An Open-World Novelty Generator for Authoring Reinforcement Learning Environment of Standardized Toolkits

Sangho Lee; Junbeom Park; Ho Suk; Taewoo Kim; Pamul Yadav; Shiho Kim

doi:10.1007/978-3-030-80253-0_3

An Open-World Novelty Generator for Authoring Reinforcement Learning Environment of Standardized Toolkits

Sangho Lee, Junbeom Park, Ho Suk, Taewoo Kim, Pamul Yadav, Shiho Kim

School of Integrated Technology

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

6 Citations (Scopus)

Abstract

Current research in Reinforcement Learning (RL) is based on closed-world learning environment where the environment remains fixed and unchanged throughout the agent’s training and application session. The fixed environment may be prone to failure when the agents incorporate under novel unseen situations. To overcome the drawback of the existing closed-world model, an Open-world learning model is required which can classify the novelty occurring in an environment in a hierarchical manner. The proposed control suite with open world novelty generator is an attempt to augment the machine learning environment for authoring the novelty in actors, interactions, and environment of standardized Reinforcement learning toolkits such as UnityML, OpenAI Gym, and DeepMind Control Suite in real-time. Such a tool will provide an opportunity to the RL researchers to simulate the Open-world learning model and test their algorithms within the standardized closed-world learning environments of the standardized RL toolkits.

Original language	English
Title of host publication	Multi-disciplinary Trends in Artificial Intelligence - 14th International Conference, MIWAI 2021, Proceedings
Editors	Phatthanaphong Chomphuwiset, Junmo Kim, Pornntiwa Pawara
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	27-33
Number of pages	7
ISBN (Print)	9783030802523
DOIs	https://doi.org/10.1007/978-3-030-80253-0_3
Publication status	Published - 2021
Event	14th International Conference on Multi-disciplinary Trends in Artificial Intelligence, MIWAI 2021 - Virtual, Online Duration: 2021 Jul 2 → 2021 Jul 3

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	12832 LNAI
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	14th International Conference on Multi-disciplinary Trends in Artificial Intelligence, MIWAI 2021
City	Virtual, Online
Period	21/7/2 → 21/7/3

Bibliographical note

Publisher Copyright:
© 2021, Springer Nature Switzerland AG.

All Science Journal Classification (ASJC) codes

Theoretical Computer Science
General Computer Science

Access to Document

10.1007/978-3-030-80253-0_3

Cite this

Lee, S., Park, J., Suk, H., Kim, T., Yadav, P., & Kim, S. (2021). An Open-World Novelty Generator for Authoring Reinforcement Learning Environment of Standardized Toolkits. In P. Chomphuwiset, J. Kim, & P. Pawara (Eds.), Multi-disciplinary Trends in Artificial Intelligence - 14th International Conference, MIWAI 2021, Proceedings (pp. 27-33). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 12832 LNAI). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-030-80253-0_3

Lee, Sangho ; Park, Junbeom ; Suk, Ho et al. / An Open-World Novelty Generator for Authoring Reinforcement Learning Environment of Standardized Toolkits. Multi-disciplinary Trends in Artificial Intelligence - 14th International Conference, MIWAI 2021, Proceedings. editor / Phatthanaphong Chomphuwiset ; Junmo Kim ; Pornntiwa Pawara. Springer Science and Business Media Deutschland GmbH, 2021. pp. 27-33 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{59a0033738e04729b7b4ef6da2b0bbe4,

title = "An Open-World Novelty Generator for Authoring Reinforcement Learning Environment of Standardized Toolkits",

abstract = "Current research in Reinforcement Learning (RL) is based on closed-world learning environment where the environment remains fixed and unchanged throughout the agent{\textquoteright}s training and application session. The fixed environment may be prone to failure when the agents incorporate under novel unseen situations. To overcome the drawback of the existing closed-world model, an Open-world learning model is required which can classify the novelty occurring in an environment in a hierarchical manner. The proposed control suite with open world novelty generator is an attempt to augment the machine learning environment for authoring the novelty in actors, interactions, and environment of standardized Reinforcement learning toolkits such as UnityML, OpenAI Gym, and DeepMind Control Suite in real-time. Such a tool will provide an opportunity to the RL researchers to simulate the Open-world learning model and test their algorithms within the standardized closed-world learning environments of the standardized RL toolkits.",

keywords = "Control suite, Novelty-generator, Open-world learning, RL toolkits, Reinforcement learning",

author = "Sangho Lee and Junbeom Park and Ho Suk and Taewoo Kim and Pamul Yadav and Shiho Kim",

note = "Publisher Copyright: {\textcopyright} 2021, Springer Nature Switzerland AG.; 14th International Conference on Multi-disciplinary Trends in Artificial Intelligence, MIWAI 2021 ; Conference date: 02-07-2021 Through 03-07-2021",

year = "2021",

doi = "10.1007/978-3-030-80253-0_3",

language = "English",

isbn = "9783030802523",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Science and Business Media Deutschland GmbH",

