Mixed Likelihood Variational Gaussian Processes

2025-03-06

AI Summary

Gaussian processes (GPs) offer well-calibrated uncertainty estimates in human-in-the-loop experiments but often neglect auxiliary human inputs—such as domain priors, Likert-scale confidence ratings, or physical constraints—leading to biased uncertainty modeling. To address this, we propose a hybrid-likelihood variational Gaussian process framework that jointly models task responses and heterogeneous auxiliary signals (categorical, regression, and constraint-based likelihoods) within a unified evidence lower bound (ELBO) objective, enabling end-to-end variational inference. Our method integrates human feedback modeling, active learning, and preference learning, thereby enhancing generalization and convergence speed. Evaluated on three real-world human studies—VR-based visual error detection, haptic roughness perception, and robotic gait preference learning—the approach significantly improves model fidelity and active sampling efficiency. Results demonstrate the effectiveness and broad applicability of cross-modal auxiliary information fusion for uncertainty-aware human-centered learning.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Reasoning under Uncertainty: Relational Probabilistic Models
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Gaussian processes (GPs) are powerful models for human-in-the-loop experiments due to their flexibility and well-calibrated uncertainty. However, GPs modeling human responses typically ignore auxiliary information, including a priori domain expertise and non-task performance information like user confidence ratings. We propose mixed likelihood variational GPs to leverage auxiliary information, which combine multiple likelihoods in a single evidence lower bound to model multiple types of data. We demonstrate the benefits of mixing likelihoods in three real-world experiments with human participants. First, we use mixed likelihood training to impose prior knowledge constraints in GP classifiers, which accelerates active learning in a visual perception task where users are asked to identify geometric errors resulting from camera position errors in virtual reality. Second, we show that leveraging Likert scale confidence ratings by mixed likelihood training improves model fitting for haptic perception of surface roughness. Lastly, we show that Likert scale confidence ratings improve human preference learning in robot gait optimization. The modeling performance improvements found using our framework across this diverse set of applications illustrates the benefits of incorporating auxiliary information into active learning and preference learning by using mixed likelihoods to jointly model multiple inputs.

Problem

• Incorporates auxiliary information into Gaussian processes for human-in-the-loop experiments.
• Improves active learning in visual perception tasks using prior knowledge constraints.
• Enhances model fitting and preference learning with Likert scale confidence ratings.

Innovation

• Mixed likelihood variational Gaussian processes model
• Combine multiple likelihoods for diverse data types
• Enhance active learning with auxiliary information