Action-model based multi-agent plan recognition

Hankz Hankui Zhuo, Qiang Yang, Subbarao Kambhampati

Research output: Chapter in Book/Report/Conference proceedingConference contribution

23 Citations (Scopus)

Abstract

Multi-Agent Plan Recognition (MAPR) aims to recognize dynamic team structures and team behaviors from the observed team traces (activity sequences) of a set of intelligent agents. Previous MAPR approaches required a library of team activity sequences (team plans) be given as input. However, collecting a library of team plans to ensure adequate coverage is often difficult and costly. In this paper, we relax this constraint, so that team plans are not required to be provided beforehand. We assume instead that a set of action models are available. Such models are often already created to describe domain physics; i.e., the preconditions and effects of effects actions. We propose a novel approach for recognizing multi-agent team plans based on such action models rather than libraries of team plans. We encode the resulting MAPR problem as a satisfiability problem and solve the problem using a state-of-the-art weighted MAX-SAT solver. Our approach also allows for incompleteness in the observed plan traces. Our empirical studies demonstrate that our algorithm is both effective and efficient in comparison to state-of-the-art MAPR methods based on plan libraries.

Original languageEnglish (US)
Title of host publicationAdvances in Neural Information Processing Systems
Pages368-376
Number of pages9
Volume1
StatePublished - 2012
Event26th Annual Conference on Neural Information Processing Systems 2012, NIPS 2012 - Lake Tahoe, NV, United States
Duration: Dec 3 2012Dec 6 2012

Other

Other26th Annual Conference on Neural Information Processing Systems 2012, NIPS 2012
CountryUnited States
CityLake Tahoe, NV
Period12/3/1212/6/12

Fingerprint

Intelligent agents
Physics

ASJC Scopus subject areas

  • Computer Networks and Communications
  • Information Systems
  • Signal Processing

Cite this

Zhuo, H. H., Yang, Q., & Kambhampati, S. (2012). Action-model based multi-agent plan recognition. In Advances in Neural Information Processing Systems (Vol. 1, pp. 368-376)

Action-model based multi-agent plan recognition. / Zhuo, Hankz Hankui; Yang, Qiang; Kambhampati, Subbarao.

Advances in Neural Information Processing Systems. Vol. 1 2012. p. 368-376.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Zhuo, HH, Yang, Q & Kambhampati, S 2012, Action-model based multi-agent plan recognition. in Advances in Neural Information Processing Systems. vol. 1, pp. 368-376, 26th Annual Conference on Neural Information Processing Systems 2012, NIPS 2012, Lake Tahoe, NV, United States, 12/3/12.
Zhuo HH, Yang Q, Kambhampati S. Action-model based multi-agent plan recognition. In Advances in Neural Information Processing Systems. Vol. 1. 2012. p. 368-376
Zhuo, Hankz Hankui ; Yang, Qiang ; Kambhampati, Subbarao. / Action-model based multi-agent plan recognition. Advances in Neural Information Processing Systems. Vol. 1 2012. pp. 368-376
@inproceedings{83da3486b77d469b868aeb1c1b7befab,
title = "Action-model based multi-agent plan recognition",
abstract = "Multi-Agent Plan Recognition (MAPR) aims to recognize dynamic team structures and team behaviors from the observed team traces (activity sequences) of a set of intelligent agents. Previous MAPR approaches required a library of team activity sequences (team plans) be given as input. However, collecting a library of team plans to ensure adequate coverage is often difficult and costly. In this paper, we relax this constraint, so that team plans are not required to be provided beforehand. We assume instead that a set of action models are available. Such models are often already created to describe domain physics; i.e., the preconditions and effects of effects actions. We propose a novel approach for recognizing multi-agent team plans based on such action models rather than libraries of team plans. We encode the resulting MAPR problem as a satisfiability problem and solve the problem using a state-of-the-art weighted MAX-SAT solver. Our approach also allows for incompleteness in the observed plan traces. Our empirical studies demonstrate that our algorithm is both effective and efficient in comparison to state-of-the-art MAPR methods based on plan libraries.",
author = "Zhuo, {Hankz Hankui} and Qiang Yang and Subbarao Kambhampati",
year = "2012",
language = "English (US)",
isbn = "9781627480031",
volume = "1",
pages = "368--376",
booktitle = "Advances in Neural Information Processing Systems",

}

TY - GEN

T1 - Action-model based multi-agent plan recognition

AU - Zhuo, Hankz Hankui

AU - Yang, Qiang

AU - Kambhampati, Subbarao

PY - 2012

Y1 - 2012

N2 - Multi-Agent Plan Recognition (MAPR) aims to recognize dynamic team structures and team behaviors from the observed team traces (activity sequences) of a set of intelligent agents. Previous MAPR approaches required a library of team activity sequences (team plans) be given as input. However, collecting a library of team plans to ensure adequate coverage is often difficult and costly. In this paper, we relax this constraint, so that team plans are not required to be provided beforehand. We assume instead that a set of action models are available. Such models are often already created to describe domain physics; i.e., the preconditions and effects of effects actions. We propose a novel approach for recognizing multi-agent team plans based on such action models rather than libraries of team plans. We encode the resulting MAPR problem as a satisfiability problem and solve the problem using a state-of-the-art weighted MAX-SAT solver. Our approach also allows for incompleteness in the observed plan traces. Our empirical studies demonstrate that our algorithm is both effective and efficient in comparison to state-of-the-art MAPR methods based on plan libraries.

AB - Multi-Agent Plan Recognition (MAPR) aims to recognize dynamic team structures and team behaviors from the observed team traces (activity sequences) of a set of intelligent agents. Previous MAPR approaches required a library of team activity sequences (team plans) be given as input. However, collecting a library of team plans to ensure adequate coverage is often difficult and costly. In this paper, we relax this constraint, so that team plans are not required to be provided beforehand. We assume instead that a set of action models are available. Such models are often already created to describe domain physics; i.e., the preconditions and effects of effects actions. We propose a novel approach for recognizing multi-agent team plans based on such action models rather than libraries of team plans. We encode the resulting MAPR problem as a satisfiability problem and solve the problem using a state-of-the-art weighted MAX-SAT solver. Our approach also allows for incompleteness in the observed plan traces. Our empirical studies demonstrate that our algorithm is both effective and efficient in comparison to state-of-the-art MAPR methods based on plan libraries.

UR - http://www.scopus.com/inward/record.url?scp=84877754763&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84877754763&partnerID=8YFLogxK

M3 - Conference contribution

SN - 9781627480031

VL - 1

SP - 368

EP - 376

BT - Advances in Neural Information Processing Systems

ER -