Differentiable User Models

TL;DR

This paper introduces differentiable surrogates for probabilistic user models, enabling efficient inference and real-time AI interaction, which was previously computationally prohibitive.

Contribution

It presents a novel approach to make cognitive user models differentiable, allowing for fast inference suitable for online applications.

Findings

Achieved comparable modeling capabilities to likelihood-free inference methods.

Reduced computational cost enabling real-time AI-user interactions.

Demonstrated online menu-search task with efficient cognitive model usage.

Abstract

Probabilistic user modeling is essential for building machine learning systems in the ubiquitous cases with humans in the loop. However, modern advanced user models, often designed as cognitive behavior simulators, are incompatible with modern machine learning pipelines and computationally prohibitive for most practical applications. We address this problem by introducing widely-applicable differentiable surrogates for bypassing this computational bottleneck; the surrogates enable computationally efficient inference with modern cognitive models. We show experimentally that modeling capabilities comparable to the only available solution, existing likelihood-free inference methods, are achievable with a computational cost suitable for online applications. Finally, we demonstrate how AI-assistants can now use cognitive models for online interaction in a menu-search task, which has so far required hours of computation during interaction.