Plan explanations as model reconciliation: Moving beyond explanation as soliloquy

Tathagata Chakraborti, Sarath Sreedharan, Yu Zhang, Subbarao Kambhampati

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

28 Citations (Scopus)

Abstract

When AI systems interact with humans in the loop, they are often called on to provide explanations for their plans and behavior. Past work on plan explanations primarily involved the AI system explaining the correctness of its plan and the rationale for its decision in terms of its own model. Such soliloquy is wholly inadequate in most realistic scenarios where the humans have domain and task models that differ significantly from that used by the AI system. We posit that the explanations are best studied in light of these differing models. In particular, we show how explanation can be seen as a "model reconciliation problem" (MRP), where the AI system in effect suggests changes to the human's model, so as to make its plan be optimal with respect to that changed human model. We will study the properties of such explanations, present algorithms for automatically computing them, and evaluate the performance of the algorithms.

Original languageEnglish (US)
Title of host publication26th International Joint Conference on Artificial Intelligence, IJCAI 2017
PublisherInternational Joint Conferences on Artificial Intelligence
Pages156-163
Number of pages8
ISBN (Electronic)9780999241103
StatePublished - 2017
Event26th International Joint Conference on Artificial Intelligence, IJCAI 2017 - Melbourne, Australia
Duration: Aug 19 2017Aug 25 2017

Other

Other26th International Joint Conference on Artificial Intelligence, IJCAI 2017
CountryAustralia
CityMelbourne
Period8/19/178/25/17

ASJC Scopus subject areas

  • Artificial Intelligence

Cite this

Chakraborti, T., Sreedharan, S., Zhang, Y., & Kambhampati, S. (2017). Plan explanations as model reconciliation: Moving beyond explanation as soliloquy. In 26th International Joint Conference on Artificial Intelligence, IJCAI 2017 (pp. 156-163). International Joint Conferences on Artificial Intelligence.

Plan explanations as model reconciliation : Moving beyond explanation as soliloquy. / Chakraborti, Tathagata; Sreedharan, Sarath; Zhang, Yu; Kambhampati, Subbarao.

26th International Joint Conference on Artificial Intelligence, IJCAI 2017. International Joint Conferences on Artificial Intelligence, 2017. p. 156-163.

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

Chakraborti, T, Sreedharan, S, Zhang, Y & Kambhampati, S 2017, Plan explanations as model reconciliation: Moving beyond explanation as soliloquy. in 26th International Joint Conference on Artificial Intelligence, IJCAI 2017. International Joint Conferences on Artificial Intelligence, pp. 156-163, 26th International Joint Conference on Artificial Intelligence, IJCAI 2017, Melbourne, Australia, 8/19/17.
Chakraborti T, Sreedharan S, Zhang Y, Kambhampati S. Plan explanations as model reconciliation: Moving beyond explanation as soliloquy. In 26th International Joint Conference on Artificial Intelligence, IJCAI 2017. International Joint Conferences on Artificial Intelligence. 2017. p. 156-163
Chakraborti, Tathagata ; Sreedharan, Sarath ; Zhang, Yu ; Kambhampati, Subbarao. / Plan explanations as model reconciliation : Moving beyond explanation as soliloquy. 26th International Joint Conference on Artificial Intelligence, IJCAI 2017. International Joint Conferences on Artificial Intelligence, 2017. pp. 156-163
@inproceedings{19bb37bfb6004808b400b187e17e413d,
title = "Plan explanations as model reconciliation: Moving beyond explanation as soliloquy",
abstract = "When AI systems interact with humans in the loop, they are often called on to provide explanations for their plans and behavior. Past work on plan explanations primarily involved the AI system explaining the correctness of its plan and the rationale for its decision in terms of its own model. Such soliloquy is wholly inadequate in most realistic scenarios where the humans have domain and task models that differ significantly from that used by the AI system. We posit that the explanations are best studied in light of these differing models. In particular, we show how explanation can be seen as a {"}model reconciliation problem{"} (MRP), where the AI system in effect suggests changes to the human's model, so as to make its plan be optimal with respect to that changed human model. We will study the properties of such explanations, present algorithms for automatically computing them, and evaluate the performance of the algorithms.",
author = "Tathagata Chakraborti and Sarath Sreedharan and Yu Zhang and Subbarao Kambhampati",
year = "2017",
language = "English (US)",
pages = "156--163",
booktitle = "26th International Joint Conference on Artificial Intelligence, IJCAI 2017",
publisher = "International Joint Conferences on Artificial Intelligence",

}

TY - GEN

T1 - Plan explanations as model reconciliation

T2 - Moving beyond explanation as soliloquy

AU - Chakraborti, Tathagata

AU - Sreedharan, Sarath

AU - Zhang, Yu

AU - Kambhampati, Subbarao

PY - 2017

Y1 - 2017

N2 - When AI systems interact with humans in the loop, they are often called on to provide explanations for their plans and behavior. Past work on plan explanations primarily involved the AI system explaining the correctness of its plan and the rationale for its decision in terms of its own model. Such soliloquy is wholly inadequate in most realistic scenarios where the humans have domain and task models that differ significantly from that used by the AI system. We posit that the explanations are best studied in light of these differing models. In particular, we show how explanation can be seen as a "model reconciliation problem" (MRP), where the AI system in effect suggests changes to the human's model, so as to make its plan be optimal with respect to that changed human model. We will study the properties of such explanations, present algorithms for automatically computing them, and evaluate the performance of the algorithms.

AB - When AI systems interact with humans in the loop, they are often called on to provide explanations for their plans and behavior. Past work on plan explanations primarily involved the AI system explaining the correctness of its plan and the rationale for its decision in terms of its own model. Such soliloquy is wholly inadequate in most realistic scenarios where the humans have domain and task models that differ significantly from that used by the AI system. We posit that the explanations are best studied in light of these differing models. In particular, we show how explanation can be seen as a "model reconciliation problem" (MRP), where the AI system in effect suggests changes to the human's model, so as to make its plan be optimal with respect to that changed human model. We will study the properties of such explanations, present algorithms for automatically computing them, and evaluate the performance of the algorithms.

UR - http://www.scopus.com/inward/record.url?scp=85031946471&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85031946471&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:85031946471

SP - 156

EP - 163

BT - 26th International Joint Conference on Artificial Intelligence, IJCAI 2017

PB - International Joint Conferences on Artificial Intelligence

ER -