Addressing the Symbol Grounding Problem with Constraints in Neuro-Symbolic Planning

In this paper, we address the Symbol Grounding Problem (SGP) in the context of neuro-symbolic planning, where the categorical vectors learned to represent high dimensional inputs suffer from instability, which poses a problem of efficiency during the planning phase. One way to alleviate the SGP is to enforce constraints — among the latent variables — by expressing them in the loss function during the learning process. Combining an existing tool for invariant search and ideas from Logic Tensor Networks (fuzzy logic), we propose to automatize the process of finding and enforcing relevant constraints. We apply our idea to LatPlan, a domain independent, image-based classical planner.