Affect-based Intrinsic Rewards for Learning General Representations

12/01/2019
by   Dean Zadok, et al.
24

Positive affect has been linked to increased interest, curiosity and satisfaction in human learning. In reinforcement learning extrinsic rewards are often sparse and difficult to define, intrinsically motivated learning can help address these challenges. We argue that positive affect is an important intrinsic reward that effectively helps drive exploration that is useful in gathering experiences critical to learning general representations. We present a novel approach leveraging a task-independent intrinsic reward function trained on spontaneous smile behavior that captures positive affect. To evaluate our approach we trained several downstream computer vision tasks on data collected with our policy and several baseline methods. We show that the policy based on intrinsic affective rewards successfully increases the duration of episodes, area explored and reduces collisions. The impact is increased speed of learning for several downstream computer vision tasks.

READ FULL TEXT

page 1

page 5

page 6

page 7

research
05/25/2018

Visceral Machines: Reinforcement Learning with Intrinsic Rewards that Mimic the Human Nervous System

The human autonomic nervous system has evolved over millions of years an...
research
07/14/2021

Experimental Evidence that Empowerment May Drive Exploration in Sparse-Reward Environments

Reinforcement Learning (RL) is known to be often unsuccessful in environ...
research
03/18/2019

Scheduled Intrinsic Drive: A Hierarchical Take on Intrinsically Motivated Exploration

Exploration in sparse reward reinforcement learning remains a difficult ...
research
10/07/2022

Generative Augmented Flow Networks

The Generative Flow Network is a probabilistic framework where an agent ...
research
04/10/2017

Stochastic Neural Networks for Hierarchical Reinforcement Learning

Deep reinforcement learning has achieved many impressive results in rece...
research
08/26/2022

Play with Emotion: Affect-Driven Reinforcement Learning

This paper introduces a paradigm shift by viewing the task of affect mod...
research
07/11/2023

Intrinsically motivated graph exploration using network theories of human curiosity

Intrinsically motivated exploration has proven useful for reinforcement ...

Please sign up or login with your details

Forgot password? Click here to reset