02/02/2021

DeepSynth: Automata Synthesis for Automatic Task Segmentation in Deep Reinforcement Learning

Mohammadhosein Hasanbeig, Natasha Yogananda Jeppu, Alessandro Abate, Tom Melham, Daniel Kroening

Keywords:

Abstract Paper Similar Papers

Abstract: This paper proposes DeepSynth, a method for effective training of deep Reinforcement Learning (RL) agents when the reward is sparse and non-Markovian, but at the same time progress towards the reward requires achieving an unknown sequence of high-level objectives. Our method employs a novel algorithm for synthesis of compact automata to uncover this sequential structure automatically. We synthesise a human-interpretable automaton from trace data collected by exploring the environment. The state space of the environment is then enriched with the synthesised automaton so that the generation of a control policy by deep RL is guided by the discovered structure encoded in the automaton. The proposed approach is able to cope with both high-dimensional, low-level features and unknown sparse non-Markovian rewards. We have evaluated DeepSynth's performance in a set of experiments that includes the Atari game Montezuma's Revenge. Compared to existing approaches, we obtain a reduction of two orders of magnitude in the number of iterations required for policy synthesis, and also a significant improvement in scalability.

The video of this talk cannot be embedded. You can watch it here:
https://slideslive.com/38949011
(Link will open in new window)
 0
 0
 0
 0
  Share    
This is an embedded video. Talk and the respective paper are published at AAAI 2021 virtual conference. If you are one of the authors of the paper and want to manage your upload, see the question "My papertalk has been externally embedded..." in the FAQ section.

Comments

Post Comment
no comments yet
code of conduct: tbd Characters remaining: 140

Similar Papers

 1
 0
 0
 0
 3:50 
 0
 0
 0
 0
 3:21 
 0
 0
 0
 0
 19:12 
 0
 0
 0
 1
 10:04 
 0
 0
 0
 0
 5:07 
 0
 0
 0
 0
 12:21 
 0
 0
 0
 0
 2:42 
 0
 0
 0
 0
 5:30 
 0
 0
 0
 0
 4:28 
 0
 0
 0
 0
 6:48 
 0
 0
 0
 0
 3:20 
 0
 0
 0
 0
 13:14 
 0
 0
 0
 0
 5:17 
 0
 0
 0
 0
 13:50 
 0
 0
 0
 0
 10:21 
 0
 0
 0
 0
 3:14 
 0
 0
 0
 0
 4:53 
 0
 0
 0
 0
 5:04 
 1
 0
 0
 0
 5:21 
 0
 0
 0
 0
 5:04 
 0
 0
 0
 0
 4:55 
 0
 0
 0
 0
 11:27 
 0
 0
 0
 0
 14:57 
 0
 0
 0
 0
 3:10 
 0
 0
 0
 0
 14:54 
 0
 0
 0
 0
 19:35 
 0
 0
 0
 0
 8:39 
 0
 0
 0
 0
 14:49 
 0
 0
 0
 0
 4:38 
 0
 0
 0
 0
 10:30 
 0
 0
 0
 0
 5:09 
 0
 0
 0
 0
 3:20 
 0
 0
 0
 0
 3:59 
 0
 0
 0
 0
 4:38 
 0
 0
 0
 0
 13:36 
 0
 0
 0
 0
 13:23 
 0
 0
 0
 0
 2:53 
 0
 0
 0
 0
 5:12 
 0
 0
 0
 0
 3:25 
 0
 0
 0
 0
 14:47