Towards machines that understand people

Andrew Howes; Jussi P. P. Jokinen; Antti Oulasvirta

doi:10.1002/aaai.12116

Towards machines that understand people

Andrew Howes^*, Jussi P. P. Jokinen, Antti Oulasvirta

^*Corresponding author for this work

Research output: Contribution to journal › Article › peer-review

131 Downloads (Pure)

Abstract

The ability to estimate the state of a human partner is an insufficient basis on which to build cooperative agents. Also needed is an ability to predict how people adapt their behavior in response to an agent's actions. We propose a new approach based on computational rationality, which models humans based on the idea that predictions can be derived by calculating policies that are approximately optimal given human‐like bounds. Computational rationality brings together reinforcement learning and cognitive modeling in pursuit of this goal, facilitating machine understanding of humans.

Original language	English
Journal	AI Magazine
Early online date	4 Sept 2023
DOIs	https://doi.org/10.1002/aaai.12116
Publication status	E-pub ahead of print - 4 Sept 2023

Access to Document

10.1002/aaai.12116Licence: Creative Commons: Attribution (CC BY)

HowesA2023TowardsFinal published version, 470 KBLicence: Creative Commons: Attribution (CC BY)

Cite this

@article{a6176053ea2c4b8b8a2dd75b831fa633,

title = "Towards machines that understand people",

abstract = "The ability to estimate the state of a human partner is an insufficient basis on which to build cooperative agents. Also needed is an ability to predict how people adapt their behavior in response to an agent's actions. We propose a new approach based on computational rationality, which models humans based on the idea that predictions can be derived by calculating policies that are approximately optimal given human‐like bounds. Computational rationality brings together reinforcement learning and cognitive modeling in pursuit of this goal, facilitating machine understanding of humans.",

author = "Andrew Howes and Jokinen, {Jussi P. P.} and Antti Oulasvirta",

year = "2023",

month = sep,

day = "4",

doi = "10.1002/aaai.12116",

language = "English",

journal = "AI Magazine",

issn = "0738-4602",

publisher = "AI Access Foundation",

}

TY - JOUR

T1 - Towards machines that understand people

AU - Howes, Andrew

AU - Jokinen, Jussi P. P.

AU - Oulasvirta, Antti

PY - 2023/9/4

Y1 - 2023/9/4

N2 - The ability to estimate the state of a human partner is an insufficient basis on which to build cooperative agents. Also needed is an ability to predict how people adapt their behavior in response to an agent's actions. We propose a new approach based on computational rationality, which models humans based on the idea that predictions can be derived by calculating policies that are approximately optimal given human‐like bounds. Computational rationality brings together reinforcement learning and cognitive modeling in pursuit of this goal, facilitating machine understanding of humans.

AB - The ability to estimate the state of a human partner is an insufficient basis on which to build cooperative agents. Also needed is an ability to predict how people adapt their behavior in response to an agent's actions. We propose a new approach based on computational rationality, which models humans based on the idea that predictions can be derived by calculating policies that are approximately optimal given human‐like bounds. Computational rationality brings together reinforcement learning and cognitive modeling in pursuit of this goal, facilitating machine understanding of humans.

U2 - 10.1002/aaai.12116

DO - 10.1002/aaai.12116

M3 - Article

SN - 0738-4602

JO - AI Magazine

JF - AI Magazine

ER -

Towards machines that understand people

Abstract

Access to Document

Fingerprint

Cite this