Skip to main content
  • Expression of Interest

    A reference framework for XAI
    Benchmark; Evaluation; Critical applications

    Here is our (not restricted) approach:

    - A focus on critical applications, to enhance the trustworthiness of systems with strong economical/health/society impacts (industry, medecine, defense, etc.) ;
    - Develop a reference framework for explainability, to offer concrete ways of integrating explainability in standards and regulations;
    - Benchmark explainable solutions, to provide a rigorous qualification of AI.

    If you are interested in joining us in the application for this call, please indicate which skill(s) you may bring forward, and your first idea (if any yet) of the approach you plan to take. From your inputs, we will try and form a coherent consortium.

    {Empty}

    In our opinion, a solid consortium will present these skills:

    • Skill A - Usecase(s) provider: critical applications where explainability is a requirement and an issue. Expertise of the domain of application of the AI solution; provides potentially testing panels;
    • Skill B - Explainability solutions expert: a strong know-how on the development of such solutions. Provides AI solutions and/or XAI solutions;
    • Skill C - Expert in human sciences (ergonomy, law, economy, etc.): they provide models to guide the design of the evaluation plans. This can address human factors and/or society factors ;
    • Skill D - Benchmarking specialist (LNE): we design the evaluation framework.

    These skills are not exclusive: partners may cover several of them.

    Different from those already involved

    As an independant evaluating and certifying body, LNE contributes to the development of standards and test methods for the qualification of AI systems. LNE wishes to explore the design of a reference framework for explainability, including the type of information to be extracted for compliance assessment, the conditions of the application of explainability solutions, but also the assessment of the performance of these solutions.

     

    {Empty}
Search