Learning to Build Qualitative Scenario Models From Natural Language

Maxwell Crouse, Clifton McFate, Kenneth D Forbus

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

Abstract

Agents often have to make decisions with incomplete knowledge and few computational resources. We argue that qualitative representations and reasoning, especially combined with analogy, provide a natural approach to performing decision-making in situations with little data, incomplete models, and under tight computational constraints. Moreover, qualitative models provide a means of recognizing and framing decision problems. This paper describes our progress in exploring these ideas to date, using examples from experiments with a system that learns to play Freeciv, an open-source strategy game.
Original languageEnglish (US)
Title of host publicationProceedings of the QR’18
EditorsZoe Falomir, George M. Coghill, Wei Pang
Pages32-39
Number of pages8
StatePublished - 2018

Fingerprint

Decision making
Experiments

Cite this

Crouse, M., McFate, C., & Forbus, K. D. (2018). Learning to Build Qualitative Scenario Models From Natural Language. In Z. Falomir, G. M. Coghill, & W. Pang (Eds.), Proceedings of the QR’18 (pp. 32-39)
Crouse, Maxwell ; McFate, Clifton ; Forbus, Kenneth D. / Learning to Build Qualitative Scenario Models From Natural Language. Proceedings of the QR’18. editor / Zoe Falomir ; George M. Coghill ; Wei Pang. 2018. pp. 32-39
@inproceedings{7c936e3da0c24f6dbc7f14f5c894d19c,
title = "Learning to Build Qualitative Scenario Models From Natural Language",
abstract = "Agents often have to make decisions with incomplete knowledge and few computational resources. We argue that qualitative representations and reasoning, especially combined with analogy, provide a natural approach to performing decision-making in situations with little data, incomplete models, and under tight computational constraints. Moreover, qualitative models provide a means of recognizing and framing decision problems. This paper describes our progress in exploring these ideas to date, using examples from experiments with a system that learns to play Freeciv, an open-source strategy game.",
author = "Maxwell Crouse and Clifton McFate and Forbus, {Kenneth D}",
note = "https://homepages.abdn.ac.uk/pang.wei/pages/QR2018/QR18-Proceedings.pdf",
year = "2018",
language = "English (US)",
pages = "32--39",
editor = "Zoe Falomir and Coghill, {George M.} and Wei Pang",
booktitle = "Proceedings of the QR’18",

}

Crouse, M, McFate, C & Forbus, KD 2018, Learning to Build Qualitative Scenario Models From Natural Language. in Z Falomir, GM Coghill & W Pang (eds), Proceedings of the QR’18. pp. 32-39.

Learning to Build Qualitative Scenario Models From Natural Language. / Crouse, Maxwell; McFate, Clifton; Forbus, Kenneth D.

Proceedings of the QR’18. ed. / Zoe Falomir; George M. Coghill; Wei Pang. 2018. p. 32-39.

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

TY - GEN

T1 - Learning to Build Qualitative Scenario Models From Natural Language

AU - Crouse, Maxwell

AU - McFate, Clifton

AU - Forbus, Kenneth D

N1 - https://homepages.abdn.ac.uk/pang.wei/pages/QR2018/QR18-Proceedings.pdf

PY - 2018

Y1 - 2018

N2 - Agents often have to make decisions with incomplete knowledge and few computational resources. We argue that qualitative representations and reasoning, especially combined with analogy, provide a natural approach to performing decision-making in situations with little data, incomplete models, and under tight computational constraints. Moreover, qualitative models provide a means of recognizing and framing decision problems. This paper describes our progress in exploring these ideas to date, using examples from experiments with a system that learns to play Freeciv, an open-source strategy game.

AB - Agents often have to make decisions with incomplete knowledge and few computational resources. We argue that qualitative representations and reasoning, especially combined with analogy, provide a natural approach to performing decision-making in situations with little data, incomplete models, and under tight computational constraints. Moreover, qualitative models provide a means of recognizing and framing decision problems. This paper describes our progress in exploring these ideas to date, using examples from experiments with a system that learns to play Freeciv, an open-source strategy game.

M3 - Conference contribution

SP - 32

EP - 39

BT - Proceedings of the QR’18

A2 - Falomir, Zoe

A2 - Coghill, George M.

A2 - Pang, Wei

ER -

Crouse M, McFate C, Forbus KD. Learning to Build Qualitative Scenario Models From Natural Language. In Falomir Z, Coghill GM, Pang W, editors, Proceedings of the QR’18. 2018. p. 32-39