Fullscreen
Loading...
 
Tao
Print

Seminar27032018

March, Tuesday 27th

14:30 (room 2014, 'Digiteo Shannon' 660 building) (see location )

Nizam Makdoud

(TAU team)

Title: Intrinsic Motivation, Exploration and Deep Reinforcement Learning


Abstract

Exploration in complex environment is a key to increase the adoption of deep reinforcement learning (in an industrial context). During this talk, I will present the work done at THALES on deep reinforcement learning applied on the internal urban simulator SESTAR. In an industrial context, state of the art algorithms struggle because “realistic” environments require the ability to explore. The question of the exploration is central and I will present Intrinsic motivations as a way to enhance traditional reinforcement algorithms.




Contact: guillaume.charpiat at inria.fr
All TAU seminars: here

Contributors to this page: guillaume .
Page last modified on Monday 26 of March, 2018 14:13:27 CEST by guillaume.