Fast Task Inference with Variational Intrinsic Successor Features

Abstract

It has been established that diverse behaviors spanning the controllable subspace of an Markov decision process can be trained by rewarding a policy for being distinguishable from other policies \citep{gregor2016variational, eysenbach2018diversity, warde2018unsupervised}. However, one limitation of this formulation is generalizing behaviors beyond the finite set being explicitly learned, as is needed for use on subsequent tasks. Successor features \citep{dayan93improving, barreto2017successor} provide an appealing solution to this generalization problem, but require defining the reward function as linear in some grounded feature space. In this paper, we show that these two techniques can be combined, and that each method solves the other's primary limitation. To do so we introduce Variational Intrinsic Successor FeatuRes (VISR), a novel algorithm which learns controllable features that can be leveraged to provide enhanced generalization and fast task inference through the successor feature framework. We empirically validate VISR on the full Atari suite, in a novel setup wherein the rewards are only exposed briefly after a long unsupervised phase. Achieving human-level performance on 14 games and beating all baselines, we believe VISR represents a step towards agents that rapidly learn from limited feedback.

Accepted for publication at ICLR 2020

References in corpus (9)

• Adam: A Method for Stochastic Optimization
• Conditional Image Generation with PixelCNN Decoders
• Semi-Supervised Learning with Ladder Networks
• IMPALA: Scalable Distributed Deep-RL with Importance Weighted Actor-Learner Architectures
• Model-Based Reinforcement Learning for Atari
• Large-Scale Study of Curiosity-Driven Learning
• Variational Information Maximisation for Intrinsically Motivated Reinforcement Learning
• Variational Intrinsic Control
• Wasserstein Dependency Measure for Representation Learning

Cited by in corpus (13)

• An information-theoretic perspective on intrinsic motivation in reinforcement learning: a survey
• SEED RL: Scalable and Efficient Deep-RL with Accelerated Central Inference
• Dynamics-Aware Quality-Diversity for Efficient Learning of Skill Repertoires
• Behavior From the Void: Unsupervised Active Pre-Training
• Reset-Free Lifelong Learning with Skill-Space Planning
• Fast Adaptation via Policy-Dynamics Value Functions
• UneVEn: Universal Value Exploration for Multi-Agent Reinforcement Learning
• Disentangled Cumulants Help Successor Representations Transfer to New Tasks
• PsiPhi-Learning: Reinforcement Learning with Demonstrations using Successor Features and Inverse Temporal Difference Learning
• Pretrained Encoders are All You Need
• Planning in Hierarchical Reinforcement Learning: Guarantees for Using Local Policies
• Discovering a set of policies for the worst case reward
• Successor Feature Neural Episodic Control