformulates report generation as an iterative evidence acquisition process rather than a single-pass encoding task. MedScribe models reporting as a sequential decision process in which a large language model dynamically invokes pathology-specific diagnostic tools to extract localized volumetric features. These structured features are used to query a multidimensional retrieval space aligned with pathology-specific textual evidence. By explicitly accumulating quantitative evidence prior to synthesis, the framework enforces fine-grained grounding and reduces unsupported claims. Without task-specific fine-tuning, MedScribe improves clinical accuracy, factual consistency, and interpretability on CT-RATE and RadChestCT compared to state-of-the-art 2D and 3D VLMs, demonstrating the value of hypothesis-driven reasoning for reliable medical image reporting.

Martin Van Waerebeke, Marco Lorenzi, Giovanni Neglia et al.

Machine Unlearning (MU) aims at removing the influence of specific data points from a trained model, striving to achieve this at a fraction of the cost of full model retraining. In this paper, we analyze the efficiency of unlearning methods and establish the first upper and lower bounds on minimax computation times for this problem, characterizing the performance of the most efficient algorithm against the most difficult objective function. Specifically, for strongly convex objective functions and under the assumption that the forget data is inaccessible to the unlearning method, we provide a phase diagram for the unlearning complexity ratio -- a novel metric that compares the computational cost of the best unlearning method to full model retraining. The phase diagram reveals three distinct regimes: one where unlearning at a reduced cost is infeasible, another where unlearning is trivial because adding noise suffices, and a third where unlearning achieves significant computational advantages over retraining. These findings highlight the critical role of factors such as data dimensionality, the number of samples to forget, and privacy constraints in determining the practical feasibility of unlearning.

Francesco Cremonesi, Lucia Innocenti, Sebastien Ourselin et al.

Background. Federated learning (FL) has gained wide popularity as a collaborative learning paradigm enabling collaborative AI in sensitive healthcare applications. Nevertheless, the practical implementation of FL presents technical and organizational challenges, as it generally requires complex communication infrastructures. In this context, consensus-based learning (CBL) may represent a promising collaborative learning alternative, thanks to the ability of combining local knowledge into a federated decision system, while potentially reducing deployment overhead. Methods. In this work we propose an extensive benchmark of the accuracy and cost-effectiveness of a panel of FL and CBL methods in a wide range of collaborative medical data analysis scenarios. The benchmark includes 7 different medical datasets, encompassing 3 machine learning tasks, 8 different data modalities, and multi-centric settings involving 3 to 23 clients. Findings. Our results reveal that CBL is a cost-effective alternative to FL. When compared across the panel of medical dataset in the considered benchmark, CBL methods provide equivalent accuracy to the one achieved by FL.Nonetheless, CBL significantly reduces training time and communication cost (resp. 15 fold and 60 fold decrease) (p < 0.05). Interpretation. This study opens a novel perspective on the deployment of collaborative AI in real-world applications, whereas the adoption of cost-effective methods is instrumental to achieve sustainability and democratisation of AI by alleviating the need for extensive computational resources.

Riccardo Taiello, Melek Önen, Francesco Capano et al.

Image registration is a key task in medical imaging applications, allowing to represent medical images in a common spatial reference frame. Current approaches to image registration are generally based on the assumption that the content of the images is usually accessible in clear form, from which the spatial transformation is subsequently estimated. This common assumption may not be met in practical applications, since the sensitive nature of medical images may ultimately require their analysis under privacy constraints, preventing to openly share the image content.In this work, we formulate the problem of image registration under a privacy preserving regime, where images are assumed to be confidential and cannot be disclosed in clear. We derive our privacy preserving image registration framework by extending classical registration paradigms to account for advanced cryptographic tools, such as secure multi-party computation and homomorphic encryption, that enable the execution of operations without leaking the underlying data. To overcome the problem of performance and scalability of cryptographic tools in high dimensions, we propose several techniques to optimize the image registration operations by using gradient approximations, and by revisiting the use of homomorphic encryption trough packing, to allow the efficient encryption and multiplication of large matrices. We demonstrate our privacy preserving framework in linear and non-linear registration problems, evaluating its accuracy and scalability with respect to standard, non-private counterparts. Our results show that privacy preserving image registration is feasible and can be adopted in sensitive medical imaging applications.

Yann Fraboni, Richard Vidal, Laetitia Kameni et al.

