Optimal Skipping Rates: Training Agents with Fine-Grained Control Using Deep Reinforcement Learning

These days game AI is one of the focused and active research areas in artificial intelligence because computer games are the best test-beds for testing theoretical ideas in AI before practically applying them in real life world. Similarly, ViZDoom is a game artificial intelligence research platform...

Full description

Saved in:
Bibliographic Details
Main Authors: Adil Khan, Jiang Feng, Shaohui Liu, Muhammad Zubair Asghar
Format: Article
Language:English
Published: Wiley 2019-01-01
Series:Journal of Robotics
Online Access:http://dx.doi.org/10.1155/2019/2970408
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1832551409407492096
author Adil Khan
Jiang Feng
Shaohui Liu
Muhammad Zubair Asghar
author_facet Adil Khan
Jiang Feng
Shaohui Liu
Muhammad Zubair Asghar
author_sort Adil Khan
collection DOAJ
description These days game AI is one of the focused and active research areas in artificial intelligence because computer games are the best test-beds for testing theoretical ideas in AI before practically applying them in real life world. Similarly, ViZDoom is a game artificial intelligence research platform based on Doom used for visual deep reinforcement learning in 3D game environments such as first-person shooters (FPS). While training, the speed of the learning agent greatly depends on the number of frames the agent is permitted to skip. In this paper, how the frame skipping rate influences the agent’s learning and final performance is proposed, particularly using deep Q-learning, experience replay memory, and the ViZDoom Game AI research platform. The agent is trained and tested on Doom’s basic scenario(s) where the results are compared and found to be 10% better compared to the existing state-of-the-art research work on Doom-based agents. The experiments show that the profitable and optimal frame skipping rate falls in the range of 3 to 11 that provides the best balance between the learning speed and the final performance of the agent which exhibits human-like behavior and outperforms an average human player and inbuilt game agents.
format Article
id doaj-art-979434ff503f4520953709d25f223aea
institution Kabale University
issn 1687-9600
1687-9619
language English
publishDate 2019-01-01
publisher Wiley
record_format Article
series Journal of Robotics
spelling doaj-art-979434ff503f4520953709d25f223aea2025-02-03T06:01:29ZengWileyJournal of Robotics1687-96001687-96192019-01-01201910.1155/2019/29704082970408Optimal Skipping Rates: Training Agents with Fine-Grained Control Using Deep Reinforcement LearningAdil Khan0Jiang Feng1Shaohui Liu2Muhammad Zubair Asghar3School of Computer Science and Technology, Harbin Institute of Technology, NO. 92, Xidazhi Street, Harbin, Heilongjiang 150001, ChinaSchool of Computer Science and Technology, Harbin Institute of Technology, NO. 92, Xidazhi Street, Harbin, Heilongjiang 150001, ChinaSchool of Computer Science and Technology, Harbin Institute of Technology, NO. 92, Xidazhi Street, Harbin, Heilongjiang 150001, ChinaInstitute of Computing and Information Technology, Gomal University, D. I. Khan, KP, PakistanThese days game AI is one of the focused and active research areas in artificial intelligence because computer games are the best test-beds for testing theoretical ideas in AI before practically applying them in real life world. Similarly, ViZDoom is a game artificial intelligence research platform based on Doom used for visual deep reinforcement learning in 3D game environments such as first-person shooters (FPS). While training, the speed of the learning agent greatly depends on the number of frames the agent is permitted to skip. In this paper, how the frame skipping rate influences the agent’s learning and final performance is proposed, particularly using deep Q-learning, experience replay memory, and the ViZDoom Game AI research platform. The agent is trained and tested on Doom’s basic scenario(s) where the results are compared and found to be 10% better compared to the existing state-of-the-art research work on Doom-based agents. The experiments show that the profitable and optimal frame skipping rate falls in the range of 3 to 11 that provides the best balance between the learning speed and the final performance of the agent which exhibits human-like behavior and outperforms an average human player and inbuilt game agents.http://dx.doi.org/10.1155/2019/2970408
spellingShingle Adil Khan
Jiang Feng
Shaohui Liu
Muhammad Zubair Asghar
Optimal Skipping Rates: Training Agents with Fine-Grained Control Using Deep Reinforcement Learning
Journal of Robotics
title Optimal Skipping Rates: Training Agents with Fine-Grained Control Using Deep Reinforcement Learning
title_full Optimal Skipping Rates: Training Agents with Fine-Grained Control Using Deep Reinforcement Learning
title_fullStr Optimal Skipping Rates: Training Agents with Fine-Grained Control Using Deep Reinforcement Learning
title_full_unstemmed Optimal Skipping Rates: Training Agents with Fine-Grained Control Using Deep Reinforcement Learning
title_short Optimal Skipping Rates: Training Agents with Fine-Grained Control Using Deep Reinforcement Learning
title_sort optimal skipping rates training agents with fine grained control using deep reinforcement learning
url http://dx.doi.org/10.1155/2019/2970408
work_keys_str_mv AT adilkhan optimalskippingratestrainingagentswithfinegrainedcontrolusingdeepreinforcementlearning
AT jiangfeng optimalskippingratestrainingagentswithfinegrainedcontrolusingdeepreinforcementlearning
AT shaohuiliu optimalskippingratestrainingagentswithfinegrainedcontrolusingdeepreinforcementlearning
AT muhammadzubairasghar optimalskippingratestrainingagentswithfinegrainedcontrolusingdeepreinforcementlearning