Artificial intelligence, uncertain reasoning
Aix-Marseille Université – Laboratoire d’Informatique et Systèmes (LIS)
Campus de Saint Jérôme
52 Av. Escadrille Normandie Niemen F13013 Marseille – France
The goal of this PhD thesis is to provide efficient tools for optimizing the number and locations of electric car charging stations in the Aix-Marseille Metropolis area. To do so, the following problems need be addressed: i) determining the trips that car users plan to do; ii) the paths that they will use for their trips and their possible congestion; iii) the reasons why they are doing such trips. The first problem is known in the literature as the origin-destination problem and has received many contributions [1,2]. The second problem is a classical trip assignment problem. To solve them, the PhD candidate will learn a Bayesian network-based model [3,4] from datasets available, notably from Aix-Marseille Metropolis, and exploit it to predict the car users’ behaviors. Here, one significant issue lies in the large size of Aix-Marseille Metropolis, which prevents classical algorithms to be used, both for learning and prediction. New algorithms will therefore need to be developed. The continuous nature of the variables as well as the non-stationary/seasonal nature of the traffic flow process will also be issues to be addressed. The third aforementioned problem (determining the reasons why the users perform their trips) can be cast as a causality problem and Causal Bayesian networks  will be used for this purpose. Here again, the high-dimensional space in which computations will be made as well as the continuous nature of the variables will be major issues to address.
 Enrique Castillo, José Marı́a Menéndez, Santos Sánchez-Cambronero (2008) “Predicting traffic flow using Bayesian networks”, Transportation Research Part B, Vol. 42, pp. 482–509.
 Shiliang Sun, Changshui Zhang, Guoqiang Yu (2006) “A Bayesian Network Approach to Traffic Flow Forecasting”, IEEE Transactions on Intelligent Transportation Systems, Vol. 7, N°1, pp. 124–132.
 Judea Pearl (1988) Probabilistic Reasoning in Intelligent Systems: Networks of Plausible Inference, Morgan-Kaufmann.
 Daphne Koller, Nir Friedman (2009) Probabilistic Graphical Models: Principles and Techniques, MIT Press.
 Judea Pearl (2009) Causality: Models, reasoning and Inference, 2nd edition, Cambridge University Press.
Monthly gross salary: 1768 euros
QUALIFICATIONS/SKILLS/EDUCATION & RESEARCH REQUIREMENTS/DUTIES
We are looking for a highly motivated candidate with a strong background in mathematics and computer science. He/She should have recently completed a master’s degree or should be about to complete it. The candidate should demonstrate very good programming skills, notably in C++ and Python. An experience or some understanding of probabilistic graphical models, e.g., of Bayesian networks, would be appreciated. We also expect the candidate to be open-minded, curious and autonomous.
Applications should be made before Tuesday, November 30.
REQUESTED DOCUMENTS OF APPLICATION
- Curriculum Vitæ
- Cover letter
- Grades and transcripts after High school, including rankings
|Job Category||Internship and training|