Poster
Hindsight Credit Assignment
Anna Harutyunyan · Will Dabney · Thomas Mesnard · Mohammad Gheshlaghi Azar · Bilal Piot · Nicolas Heess · Hado van Hasselt · Gregory Wayne · Satinder Singh · Doina Precup · Remi Munos

Tue Dec 10th 05:30 -- 07:30 PM @ East Exhibition Hall B + C #204
in Reinforcement Learning and Planning -- Reinforcement Learning »

We consider the problem of efficient credit assignment in reinforcement learning. In order to efficiently and meaningfully utilize new data, we propose to explicitly assign credit to past decisions based on the likelihood of them having led to the observed outcome. This approach uses new information in hindsight, rather than employing foresight. Somewhat surprisingly, we show that value functions can be rewritten through this lens, yielding a new family of algorithms. We study the properties of these algorithms, and empirically show that they successfully address important credit assignment challenges, through a set of illustrative tasks.

Keywords: Reinforcement Learning and Planning -> Reinforcement Learning · Reinforcement Learning and Planning -> Decision and Control; Reinforcement Learning and Planning -> Markov Decision Processes

Author Information

Anna Harutyunyan (DeepMind)
Will Dabney (DeepMind)
Thomas Mesnard (DeepMind)
Mohammad Gheshlaghi Azar (DeepMind)
Bilal Piot (DeepMind)
Nicolas Heess (Google DeepMind)
Hado van Hasselt (DeepMind)
Greg Wayne (Google DeepMind)
Satinder Singh (DeepMind)
Doina Precup (DeepMind)
Remi Munos (DeepMind)

Related Events (a corresponding poster, oral, or spotlight)

More from the Same Authors