We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.HC

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Machine Learning

Title: Differentiable User Models

Authors: Alex Hämäläinen (1), Mustafa Mert Çelikok (1), Samuel Kaski (1, 2) ((1) Department of Computer Science, Aalto University, (2) Department of Computer Science, University of Manchester)
Abstract: Probabilistic user modeling is essential for building collaborative AI systems within probabilistic frameworks. However, modern advanced user models, often designed as cognitive behavior simulators, are computationally prohibitive for interactive use in cooperative AI assistants. In this extended abstract, we address this problem by introducing widely-applicable differentiable surrogates for bypassing this computational bottleneck; the surrogates enable using modern behavioral models with online computational cost which is independent of their original computational cost. We show experimentally that modeling capabilities comparable to likelihood-free inference methods are achievable, with over eight orders of magnitude reduction in computational time. Finally, we demonstrate how AI-assistants can computationally feasibly use cognitive models in a previously studied menu-search task.
Comments: This is an extended abstract accepted for presentation in NeurIPS 2022 HILL workshop
Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC)
Cite as: arXiv:2211.16277 [cs.LG]
  (or arXiv:2211.16277v1 [cs.LG] for this version)

Submission history

From: Alex Hämäläinen [view email]
[v1] Tue, 29 Nov 2022 15:09:18 GMT (305kb,D)

Link back to: arXiv, form interface, contact.