Amortised Experimental Design and Parameter Estimation for User Models of Pointing
This work addresses the challenge of minimizing human data needs for parameter estimation in user modeling, which is incremental by building on prior efficient experimental design methods.
The paper tackles the problem of efficiently estimating parameters for user models in interaction design by amortizing the computational cost of experimental design through a policy trained with simulated participants, demonstrating the approach on three pointing models with reduced data requirements.
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.