pages = "27--33",

editor = "Phatthanaphong Chomphuwiset and Junmo Kim and Pornntiwa Pawara",

booktitle = "Multi-disciplinary Trends in Artificial Intelligence - 14th International Conference, MIWAI 2021, Proceedings",

address = "Germany",

}

Lee, S, Park, J, Suk, H, Kim, T, Yadav, P & Kim, S 2021, An Open-World Novelty Generator for Authoring Reinforcement Learning Environment of Standardized Toolkits. in P Chomphuwiset, J Kim & P Pawara (eds), Multi-disciplinary Trends in Artificial Intelligence - 14th International Conference, MIWAI 2021, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 12832 LNAI, Springer Science and Business Media Deutschland GmbH, pp. 27-33, 14th International Conference on Multi-disciplinary Trends in Artificial Intelligence, MIWAI 2021, Virtual, Online, 21/7/2. https://doi.org/10.1007/978-3-030-80253-0_3

An Open-World Novelty Generator for Authoring Reinforcement Learning Environment of Standardized Toolkits. / Lee, Sangho; Park, Junbeom; Suk, Ho et al.
Multi-disciplinary Trends in Artificial Intelligence - 14th International Conference, MIWAI 2021, Proceedings. ed. / Phatthanaphong Chomphuwiset; Junmo Kim; Pornntiwa Pawara. Springer Science and Business Media Deutschland GmbH, 2021. p. 27-33 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 12832 LNAI).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - An Open-World Novelty Generator for Authoring Reinforcement Learning Environment of Standardized Toolkits

AU - Lee, Sangho

AU - Park, Junbeom

AU - Suk, Ho

AU - Kim, Taewoo

AU - Yadav, Pamul

AU - Kim, Shiho

PY - 2021

Y1 - 2021

N2 - Current research in Reinforcement Learning (RL) is based on closed-world learning environment where the environment remains fixed and unchanged throughout the agent’s training and application session. The fixed environment may be prone to failure when the agents incorporate under novel unseen situations. To overcome the drawback of the existing closed-world model, an Open-world learning model is required which can classify the novelty occurring in an environment in a hierarchical manner. The proposed control suite with open world novelty generator is an attempt to augment the machine learning environment for authoring the novelty in actors, interactions, and environment of standardized Reinforcement learning toolkits such as UnityML, OpenAI Gym, and DeepMind Control Suite in real-time. Such a tool will provide an opportunity to the RL researchers to simulate the Open-world learning model and test their algorithms within the standardized closed-world learning environments of the standardized RL toolkits.

AB - Current research in Reinforcement Learning (RL) is based on closed-world learning environment where the environment remains fixed and unchanged throughout the agent’s training and application session. The fixed environment may be prone to failure when the agents incorporate under novel unseen situations. To overcome the drawback of the existing closed-world model, an Open-world learning model is required which can classify the novelty occurring in an environment in a hierarchical manner. The proposed control suite with open world novelty generator is an attempt to augment the machine learning environment for authoring the novelty in actors, interactions, and environment of standardized Reinforcement learning toolkits such as UnityML, OpenAI Gym, and DeepMind Control Suite in real-time. Such a tool will provide an opportunity to the RL researchers to simulate the Open-world learning model and test their algorithms within the standardized closed-world learning environments of the standardized RL toolkits.

KW - Control suite

KW - Novelty-generator

KW - Open-world learning

KW - RL toolkits

KW - Reinforcement learning

UR - http://www.scopus.com/inward/record.url?scp=85111966969&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85111966969&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-80253-0_3

DO - 10.1007/978-3-030-80253-0_3

M3 - Conference contribution

AN - SCOPUS:85111966969

SN - 9783030802523

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 27

EP - 33

BT - Multi-disciplinary Trends in Artificial Intelligence - 14th International Conference, MIWAI 2021, Proceedings

A2 - Chomphuwiset, Phatthanaphong

A2 - Kim, Junmo

A2 - Pawara, Pornntiwa

PB - Springer Science and Business Media Deutschland GmbH

T2 - 14th International Conference on Multi-disciplinary Trends in Artificial Intelligence, MIWAI 2021

Y2 - 2 July 2021 through 3 July 2021

ER -

Lee S, Park J, Suk H, Kim T, Yadav P, Kim S. An Open-World Novelty Generator for Authoring Reinforcement Learning Environment of Standardized Toolkits. In Chomphuwiset P, Kim J, Pawara P, editors, Multi-disciplinary Trends in Artificial Intelligence - 14th International Conference, MIWAI 2021, Proceedings. Springer Science and Business Media Deutschland GmbH. 2021. p. 27-33. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-030-80253-0_3

An Open-World Novelty Generator for Authoring Reinforcement Learning Environment of Standardized Toolkits

Abstract

Publication series

Conference

Bibliographical note

All Science Journal Classification (ASJC) codes

Access to Document

Other files and links

Fingerprint

Cite this