- MBDA's aim is to support staff with trusted AI-enabled devices.
- In the short term: it will be supporting decision on the battlefield;
- In the medium term: integration into on-board functional chains at the proper level.
- To achieve these goals, MBDA worked with Numalis and BPI through an AI Challenge.
- The aim is twofold: demonstrating the operational capabilities of Numalis' formal approaches to underline the robustness of Machine Learning algorithms; and working on the ability to cover and visualize the system's risk coverage in its field of use.
- This work has helped to define AI algorithm validation chains in line with developed standards at the international level (with a part supported by Numalis).
- The use case will illustrate the approach, the obtained results and how it could be adapted to other sectors also sharing an imperative need for reliable AI systems.