AI RESEARCH
Semantic Interaction Information mediates compositional generalization in latent space
arXiv CS.LG
•
ArXi:2603.27134v1 Announce Type: new Are there still barriers to generalization once all relevant variables are known? We address this question via a framework that casts compositional generalization as a variational inference problem over latent variables with parametric interactions. To explore this, we develop the Cognitive Gridworld, a stationary Partially Observable Marko Decision Process (POMDP) where observations are generated jointly by multiple latent variables, yet feedback is provided for only a single goal variable.