HAPI: A Model for Learning Robot Facial Expressions from Human Preferences

2025-03-21

AI Summary

This work addresses the problem of robotic facial expressions lacking naturalness and social authenticity. To this end, we propose HAPI, a human-preference-driven learning-to-rank framework. HAPI introduces a novel Siamese RankNet architecture trained on pairwise preference annotations, directly embedding human affective intuition into the expression generation loop. The framework integrates human-in-the-loop annotation, Bayesian optimization, and online evaluation on a 35-DOF anthropomorphic robot platform. Empirical results demonstrate significant improvements over baseline and expert-designed methods across three prototypical expressions—anger, happiness, and surprise. A user study confirms that HAPI-generated expressions achieve a 42% increase in perceived social resonance and realism (p < 0.01), validating the efficacy of preference modeling for enhancing the naturalness of human–robot interaction.

Technology Category

• Humans and AI: Learning Human Values and Preferences
• Intelligent Robots: Human-Robot Interaction
• Machine Learning: Learning Preferences or Rankings

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Automatic robotic facial expression generation is crucial for human-robot interaction, as handcrafted methods based on fixed joint configurations often yield rigid and unnatural behaviors. Although recent automated techniques reduce the need for manual tuning, they tend to fall short by not adequately bridging the gap between human preferences and model predictions-resulting in a deficiency of nuanced and realistic expressions due to limited degrees of freedom and insufficient perceptual integration. In this work, we propose a novel learning-to-rank framework that leverages human feedback to address this discrepancy and enhanced the expressiveness of robotic faces. Specifically, we conduct pairwise comparison annotations to collect human preference data and develop the Human Affective Pairwise Impressions (HAPI) model, a Siamese RankNet-based approach that refines expression evaluation. Results obtained via Bayesian Optimization and online expression survey on a 35-DOF android platform demonstrate that our approach produces significantly more realistic and socially resonant expressions of Anger, Happiness, and Surprise than those generated by baseline and expert-designed methods. This confirms that our framework effectively bridges the gap between human preferences and model predictions while robustly aligning robotic expression generation with human affective responses.

Problem

• Bridging human preferences and robotic facial expression predictions
• Enhancing expressiveness of robotic faces using human feedback
• Generating realistic and socially resonant robotic expressions

Innovation

• Learning-to-rank framework with human feedback
• Siamese RankNet-based HAPI model
• Bayesian Optimization for expression refinement