HiVAE: Hierarchical Latent Variables for Scalable Theory of Mind
This addresses the problem of scaling theory of mind for AI systems, though it identifies critical limitations in grounding learned representations.
The authors tackled scaling theory of mind reasoning to realistic domains by introducing HiVAE, a hierarchical variational architecture inspired by human cognition, achieving substantial performance improvements on a 3,185-node campus navigation task.
Theory of mind (ToM) enables AI systems to infer agents' hidden goals and mental states, but existing approaches focus mainly on small human understandable gridworld spaces. We introduce HiVAE, a hierarchical variational architecture that scales ToM reasoning to realistic spatiotemporal domains. Inspired by the belief-desire-intention structure of human cognition, our three-level VAE hierarchy achieves substantial performance improvements on a 3,185-node campus navigation task. However, we identify a critical limitation: while our hierarchical structure improves prediction, learned latent representations lack explicit grounding to actual mental states. We propose self-supervised alignment strategies and present this work to solicit community feedback on grounding approaches.