Superstition In The Network: Deep Reinforcement Learning Plays Deceptive Games

Abstract

Deep reinforcement learning has learned to play many games well, but failed on others. To better characterize the modes and reasons of failure of deep reinforcement learners, we test the widely used Asynchronous Actor-Critic (A2C) algorithm on four deceptive games, which are specially designed to provide challenges to game-playing agents. These games are implemented in the General Video Game AI framework, which allows us to compare the behavior of reinforcement learning-based agents with planning agents based on tree search. We find that several of these games reliably deceive deep reinforcement learners, and that the resulting behavior highlights the shortcomings of the learning algorithm. The particular ways in which agents fail differ from how planning-based agents fail, further illuminating the character of these algorithms. We propose an initial typology of deceptions which could help us better understand pitfalls and failure modes of (deep) reinforcement learning.

Stats

• citations5
• S2 citations—
• github stars0
• HF likes0
• heat score5.84
• arxiv keybontrager2019superstition

Related papers

• Deception In Social Learning: A Multi-agent Reinforcement Learning Perspective (2021)0.00
• When Your Ais Deceive You: Challenges Of Partial Observability In Reinforcement Learning From Human Feedback (2024)0.00
• Honesty Is The Best Policy: Defining And Mitigating AI Deception (2023)0.00
• Reinforcing Competitive Multi-agents For Playing 'so Long Sucker' (2024)0.00
• Diagnosing And Exploiting The Computational Demands Of Videos Games For Deep Reinforcement Learning (2023)0.00
• Simplified Action Decoder For Deep Multi-agent Reinforcement Learning (2019)4.03
• Tactics Of Adversarial Attack On Deep Reinforcement Learning Agents (2017)17.32
• Impartial Games: A Challenge For Reinforcement Learning (2022)0.00