Context as Prior: Bayesian-Inspired Intent Inference for Non-Speaking Agents with a Household Cat Testbed

TL;DR

This paper introduces CatSignal, a Bayesian-inspired framework for inferring intent in non-speaking agents like household cats, effectively integrating context and behavior to improve accuracy and reduce shortcut errors.

Contribution

It presents a novel context-gated Product-of-Experts model for intent inference, demonstrating superior performance in a household cat dataset compared to traditional fusion methods.

Findings

Achieved 77.72% accuracy in intent inference on a domestic cat dataset.

Reduced shortcut failures caused by context ambiguity.

Outperformed feature concatenation and late-fusion baselines.

Abstract

Many agents in real-world environments cannot reliably communicate their goals through language, including household pets, pre-verbal infants, and other non-speaking embodied agents. In such settings, intent must be inferred from incomplete behavioral observations in context-rich environments. This creates a core ambiguity: observable behavior is often noisy or underspecified, while context provides strong prior information but can also induce brittle shortcut predictions if used naively. We present CatSignal, a Bayesian-inspired probabilistic framework for multimodal intent inference that models spatial context as a prior-like constraint and behavioral observations as evidence. Rather than treating context as an ordinary input feature, our method uses a context-gated Product-of-Experts formulation to compute posterior-like intent distributions from context, pose dynamics, and acoustic cues. We instantiate this formulation in a household cat setting as a focused proof-of-concept for intent inference in non-speaking agents. Under Leave-One-Video-Out evaluation on a multimodal domestic cat dataset, the proposed prior-guided fusion achieves the best overall accuracy of 77.72%, outperforming feature concatenation (71.83%) and stronger late-fusion baselines. More importantly, it substantially reduces context-driven shortcut failures in ambiguous cases. While simpler fusion strategies remain competitive in Macro-F1 and selective prediction, the proposed model provides the strongest overall accuracy and the best suppression of context-based shortcut collapse.