Learning Spatial Search using Submodular Inverse Reinforcement Learning

Ji Jie Wu; Kuo Shih Tseng

doi:10.1109/SSRR50563.2020.9292603

Learning Spatial Search using Submodular Inverse Reinforcement Learning

Ji Jie Wu, Kuo Shih Tseng

Department of Mathematics

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

1 Scopus citations

Abstract

Finding an optimal search path is a NP-hard problem. Since search is one of human central activities, learning spatial search behavior from human operators is a way to solve search problems. Utilizing the submodularity of search problems, this research proposes a submodular inverse reinforcement learning (SIRL) algorithm to learn humans' search behavior. The experiments demonstrate that the performance of the learned search paths outperform that of state of the art approaches (e.g., MaxEnt IRL and DIRL).

Original language	English
Title of host publication	2020 IEEE International Symposium on Safety, Security, and Rescue Robotics, SSRR 2020
Editors	Lino Marques, Majid Khonji, Jorge Dias
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	7-14
Number of pages	8
ISBN (Electronic)	9781665403900
DOIs	https://doi.org/10.1109/SSRR50563.2020.9292603
State	Published - 4 Nov 2020
Event	2020 IEEE International Symposium on Safety, Security, and Rescue Robotics, SSRR 2020 - Abu Dhabi, United Arab Emirates Duration: 4 Nov 2020 → 6 Nov 2020

Publication series

Name	2020 IEEE International Symposium on Safety, Security, and Rescue Robotics, SSRR 2020

Conference

Conference	2020 IEEE International Symposium on Safety, Security, and Rescue Robotics, SSRR 2020
Country/Territory	United Arab Emirates
City	Abu Dhabi
Period	4/11/20 → 6/11/20

Access to Document

10.1109/SSRR50563.2020.9292603

Deep Inverse Reinforcement Learning for Informative Path Planning(2/3)
Tseng, K.
1/08/20 → 31/07/21
Project: Research
Near-Optimal Search Using Uavs in 3d Environments
Tseng, K.
1/10/18 → 30/09/19
Project: Research

Cite this

Wu, J. J., & Tseng, K. S. (2020). Learning Spatial Search using Submodular Inverse Reinforcement Learning. In L. Marques, M. Khonji, & J. Dias (Eds.), 2020 IEEE International Symposium on Safety, Security, and Rescue Robotics, SSRR 2020 (pp. 7-14). [9292603] (2020 IEEE International Symposium on Safety, Security, and Rescue Robotics, SSRR 2020). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/SSRR50563.2020.9292603

Wu, Ji Jie ; Tseng, Kuo Shih. / Learning Spatial Search using Submodular Inverse Reinforcement Learning. 2020 IEEE International Symposium on Safety, Security, and Rescue Robotics, SSRR 2020. editor / Lino Marques ; Majid Khonji ; Jorge Dias. Institute of Electrical and Electronics Engineers Inc., 2020. pp. 7-14 (2020 IEEE International Symposium on Safety, Security, and Rescue Robotics, SSRR 2020).

@inproceedings{da520fad0ec34c759d571fc181b46131,

title = "Learning Spatial Search using Submodular Inverse Reinforcement Learning",

abstract = "Finding an optimal search path is a NP-hard problem. Since search is one of human central activities, learning spatial search behavior from human operators is a way to solve search problems. Utilizing the submodularity of search problems, this research proposes a submodular inverse reinforcement learning (SIRL) algorithm to learn humans' search behavior. The experiments demonstrate that the performance of the learned search paths outperform that of state of the art approaches (e.g., MaxEnt IRL and DIRL). ",

author = "Wu, {Ji Jie} and Tseng, {Kuo Shih}",

note = "Publisher Copyright: {\textcopyright} 2020 IEEE.; 2020 IEEE International Symposium on Safety, Security, and Rescue Robotics, SSRR 2020 ; Conference date: 04-11-2020 Through 06-11-2020",

year = "2020",

month = nov,

day = "4",

doi = "10.1109/SSRR50563.2020.9292603",

language = "???core.languages.en_GB???",

series = "2020 IEEE International Symposium on Safety, Security, and Rescue Robotics, SSRR 2020",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

pages = "7--14",

editor = "Lino Marques and Majid Khonji and Jorge Dias",

booktitle = "2020 IEEE International Symposium on Safety, Security, and Rescue Robotics, SSRR 2020",

