Research

Separating value functions across time-scales

June 10, 2019

Abstract

In many finite horizon episodic reinforcement learning (RL) settings, it is desirable to optimize for the undiscounted return – in settings like Atari, for instance, the goal is to collect the most points while staying alive in the long run. Yet, it may be difficult (or even intractable) mathematically to learn with this target. As such, temporal discounting is often applied to optimize over a shorter effective planning horizon. This comes at the cost of potentially biasing the optimization target away from the undiscounted goal. In settings where this bias is unacceptable – where the system must optimize for longer horizons at higher discounts – the target of the value function approximator may increase in variance leading to difficulties in learning. We present an extension of temporal difference (TD) learning, which we call TD(∆), that breaks down a value function into a series of components based on the differences between value functions with smaller discount factors. The separation of a longer horizon value function into these components has useful properties in scalability and performance. We discuss these properties and show theoretic and empirical improvements over standard TD learning in certain settings.

Download the Paper

AUTHORS

Written by

Joshua Romoff

Peter Henderson

Ahmed Touati

Emma Brunskill

Joelle Pineau

Yann Ollivier

Publisher

International Conference on Machine Learning (ICML)

Related Publications

June 14, 2020

Iterative Answer Prediction with Pointer-Augmented Multimodal Transformers for TextVQA | Facebook AI Research

Ronghang Hu, Amanpreet Singh, Trevor Darrell, Marcus Rohrbach

June 14, 2020

June 17, 2019

Computer Vision

DMC-Net: Generating Discriminative Motion Cues for Fast Compressed Video Action Recognition | Facebook AI Research

Zheng Shou, Xudong Lin, Yannis Kalantidis, Laura Sevilla-Lara, Marcus Rohrbach, Shih-Fu Chang, Zhicheng Yan

June 17, 2019

June 18, 2019

Computer Vision

Embodied Question Answering in Photorealistic Environments with Point Cloud Perception | Facebook AI Research

Erik Wijmans, Samyak Datta, Oleksandr Maksymets, Abhishek Das, Georgia Gkioxari, Stefan Lee, Irfan Essa, Devi Parikh, Dhruv Batra

June 18, 2019

August 01, 2019

NLP

Simple and Effective Curriculum Pointer-Generator Networks for Reading Comprehension over Long Narratives | Facebook AI Research

Yi Tay, Shuohang Wang, Luu Anh Tuan, Jie Fu, Minh C. Phan, Xingdi Yuan, Jinfeng Rao, Siu Cheung Hui, Aston Zhang

August 01, 2019

Help Us Pioneer The Future of AI

We share our open source frameworks, tools, libraries, and models for everything from research exploration to large-scale production deployment.