In many reasoning problems, premises are not observed as discrete symbols but must be inferred from high-dimensional inputs. The predicate vocabulary, argument structure, and trusted evidence are provided by a Knowledge Graph (KG) or rule definitions. Classical neuro-symbolic pipelines have a discrete interface between perception and deduction. We propose a neuro-soft-symbolic architecture for differentiable deductive reasoning over latent perceptual facts and knowledge-provided predicates.
SoftReason eliminates the gradient gap by representing the deductive state as a local soft interpretation tensor over candidate constants and predicates. Perception proposes probabilistic base facts, KG triples enter as high-confidence soft evidence, and every query anchor, predicate choice, and closure update remains differentiable. Our core innovation is a learned differentiable lift of the immediate-consequence operator. It utilizes predicate-definition embeddings and latent composition channels to form soft body-predicate mixtures, aggregate over all possible witnesses, propose query-conditioned head facts, and update the interpretation through a monotone probabilistic OR.
We instantiate the framework on Knowledge-aware Visual Question Answering (KVQA) and demonstrate how SoftReason supports end-to-end perceptual grounding, KG evidence injection, and differentiable deductive closure in one trainable architecture.