}

Wu, JJ & Tseng, KS 2020, Learning Spatial Search using Submodular Inverse Reinforcement Learning. in L Marques, M Khonji & J Dias (eds), 2020 IEEE International Symposium on Safety, Security, and Rescue Robotics, SSRR 2020., 9292603, 2020 IEEE International Symposium on Safety, Security, and Rescue Robotics, SSRR 2020, Institute of Electrical and Electronics Engineers Inc., pp. 7-14, 2020 IEEE International Symposium on Safety, Security, and Rescue Robotics, SSRR 2020, Abu Dhabi, United Arab Emirates, 4/11/20. https://doi.org/10.1109/SSRR50563.2020.9292603

Learning Spatial Search using Submodular Inverse Reinforcement Learning. / Wu, Ji Jie; Tseng, Kuo Shih.
2020 IEEE International Symposium on Safety, Security, and Rescue Robotics, SSRR 2020. ed. / Lino Marques; Majid Khonji; Jorge Dias. Institute of Electrical and Electronics Engineers Inc., 2020. p. 7-14 9292603 (2020 IEEE International Symposium on Safety, Security, and Rescue Robotics, SSRR 2020).

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

TY - GEN

T1 - Learning Spatial Search using Submodular Inverse Reinforcement Learning

AU - Wu, Ji Jie

AU - Tseng, Kuo Shih

PY - 2020/11/4

Y1 - 2020/11/4

N2 - Finding an optimal search path is a NP-hard problem. Since search is one of human central activities, learning spatial search behavior from human operators is a way to solve search problems. Utilizing the submodularity of search problems, this research proposes a submodular inverse reinforcement learning (SIRL) algorithm to learn humans' search behavior. The experiments demonstrate that the performance of the learned search paths outperform that of state of the art approaches (e.g., MaxEnt IRL and DIRL).

AB - Finding an optimal search path is a NP-hard problem. Since search is one of human central activities, learning spatial search behavior from human operators is a way to solve search problems. Utilizing the submodularity of search problems, this research proposes a submodular inverse reinforcement learning (SIRL) algorithm to learn humans' search behavior. The experiments demonstrate that the performance of the learned search paths outperform that of state of the art approaches (e.g., MaxEnt IRL and DIRL).

UR - http://www.scopus.com/inward/record.url?scp=85099455867&partnerID=8YFLogxK

U2 - 10.1109/SSRR50563.2020.9292603

DO - 10.1109/SSRR50563.2020.9292603

M3 - 會議論文篇章

AN - SCOPUS:85099455867

T3 - 2020 IEEE International Symposium on Safety, Security, and Rescue Robotics, SSRR 2020

SP - 7

EP - 14

BT - 2020 IEEE International Symposium on Safety, Security, and Rescue Robotics, SSRR 2020

A2 - Marques, Lino

A2 - Khonji, Majid

A2 - Dias, Jorge

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 2020 IEEE International Symposium on Safety, Security, and Rescue Robotics, SSRR 2020

Y2 - 4 November 2020 through 6 November 2020

ER -

Wu JJ, Tseng KS. Learning Spatial Search using Submodular Inverse Reinforcement Learning. In Marques L, Khonji M, Dias J, editors, 2020 IEEE International Symposium on Safety, Security, and Rescue Robotics, SSRR 2020. Institute of Electrical and Electronics Engineers Inc. 2020. p. 7-14. 9292603. (2020 IEEE International Symposium on Safety, Security, and Rescue Robotics, SSRR 2020). doi: 10.1109/SSRR50563.2020.9292603

Learning Spatial Search using Submodular Inverse Reinforcement Learning

Abstract

Publication series

Conference

Access to Document

Fingerprint

Projects

Deep Inverse Reinforcement Learning for Informative Path Planning(2/3)

Near-Optimal Search Using Uavs in 3d Environments

Cite this