- Barcelona Supercomputing Center (BSC)
- IKERLAN S. Coop.
- AIKO SRL
- RISE RESEARCH INSTITUTES OF SWEDEN AB
- NAVINFO EUROPE BV
- EXIDA DEVELOPMENT SRL
2022 - 2025
3 891 875 €
European programmes: Horizon Europe
SAFEXPLAIN aims to reduce the current fundamental gap between Functional Safety (FUSA) requirements of Critical Autonomous Artificial Intelligence (AI) based Systems and the complex nature of deep learning (DL) solutions. The lack of transparency (mainly explainability and traceability), and the data-dependent and stochastic nature of DL software clash against the need for deterministic, verifiable and pass/fail test-based software solutions for critical systems.
SAFEXPLAIN tackles this challenge by providing a novel and flexible approach to work towards the certification – hence adoption – of DL-based solutions in critical systems by architecting transparent DL solutions that allow explaining why they satisfy functional safety requirements, with end-to-end traceability, with specific approaches to explain whether predictions can be trusted, and with strategies to reach (and prove) correct operation, in accordance with certification standards. SAFEXPLAIN will also devise alternative and increasingly complex functional safety design safety patterns for different DL usage levels (i.e. with varying safety requirements) that will allow using DL in a wide set of critical systems, for varying levels of criticality and fault tolerance.
IKERLAN will participate with a multidisciplinary team of functional safety engineers, Data Analytics and Artificial intelligence specialists and software engineers with high expertise in the railway domain.
This team will cover several aspects of the project including the leadership of Safety Assessment work package, which aims to identify the needs and create the required designs and arguments to build a safe DL-based solution. This work is complemented by the study and adaptation of DL models and libraries that ease the compliance of functional safety standards. Finally, all this work will be evaluated in a Railway case study that aims to prevent the collision of autonomous trains with objects or persons in the track with DL-based object detection algorithms.