A method for automating token causal explanation and discovery

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

Abstract

Explaining why events occur is key to making decisions, assigning blame, and enacting policies. Despite the need, few methods can compute explanations in an automated way. Existing solutions start with a type-level model (e.g. factors affecting risk of disease), and use this to explain token-level events (e.g. cause of an individual's illness). This is limiting, since an individual's illness may be due to a previously unknown drug interaction. We propose a hybrid method for token explanation that uses known type-level models while also discovering potentially novel explanations. On simulated data with ground truth, the approach finds accurate explanations when observations match what is known, and correctly finds novel relationships when they do not. On real world data, our approach finds explanations consistent with intuition.

Original languageEnglish (US)
Title of host publicationFLAIRS 2017 - Proceedings of the 30th International Florida Artificial Intelligence Research Society Conference
PublisherAAAI Press
Pages176-181
Number of pages6
ISBN (Electronic)9781577357872
StatePublished - Jan 1 2017
Event30th International Florida Artificial Intelligence Research Society Conference, FLAIRS 2017 - Marco Island, United States
Duration: May 22 2017May 24 2017

Other

Other30th International Florida Artificial Intelligence Research Society Conference, FLAIRS 2017
CountryUnited States
CityMarco Island
Period5/22/175/24/17

Fingerprint

Drug interactions
Decision making

Cite this

Zheng, M., & Kleinberg, S. (2017). A method for automating token causal explanation and discovery. In FLAIRS 2017 - Proceedings of the 30th International Florida Artificial Intelligence Research Society Conference (pp. 176-181). AAAI Press.
Zheng, Min ; Kleinberg, Samantha. / A method for automating token causal explanation and discovery. FLAIRS 2017 - Proceedings of the 30th International Florida Artificial Intelligence Research Society Conference. AAAI Press, 2017. pp. 176-181
@inproceedings{28d831a4d5be4404957f6d689fa21c6e,
title = "A method for automating token causal explanation and discovery",
abstract = "Explaining why events occur is key to making decisions, assigning blame, and enacting policies. Despite the need, few methods can compute explanations in an automated way. Existing solutions start with a type-level model (e.g. factors affecting risk of disease), and use this to explain token-level events (e.g. cause of an individual's illness). This is limiting, since an individual's illness may be due to a previously unknown drug interaction. We propose a hybrid method for token explanation that uses known type-level models while also discovering potentially novel explanations. On simulated data with ground truth, the approach finds accurate explanations when observations match what is known, and correctly finds novel relationships when they do not. On real world data, our approach finds explanations consistent with intuition.",
author = "Min Zheng and Samantha Kleinberg",
year = "2017",
month = "1",
day = "1",
language = "English (US)",
pages = "176--181",
booktitle = "FLAIRS 2017 - Proceedings of the 30th International Florida Artificial Intelligence Research Society Conference",
publisher = "AAAI Press",

}

Zheng, M & Kleinberg, S 2017, A method for automating token causal explanation and discovery. in FLAIRS 2017 - Proceedings of the 30th International Florida Artificial Intelligence Research Society Conference. AAAI Press, pp. 176-181, 30th International Florida Artificial Intelligence Research Society Conference, FLAIRS 2017, Marco Island, United States, 5/22/17.

A method for automating token causal explanation and discovery. / Zheng, Min; Kleinberg, Samantha.

FLAIRS 2017 - Proceedings of the 30th International Florida Artificial Intelligence Research Society Conference. AAAI Press, 2017. p. 176-181.

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

TY - GEN

T1 - A method for automating token causal explanation and discovery

AU - Zheng, Min

AU - Kleinberg, Samantha

PY - 2017/1/1

Y1 - 2017/1/1

N2 - Explaining why events occur is key to making decisions, assigning blame, and enacting policies. Despite the need, few methods can compute explanations in an automated way. Existing solutions start with a type-level model (e.g. factors affecting risk of disease), and use this to explain token-level events (e.g. cause of an individual's illness). This is limiting, since an individual's illness may be due to a previously unknown drug interaction. We propose a hybrid method for token explanation that uses known type-level models while also discovering potentially novel explanations. On simulated data with ground truth, the approach finds accurate explanations when observations match what is known, and correctly finds novel relationships when they do not. On real world data, our approach finds explanations consistent with intuition.

AB - Explaining why events occur is key to making decisions, assigning blame, and enacting policies. Despite the need, few methods can compute explanations in an automated way. Existing solutions start with a type-level model (e.g. factors affecting risk of disease), and use this to explain token-level events (e.g. cause of an individual's illness). This is limiting, since an individual's illness may be due to a previously unknown drug interaction. We propose a hybrid method for token explanation that uses known type-level models while also discovering potentially novel explanations. On simulated data with ground truth, the approach finds accurate explanations when observations match what is known, and correctly finds novel relationships when they do not. On real world data, our approach finds explanations consistent with intuition.

UR - http://www.scopus.com/inward/record.url?scp=85029514096&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85029514096&partnerID=8YFLogxK

M3 - Conference contribution

SP - 176

EP - 181

BT - FLAIRS 2017 - Proceedings of the 30th International Florida Artificial Intelligence Research Society Conference

PB - AAAI Press

ER -

Zheng M, Kleinberg S. A method for automating token causal explanation and discovery. In FLAIRS 2017 - Proceedings of the 30th International Florida Artificial Intelligence Research Society Conference. AAAI Press. 2017. p. 176-181