Towards Providing Explanations for AI Planner Decisions

Rita Borgo; Michael Cashmore; Daniele Magazzeni

Towards Providing Explanations for AI Planner Decisions

Rita Borgo, Michael Cashmore, Daniele Magazzeni

Computer And Information Sciences

Research output: Contribution to conference › Paper › peer-review

197 Downloads (Pure)

Abstract

In order to engender trust in AI, humans must understand what an AI system is trying to achieve, and why. To overcome this problem, the underlying AI process must produce justifications and explanations that are both transparent and comprehensible to the user. AI Planning is well placed to be able to address this challenge. In this paper we present a methodology to provide initial explanations for the decisions made by the planner. Explanations are created by allowing the user to suggest alternative actions in plans and then compare the resulting plans with the one found by the planner. The methodology is implemented in the new XAI-Plan framework.

Original language	English
Pages	11-17
Number of pages	7
Publication status	Published - 13 Jul 2018
Event	IJCAI/ECAI 2018 Workshop on Explainable Artificial Intelligence (XAI) - Stockholmsmässan , Stockholm, Sweden Duration: 13 Jul 2018 → 13 Jul 2018

Conference

Conference	IJCAI/ECAI 2018 Workshop on Explainable Artificial Intelligence (XAI)
Country/Territory	Sweden
City	Stockholm
Period	13/07/18 → 13/07/18

Keywords

contingency planning
artifical intelligence
AI
explainable planning

Access to Document

Borgo-etal-XAI-2018-Towards-providing-explanations-for-AI-planner-decisionsAccepted author manuscript, 775 KB

Cite this

@conference{cfbf1df33e724c6eb455a784b5cebd73,

title = "Towards Providing Explanations for AI Planner Decisions",

abstract = "In order to engender trust in AI, humans must understand what an AI system is trying to achieve, and why. To overcome this problem, the underlying AI process must produce justifications and explanations that are both transparent and comprehensible to the user. AI Planning is well placed to be able to address this challenge. In this paper we present a methodology to provide initial explanations for the decisions made by the planner. Explanations are created by allowing the user to suggest alternative actions in plans and then compare the resulting plans with the one found by the planner. The methodology is implemented in the new XAI-Plan framework. ",

keywords = "contingency planning, artifical intelligence, AI, explainable planning",

author = "Rita Borgo and Michael Cashmore and Daniele Magazzeni",

note = "Presented at the IJCAI/ECAI 2018 Workshop on Explainable Artificial Intelligence (XAI). Stockholm, July 2018; IJCAI/ECAI 2018 Workshop on Explainable Artificial Intelligence (XAI) ; Conference date: 13-07-2018 Through 13-07-2018",

year = "2018",

month = jul,

day = "13",

language = "English",

pages = "11--17",

}

TY - CONF

T1 - Towards Providing Explanations for AI Planner Decisions

AU - Borgo, Rita

AU - Cashmore, Michael

AU - Magazzeni, Daniele

N1 - Presented at the IJCAI/ECAI 2018 Workshop on Explainable Artificial Intelligence (XAI). Stockholm, July 2018

PY - 2018/7/13

Y1 - 2018/7/13

N2 - In order to engender trust in AI, humans must understand what an AI system is trying to achieve, and why. To overcome this problem, the underlying AI process must produce justifications and explanations that are both transparent and comprehensible to the user. AI Planning is well placed to be able to address this challenge. In this paper we present a methodology to provide initial explanations for the decisions made by the planner. Explanations are created by allowing the user to suggest alternative actions in plans and then compare the resulting plans with the one found by the planner. The methodology is implemented in the new XAI-Plan framework.

AB - In order to engender trust in AI, humans must understand what an AI system is trying to achieve, and why. To overcome this problem, the underlying AI process must produce justifications and explanations that are both transparent and comprehensible to the user. AI Planning is well placed to be able to address this challenge. In this paper we present a methodology to provide initial explanations for the decisions made by the planner. Explanations are created by allowing the user to suggest alternative actions in plans and then compare the resulting plans with the one found by the planner. The methodology is implemented in the new XAI-Plan framework.

KW - contingency planning

KW - artifical intelligence

KW - AI

KW - explainable planning

M3 - Paper

SP - 11

EP - 17

T2 - IJCAI/ECAI 2018 Workshop on Explainable Artificial Intelligence (XAI)

Y2 - 13 July 2018 through 13 July 2018

ER -

Towards Providing Explanations for AI Planner Decisions

Abstract

Conference

Keywords

Access to Document

Fingerprint

Cite this