Un événement imaginé par
INRIA - Stage - SPADES Causal Explanations for Embedded Systems

Description
Internship L3 Master 1/ 2 - Required level: BAC +3, 4 ou 5
Who we are?
Inria is the French national research institute dedicated to digital science and technology. World-class research, technological innovation and entrepreneurial risk are its DNA. Its 215 agile project teams, most of which are joint with academic partners, involve more than 3,900 scientists in tackling the challenges of digital technology, often at the interface with other disciplines.
https://inria.fr/fr
Inria is headquartered in Rocquencourt and has 9 research centres.
Inria Grenoble is active in the fields of high-performance computing, verification and embedded systems, modeling of the environment at multiple levels, and data science and artificial intelligence. The center is a top-level scientific institute with an extensive network of international collaborations in Europe and the rest of the world.
Context:
In safety-critical embedded systems, a recently developed theory of causal explanations allows, in case of a failure, to pinpoint the cause(s) of the failure so as to help in debugging, take appropriate
actions, or help establish liability. Such explanations are needed in particular for designing explainable intelligent systems that should not only be safe and trustworthy, but also make explainable decisions.
We have recently proposed an approach to construct causal explanations based on a robustness function that measures how close the system is to the violation of an expected safety property.
Objective of the internship:
The goal of the internship is to continue to develop the theory and/or implement and validate the approach:
The robustness function is formalized as a fixpoint. A first goal is to implement this fixpoint computation into a prototype tool, and apply it to textbook examples and practical case studies.
A second, more theoretical outcome would be to prove whether the computation terminates in all cases, or find a counter-example.
Third, we are interested in extending the approach so as to quantify the responsibility of a single component in the system failure, and applying the results to models of actual systems that have failed.
Assignment:
The internship may be followed-up by a PhD thesis.
Profile and Skills:
Candidates should have solid programming skills and some background in linear algebra or formal methods. Interest in causal reasoning would be greatly appreciated.
Useful link: https://team.inria.fr/spades/
Additional information:
- Périod: flexible, to be defined with the candidate
- Duration: at least 5 months - Location: Montbonnot-Saint-Martin
- Contact to Apply: The intern will be co-advised by G. Goessler (SPADES team, INRIA Grenoble) and T. Dang (MOHYTOS team, Verimag, UGA).
Expérience
Thématique
Localisation du poste
Nombre de poste à pourvoir
1