Internship proposal.
- Title: Bridging deductive reasoning and induction
- Supervisor: Philippe Preux
- Duration: 5 to 6 months
- When: Spring-Summer 2021
- Where: ScooL, Inria Lille, Villeneuve d'Ascq, France, if sanitary conditions permit it, remotely otherwise.
- Keywords: machine learning, ontology, symbolic reasoning, deductive reasoning, knwoledge-based reasoning
- Context:
Human beings perform at once deductive reasoning (if A, and A →B, then B) and induction (given a few examples, sometimes a single one, of a given type of objects, one can recognize most of, if not all, the objects of the same kind).
In AI, these two modes of creation of new knowledge belong to two unconnected worlds. This internship is drawn by the feeling that we should try to connect these two worlds. Many limitations of today's AI rest on this disconnection.
- What:
The goal of this internship is to study how one could bridge these two worlds. This study will be of fundamental nature, but we also want that a software program will be developed to perform an experimental study of the ideas being proposed so as to assess their relevance.
- Who:
This internship is tailored as the final project of a master degree in computer science. We expect a strong background either in symbolic AI (logics, ontology, ...), or in machine learning, if not both (which is better). In either case, the intern should be very curious about the other world that (s)he will have to explore.
The intern will read English w/o any difficulty and (s)he is able to work in English, make a scientific presentation in English, interact in English more generally. (We do not expect that the intern can speak French.)
- Bibliography:
- Besold et al., Neural-Symbolic Learning and Reasoning: A Survey and Interpretation, arxiv 1711.03902.
- Working environment: Scool is a well-known research group in reinforcement learning and bandits. It is composed of 6 permanent researchers, 20+ PhD students, a couple of post-docs and engineers. Scool provides a very rich and stimulating for doing cutting-edge research.
Back to homepage.