DeepAI AI Chat
Log In Sign Up

Interactive Explanations: Diagnosis and Repair of Reinforcement Learning Based Agent Behaviors

by   Christian Arzate Cruz, et al.

Reinforcement learning techniques successfully generate convincing agent behaviors, but it is still difficult to tailor the behavior to align with a user's specific preferences. What is missing is a communication method for the system to explain the behavior and for the user to repair it. In this paper, we present a novel interaction method that uses interactive explanations using templates of natural language as a communication method. The main advantage of this interaction method is that it enables a two-way communication channel between users and the agent; the bot can explain its thinking procedure to the users, and the users can communicate their behavior preferences to the bot using the same interactive explanations. In this manner, the thinking procedure of the bot is transparent, and users can provide corrections to the bot that include a suggested action to take, a goal to achieve, and the reasons behind these decisions. We tested our proposed method in a clone of the video game named Super Mario Bros., and the results demonstrate that our interactive explanation approach is effective at diagnosing and repairing bot behaviors.


page 1

page 3

page 7


ASQ-IT: Interactive Explanations for Reinforcement-Learning Agents

As reinforcement learning methods increasingly amass accomplishments, th...

MarioMix: Creating Aligned Playstyles for Bots with Interactive Reinforcement Learning

In this paper, we propose a generic framework that enables game develope...

Understanding the Unforeseen via the Intentional Stance

We present an architecture and system for understanding novel behaviors ...

Contrastive Explanations for Reinforcement Learning in terms of Expected Consequences

Machine Learning models become increasingly proficient in complex tasks....

Cases for Explainable Software Systems:Characteristics and Examples

The need for systems to explain behavior to users has become more eviden...

T-REx: Table Repair Explanations

Data repair is a common and crucial step in many frameworks today, as ap...

Autonomous Self-Explanation of Behavior for Interactive Reinforcement Learning Agents

In cooperation, the workers must know how co-workers behave. However, an...