Site Université Toulouse Capitole
We are searching for a Master student with background in logic and computer science. An interest for interdisciplinary work is helpful.
Project Description
We aim at developing formal methods that will facilitate the handling of exceptions in legal reasoning to law practitioners, in particular exceptions that are not explicitly mentioned in legal rules. To do so, we will combine two approaches in artificial intelligence: default logic and case-based reasoning. We will study formal properties of this combination and propose a prototype of software for legal decision support based on this theoretical work. This software prototype will benefit from the open data of justice decisions.
Context
Non-monotonic reasoning formalisms like default logic are a way to represent information with exceptions and are then particularly well suited to legal reasoning. However, these formalisms require to anticipate the exceptions that may arise. For unplanned exceptions, the agent that takes a decision (e.g. a judge who must decide about a case, or a lawyer who prepares his client’s defense) faces a new situation, he applies (by default) the decision that must be taken in the general case. Let us illustrate this with a simple example. Assume there is a rule that forbids vehicles to go through a park, and that the driver of an ambulance decides to go through the park to reach the hospital faster (and thus save the life of the patient who’s in the ambulance). If it is not explicitly stated that there is an exception for this situation, then the driver of the ambulance should be fined for violating the rule that forbids going through the park.
For this reason, we propose to study a hybrid decision mechanism, which uses default logic to handle exceptional situations, and case-based reasoning to determine whether a new situation is like a known exception or whether it corresponds to the general case. Let us continue the example of the ambulance driver and assume that there is a rule regarding red lights (it is forbidden to cross the road when there is a red light) with some exception (it is permitted to some vehicles –firefighters, police, ambulances—to cross a red light when there is an emergency). Using case-based reasoning, we can detect the similarity between both situations (a general interdiction to vehicles, and an exception for one of them in cases of emergency). The case-based reasoning system can identify this similarity and propose to the lawyer to adapt the exception of the red-light rule and apply it to the park situation, creating a new exception (it is permitted to firefighters, police and ambulances to go through the park in situations of emergency). The lawyer can then decide to take this exception into account (in which case it is added to the system based on default logic). Otherwise, the proposition of new exception is ignored, and the default rule is applied.
Supervision
The research internship will be supervised by Mathieu Carpentier (Institut Maurice Hauriou, École de Droit de Toulouse Recherche), Sylvie Doutre (Institut de Recherche en Informatique de Toulouse) and Jean-Guy Mailly (Institut de Recherche en Informatique de Toulouse), at Université Toulouse Capitole. It is related to the chair AIDAL.
Application
To apply, send to the supervisors (mathieu.carpentier@ut-capitole.fr, sylvie.doutre@ut-capitole.fr, jean-guy.mailly@ut-capitole.fr) your resume, your academic record and any other information that can be relevant.