Amortised Experimental Design and Parameter Estimation for User Models of Pointing

Loading...
Thumbnail Image

Access rights

openAccess

URL

Journal Title

Journal ISSN

Volume Title

A4 Artikkeli konferenssijulkaisussa

Date

2023-04-19

Major/Subject

Mcode

Degree programme

Language

en

Pages

Series

CHI 2023 - Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems

Abstract

User models play an important role in interaction design, supporting automation of interaction design choices. In order to do so, model parameters must be estimated from user data. While very large amounts of user data are sometimes required, recent research has shown how experiments can be designed so as to gather data and infer parameters as efficiently as possible, thereby minimising the data requirement. In the current article, we investigate a variant of these methods that amortises the computational cost of designing experiments by training a policy for choosing experimental designs with simulated participants. Our solution learns which experiments provide the most useful data for parameter estimation by interacting with in-silico agents sampled from the model space thereby using synthetic data rather than vast amounts of human data. The approach is demonstrated for three progressively complex models of pointing.

Description

Funding Information: This work was supported by the Academy of Finland (Flagship programme: Finnish Center for Artifcial Intelligence FCAI) and ELISE Networks of Excellence Centres (EU Horizon:2020 grant agreement 951847) and Bitville Oy. The authors want to thank the Probabilistic Machine Learning (PML) and the User Interfaces research groups at Aalto University for fruitful discussions and feedback. Publisher Copyright: © 2023 Owner/Author. | openaire: EC/H2020/951847/EU//ELISE

Keywords

active inference, adaptive experiment design, computational rationality, parameter estimation, user models

Other note

Citation

Keurulainen, A, Westerlund, I R, Keurulainen, O & Howes, A 2023, Amortised Experimental Design and Parameter Estimation for User Models of Pointing . in CHI 2023 - Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems ., 772, ACM, ACM SIGCHI Annual Conference on Human Factors in Computing Systems, Hamburg, Germany, 23/04/2023 . https://doi.org/10.1145/3544548.3581483