While client sampling is a central operation of current state-of-the-art federated learning (FL) approaches, the impact of this procedure on the convergence and speed of FL remains under-investigated. In this work, we provide a general theoretical framework to quantify the impact of a client sampling scheme and of the clients heterogeneity on the federated optimization. First, we provide a unified theoretical ground for previously reported sampling schemes experimental results on the relationship between FL convergence and the variance of the aggregation weights. Second, we prove for the first time that the quality of FL convergence is also impacted by the resulting covariance between aggregation weights. Our theory is general, and is here applied to Multinomial Distribution (MD) and Uniform sampling, two default unbiased client sampling schemes of FL, and demonstrated through a series of experiments in non-iid and unbalanced scenarios. Our results suggest that MD sampling should be used as default sampling scheme, due to the resilience to the changes in data ratio during the learning process, while Uniform sampling is superior only in the special case when clients have the same amount of data.

Yann Fraboni, Richard Vidal, Laetitia Kameni et al.

This work addresses the problem of optimizing communications between server and clients in federated learning (FL). Current sampling approaches in FL are either biased, or non optimal in terms of server-clients communications and training stability. To overcome this issue, we introduce \textit{clustered sampling} for clients selection. We prove that clustered sampling leads to better clients representatitivity and to reduced variance of the clients stochastic aggregation weights in FL. Compatibly with our theory, we provide two different clustering approaches enabling clients aggregation based on 1) sample size, and 2) models similarity. Through a series of experiments in non-iid and unbalanced scenarios, we demonstrate that model aggregation through clustered sampling consistently leads to better training convergence and variability when compared to standard sampling approaches. Our approach does not require any additional operation on the clients side, and can be seamlessly integrated in standard FL implementations. Finally, clustered sampling is compatible with existing methods and technologies for privacy enhancement, and for communication reduction through model compression.

Vien Ngoc Dang, Francesco Galati, Rosa Cortese et al.

Deep learning techniques for 3D brain vessel image segmentation have not been as successful as in the segmentation of other organs and tissues. This can be explained by two factors. First, deep learning techniques tend to show poor performances at the segmentation of relatively small objects compared to the size of the full image. Second, due to the complexity of vascular trees and the small size of vessels, it is challenging to obtain the amount of annotated training data typically needed by deep learning methods. To address these problems, we propose a novel annotation-efficient deep learning vessel segmentation framework. The framework avoids pixel-wise annotations, only requiring weak patch-level labels to discriminate between vessel and non-vessel 2D patches in the training set, in a setup similar to the CAPTCHAs used to differentiate humans from bots in web applications. The user-provided weak annotations are used for two tasks: 1) to synthesize pixel-wise pseudo-labels for vessels and background in each patch, which are used to train a segmentation network, and 2) to train a classifier network. The classifier network allows to generate additional weak patch labels, further reducing the annotation burden, and it acts as a noise filter for poor quality images. We use this framework for the segmentation of the cerebrovascular tree in Time-of-Flight angiography (TOF) and Susceptibility-Weighted Images (SWI). The results show that the framework achieves state-of-the-art accuracy, while reducing the annotation time by ~77% w.r.t. learning-based segmentation methods using pixel-wise labels for training.

Jaume Banus, Maxime Sermesant, Oscar Camara et al.

The use of mechanistic models in clinical studies is limited by the lack of multi-modal patients data representing different anatomical and physiological processes. For example, neuroimaging datasets do not provide a sufficient representation of heart features for the modeling of cardiovascular factors in brain disorders. To tackle this problem we introduce a probabilistic framework for joint cardiac data imputation and personalisation of cardiovascular mechanistic models, with application to brain studies with incomplete heart data. Our approach is based on a variational framework for the joint inference of an imputation model of cardiac information from the available features, along with a Gaussian Process emulator that can faithfully reproduce personalised cardiovascular dynamics. Experimental results on UK Biobank show that our model allows accurate imputation of missing cardiac features in datasets containing minimal heart information, e.g. systolic and diastolic blood pressures only, while jointly estimating the emulated parameters of the lumped model. This allows a novel exploration of the heart-brain joint relationship through simulation of realistic cardiac dynamics corresponding to different conditions of brain anatomy.

