The URBEX project aims at developing a validated, breakthrough, fast-running, meshless model for the propagation of blast waves in urban configurations, accounting for all urban effects: multiple reflections, diffractions, channeling in urban canyons and urban canopy bypassing. The project intends to fill the gap identified between empirical operational or normative approaches (circular danger zones) and the use of 3D numerical codes requiring specific expertise and significant computing resources. Its applications concern global security and industrial security, as well as more generally the protection of people and goods. The URB(EX)3 model will be able to compute a complex overpressure wave shape at any point in the zone of interest. Consequences on people and infrastructures will be computed using both regulatory overpressure thresholds and probabilistic consequence models from the literature. The model will be embedded in an existing user-friendly platform, DEMOCRITE, developed during the eponymous ANR project and already tested by several organizations. Model inversion will also be addressed for two applications: 1. Definition of a protection perimeter / a vehicle exclusion zone around a building or a user-defined zone for a given threat level; 2. Forensic analysis of the observed damage to estimate the likely equivalent explosive mass. This model will be supported by series of analytical, high-quality, small-scale experiments in order to: - Guide the fitting of model parameters, when necessary, - Investigate specific phenomena, for instance blast channeling in city streets, - Help quantifying model uncertainties and safety margins, - Test the model on more global configurations, - Prepare for further extensions (terrain effects, height-of-burst explosions…). The only assumptions are: - The explosion takes place at ground level, assuming hemispherical high-explosive shape, - Free-field functions for blast parameters as a function of reduced distance: P(Z), I+(Z) , etc. are supposed to be known for the explosive under study, - The urban configuration is given (for instance in the shapefile format) by a set of buildings, each of them described by its polygonal footprint and its height (this is the standard for the IGN BD TOPO® v3 database, and corresponds also to the CityGML level of detail 1 description), - The terrain is flat. However, some of these limiting assumptions (height of burst explosion, actual vs. simplified building façades, terrain slope, clearing effects…) will be questioned during the project through dedicated experiments and/or numerical simulations. Finally, the partners have collected letters of support from several organizations, which will be part of the project's advisory board: IRCGN, LCPP, EURENCO, IRSN and CETID. The objective is to ensure that the project meets the actual needs of various potential users.

Federated learning (FL) is a promising paradigm that is gaining grip in the context of privacy-preserving machine learning for edge computing systems. Thanks to FL, several data owners called clients (e.g., organizations in cross-silo FL) can collaboratively train a model on their private data, without having to send their raw data to external service providers. FL was rapidly adopted in several thriving applications such as digital healthcare, that is generating the world’s largest volume of data. In healthcare systems, the problems of privacy and bias are particularly important. Although FL is a first step towards privacy by keeping the data local to each client, this is not sufficient since the model parameters shared by FL are vulnerable to privacy attacks, as shown in a line of recent literature. Thus, there is a need to design new FL protocols that are robust to such privacy attacks. Furthermore, FL clients may have very heterogeneous and imbalanced data, which may incur FL model bias, with disparities among socioeconomic and demographic groups. Recent studies show that the use of AI may further exacerbate disparities between groups, and that FL may be a vector of bias propagation among different FL client. In this context, recent works appeared in ICDE, NDSS and AAAI show that bias, privacy and data curation and preparation (i.e., for correcting missing or duplicate values) compete; handling them independently – as done usually – may have negative side-effects on each other. Therefore, there is a need for a novel multi-objective method for FL data preparation and cleaning, bias mitigation, and protection against privacy threats. This is particularly challenging in FL where no global knowledge about statistical information of the overall heterogeneous data is available, a knowledge that is necessary in classical state-of-the-art techniques. CITADEL project tackles this challenge and aims to precisely handle the issues raised at the intersection of FL data cleaning, their privacy and their bias, through: (i) Novel distributed FL protocols; (ii) A multi-objective approach to take into account privacy, fairness and quality aspects, these objectives being antagonistic; (ii) Applying these techniques in two use cases of FL-based digital health with real medical data.