Dear all,
We are inviting applications for a PhD position at LORIA/INRIA in Nancy,
France.
*Title*: Information-Oriented Planning with rho-POMDPs
The objective of the PhD thesis is to propose new algorithms for solving
information-oriented planning problems, i.e., deciding which sequence of
actions to perform in order to gather information regarding certain
hidden variables. Examples include: a robot willing to localize itself;
a museum surveillance system willing to track visitors; or a medical
system willing to monitor the state of a patient.
Such problems are formalized as rho-POMDPs, an extension of POMDPs with
belief-dependent rewards. First algorithms have already been proposed,
but they rely on specific properties of the reward function. This PhD
thesis will investigate Monte Carlo Tree Search (MCTS) approaches to do
without these assumptions and only requiring a generative model (a
simulator).
*Keywords*: Artificial Intelligence, Partially Observable Markov
Decision Processes (POMDPs), Information-Oriented Control, Monte Carlo
Tree Search (MCTS).
*Supervisors*: Olivier Buffet and Vincent Thomas
*Closing date*: May 1.
*More details + How to apply*:
https://jobs.inria.fr/public/classic/en/offres/2019-01556
Best regards,
Olivier Buffet and Vincent Thomas
_______________________________________________
uai mailing list
uai@ENGR.ORST.EDU
https://secure.engr.oregonstate.edu/mailman/listinfo/uai
