AIDA

Using explanatory models for deep learning algorithms

The aim of the AIDA project is to develop robust and rigorous techniques for actionable explanations in AI (aXAI). To improve the transparency of spatio-temporal data and models, AIDA proposes a causal modelling framework for multivariate time series that captures the underlying process of temporal data generation.

Main funder: ANR PRC

Nature: Multidisciplinary

UT Capitole role: Partner

Project: National

Impact: societal

Stakeholder: IRIT - Julien Aligon