SoftReason: A Fully Differentiable Neuro-Soft-Symbolic Deductive Reasoning Architecture over High-Dimensional Perceptual Data
2026-07-22 • Artificial Intelligence
Artificial Intelligence
AI summaryⓘ
The authors present SoftReason, a new system that blends perception and logical reasoning into a single learnable process. Unlike traditional methods that handle perception and logic separately, SoftReason treats reasoning as a smooth, adjustable operation over uncertain facts derived from images and knowledge graphs. This allows it to perform complex question answering tasks about images while learning how to combine visual evidence and background knowledge. Their approach is shown in a visual question answering setup where all components are trained together.
Neuro-symbolic reasoningKnowledge GraphDifferentiable reasoningVisual Question AnsweringImmediate consequence operatorPredicate logicLatent representationProbabilistic reasoningEnd-to-end learning
Authors
Wael AbdAlmageed
Abstract
In many reasoning problems, the premises are not observed as discrete symbols, but must be inferred from high-dimensional inputs. Further, the predicate vocabulary, argument structure, and trusted evidence are supplied by a Knowledge Graph (KG), or rule definitions. Classical neuro-symbolic pipelines have a discrete interface between perception and deduction. We present a neuro-soft-symbolic architecture for differentiable deductive reasoning over latent perceptual facts and knowledge-provided predicates. SoftReason removes 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 uses 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 demonstrates how SoftReason supports end-to-end perceptual grounding, KG evidence injection, and differentiable deductive closure in one trainable architecture.