Timezone: »

 
Poster
Zero-Shot Transfer with Deictic Object-Oriented Representation in Reinforcement Learning
Ofir Marom · Benjamin Rosman

Wed Dec 05 02:00 PM -- 04:00 PM (PST) @ Room 517 AB #151
in Wed Poster Session B »

Object-oriented representations in reinforcement learning have shown promise in transfer learning, with previous research introducing a propositional object-oriented framework that has provably efficient learning bounds with respect to sample complexity. However, this framework has limitations in terms of the classes of tasks it can efficiently learn. In this paper we introduce a novel deictic object-oriented framework that has provably efficient learning bounds and can solve a broader range of tasks. Additionally, we show that this framework is capable of zero-shot transfer of transition dynamics across tasks and demonstrate this empirically for the Taxi and Sokoban domains.

Keywords: Multitask and Transfer Learning · Model-Based RL

Author Information

Ofir Marom (University of the Witwatersrand)
Benjamin Rosman (University of the Witwatersrand / CSIR)

More from the Same Authors