Yann Fraboni, Richard Vidal, Marco Lorenzi

Free-rider attacks against federated learning consist in dissimulating participation to the federated learning process with the goal of obtaining the final aggregated model without actually contributing with any data. This kind of attacks is critical in sensitive applications of federated learning, where data is scarce and the model has high commercial value. We introduce here the first theoretical and experimental analysis of free-rider attacks on federated learning schemes based on iterative parameters aggregation, such as FedAvg or FedProx, and provide formal guarantees for these attacks to converge to the aggregated models of the fair participants. We first show that a straightforward implementation of this attack can be simply achieved by not updating the local parameters during the iterative federated optimization. As this attack can be detected by adopting simple countermeasures at the server level, we subsequently study more complex disguising schemes based on stochastic updates of the free-rider parameters. We demonstrate the proposed strategies on a number of experimental scenarios, in both iid and non-iid settings. We conclude by providing recommendations to avoid free-rider attacks in real world applications of federated learning, especially in sensitive domains where security of data and models is critical.

Raphaël Sivera, Hervé Delingette, Marco Lorenzi et al.

In this study we propose a deformation-based framework to jointly model the influence of aging and Alzheimer's disease (AD) on the brain morphological evolution. Our approach combines a spatio-temporal description of both processes into a generative model. A reference morphology is deformed along specific trajectories to match subject specific morphologies. It is used to define two imaging progression markers: 1) a morphological age and 2) a disease score. These markers can be computed locally in any brain region. The approach is evaluated on brain structural magnetic resonance images (MRI) from the ADNI database. The generative model is first estimated on a control population, then, for each subject, the markers are computed for each acquisition. The longitudinal evolution of these markers is then studied in relation with the clinical diagnosis of the subjects and used to generate possible morphological evolution. In the model, the morphological changes associated with normal aging are mainly found around the ventricles, while the Alzheimer's disease specific changes are more located in the temporal lobe and the hippocampal area. The statistical analysis of these markers highlights differences between clinical conditions even though the inter-subject variability is quiet high. In this context, the model can be used to generate plausible morphological trajectories associated with the disease. Our method gives two interpretable scalar imaging biomarkers assessing the effects of aging and disease on brain morphology at the individual and population level. These markers confirm an acceleration of apparent aging for Alzheimer's subjects and can help discriminate clinical conditions even in prodromal stages. More generally, the joint modeling of normal and pathological evolutions shows promising results to describe age-related brain diseases over long time scales.

Clement Abi Nader, Nicholas Ayache, Philippe Robert et al.

We introduce a probabilistic generative model for disentangling spatio-temporal disease trajectories from series of high-dimensional brain images. The model is based on spatio-temporal matrix factorization, where inference on the sources is constrained by anatomically plausible statistical priors. To model realistic trajectories, the temporal sources are defined as monotonic and time-reparametrized Gaussian Processes. To account for the non-stationarity of brain images, we model the spatial sources as sparse codes convolved at multiple scales. The method was tested on synthetic data favourably comparing with standard blind source separation approaches. The application on large-scale imaging data from a clinical study allows to disentangle differential temporal progression patterns mapping brain regions key to neurodegeneration, while revealing a disease-specific time scale associated to the clinical diagnosis.

Santiago Silva, Boris Gutman, Eduardo Romero et al.

At this moment, databanks worldwide contain brain images of previously unimaginable numbers. Combined with developments in data science, these massive data provide the potential to better understand the genetic underpinnings of brain diseases. However, different datasets, which are stored at different institutions, cannot always be shared directly due to privacy and legal concerns, thus limiting the full exploitation of big data in the study of brain disorders. Here we propose a federated learning framework for securely accessing and meta-analyzing any biomedical data without sharing individual information. We illustrate our framework by investigating brain structural relationships across diseases and clinical cohorts. The framework is first tested on synthetic data and then applied to multi-centric, multi-database studies including ADNI, PPMI, MIRIAD and UK Biobank, showing the potential of the approach for further applications in distributed analysis of multi-centric cohorts

Marco Lorenzi, Maurizio Filippone

We introduce a novel generative formulation of deep probabilistic models implementing "soft" constraints on their function dynamics. In particular, we develop a flexible methodological framework where the modeled functions and derivatives of a given order are subject to inequality or equality constraints. We then characterize the posterior distribution over model and constraint parameters through stochastic variational inference. As a result, the proposed approach allows for accurate and scalable uncertainty quantification on the predictions and on all parameters. We demonstrate the application of equality constraints in the challenging problem of parameter inference in ordinary differential equation models, while we showcase the application of inequality constraints on the problem of monotonic regression of count data. The proposed approach is extensively tested in several experimental settings, leading to highly competitive results in challenging modeling applications, while offering high expressiveness, flexibility and scalability.