Landmark-based approaches for goal recognition as planning

Ramon Fraga Pereira*, Nir Oren, Felipe Meneguzzi

*Corresponding author for this work

Research output: Contribution to journalArticle

Abstract

Recognizing goals and plans from complete or partial observations can be efficiently achieved through automated planning techniques. In many applications, it is important to recognize goals and plans not only accurately, but also quickly. To address this challenge, we develop novel goal recognition approaches based on planning techniques that rely on planning landmarks. In automated planning, landmarks are properties (or actions) that cannot be avoided to achieve a goal. We show the applicability of a number of planning techniques with an emphasis on landmarks for goal recognition tasks in two settings: (1) we use the concept of landmarks to develop goal recognition heuristics; and (2) we develop a landmark-based filtering method to refine existing planning-based goal and plan recognition approaches. These recognition approaches are empirically evaluated in experiments over several classical planning domains. We show that our goal recognition approaches yield not only accuracy comparable to (and often higher than) other state-of-the-art techniques, but also result in substantially faster recognition time over existing techniques.
Original languageEnglish
Article number103217
JournalArtificial Intelligence
Volume279
Early online date4 Dec 2019
DOIs
Publication statusE-pub ahead of print - 4 Dec 2019

Fingerprint

planning procedure
Planning
planning
heuristics
Landmarks
experiment
Experiments

Keywords

  • goal recognition
  • AI planning
  • landmarks

Cite this

Landmark-based approaches for goal recognition as planning. / Pereira, Ramon Fraga; Oren, Nir; Meneguzzi, Felipe.

In: Artificial Intelligence, Vol. 279, 103217, 02.2020.

Research output: Contribution to journalArticle

Pereira, Ramon Fraga ; Oren, Nir ; Meneguzzi, Felipe. / Landmark-based approaches for goal recognition as planning. In: Artificial Intelligence. 2020 ; Vol. 279.
@article{54c43d8ce6414738a5a4affd942c8d92,
title = "Landmark-based approaches for goal recognition as planning",
abstract = "Recognizing goals and plans from complete or partial observations can be efficiently achieved through automated planning techniques. In many applications, it is important to recognize goals and plans not only accurately, but also quickly. To address this challenge, we develop novel goal recognition approaches based on planning techniques that rely on planning landmarks. In automated planning, landmarks are properties (or actions) that cannot be avoided to achieve a goal. We show the applicability of a number of planning techniques with an emphasis on landmarks for goal recognition tasks in two settings: (1) we use the concept of landmarks to develop goal recognition heuristics; and (2) we develop a landmark-based filtering method to refine existing planning-based goal and plan recognition approaches. These recognition approaches are empirically evaluated in experiments over several classical planning domains. We show that our goal recognition approaches yield not only accuracy comparable to (and often higher than) other state-of-the-art techniques, but also result in substantially faster recognition time over existing techniques.",
keywords = "goal recognition, AI planning, landmarks",
author = "Pereira, {Ramon Fraga} and Nir Oren and Felipe Meneguzzi",
note = "This article is a revised and extended version of two papers published at AAAI 2017 (Pereira et al., 2017b) and ECAI 2016 (Pereira and Meneguzzi, 2016). We thank the anonymous reviewers that helped improve the research in this article. The authors thank Shirin Sohrabi for discussing the way in which the algorithms of Sohrabi et al. (2016) should be configured, and Yolanda Escudero-Martın for providing code for the approach of E-Martın et al. (2015) and engaging with us. We also thank Miquel Ramırez and Mor Vered for various discussions, and Andre Grahl Pereira for a discussion of properties of our algorithm. Felipe thanks CNPq for partial financial support under its PQ fellowship, grant number 305969/2016-1.",
year = "2019",
month = "12",
day = "4",
doi = "10.1016/j.artint.2019.103217",
language = "English",
volume = "279",
journal = "Artificial Intelligence",
issn = "0004-3702",
publisher = "Elsevier",

}

TY - JOUR

T1 - Landmark-based approaches for goal recognition as planning

AU - Pereira, Ramon Fraga

AU - Oren, Nir

AU - Meneguzzi, Felipe

N1 - This article is a revised and extended version of two papers published at AAAI 2017 (Pereira et al., 2017b) and ECAI 2016 (Pereira and Meneguzzi, 2016). We thank the anonymous reviewers that helped improve the research in this article. The authors thank Shirin Sohrabi for discussing the way in which the algorithms of Sohrabi et al. (2016) should be configured, and Yolanda Escudero-Martın for providing code for the approach of E-Martın et al. (2015) and engaging with us. We also thank Miquel Ramırez and Mor Vered for various discussions, and Andre Grahl Pereira for a discussion of properties of our algorithm. Felipe thanks CNPq for partial financial support under its PQ fellowship, grant number 305969/2016-1.

PY - 2019/12/4

Y1 - 2019/12/4

N2 - Recognizing goals and plans from complete or partial observations can be efficiently achieved through automated planning techniques. In many applications, it is important to recognize goals and plans not only accurately, but also quickly. To address this challenge, we develop novel goal recognition approaches based on planning techniques that rely on planning landmarks. In automated planning, landmarks are properties (or actions) that cannot be avoided to achieve a goal. We show the applicability of a number of planning techniques with an emphasis on landmarks for goal recognition tasks in two settings: (1) we use the concept of landmarks to develop goal recognition heuristics; and (2) we develop a landmark-based filtering method to refine existing planning-based goal and plan recognition approaches. These recognition approaches are empirically evaluated in experiments over several classical planning domains. We show that our goal recognition approaches yield not only accuracy comparable to (and often higher than) other state-of-the-art techniques, but also result in substantially faster recognition time over existing techniques.

AB - Recognizing goals and plans from complete or partial observations can be efficiently achieved through automated planning techniques. In many applications, it is important to recognize goals and plans not only accurately, but also quickly. To address this challenge, we develop novel goal recognition approaches based on planning techniques that rely on planning landmarks. In automated planning, landmarks are properties (or actions) that cannot be avoided to achieve a goal. We show the applicability of a number of planning techniques with an emphasis on landmarks for goal recognition tasks in two settings: (1) we use the concept of landmarks to develop goal recognition heuristics; and (2) we develop a landmark-based filtering method to refine existing planning-based goal and plan recognition approaches. These recognition approaches are empirically evaluated in experiments over several classical planning domains. We show that our goal recognition approaches yield not only accuracy comparable to (and often higher than) other state-of-the-art techniques, but also result in substantially faster recognition time over existing techniques.

KW - goal recognition

KW - AI planning

KW - landmarks

U2 - 10.1016/j.artint.2019.103217

DO - 10.1016/j.artint.2019.103217

M3 - Article

VL - 279

JO - Artificial Intelligence

JF - Artificial Intelligence

SN - 0004-3702

M1 - 103217

ER -