Grounding vs. Compositionality: On the Non-Complementarity of Reasoning in Neuro-Symbolic Systems
For neuro-symbolic AI researchers, this work provides empirical evidence that reasoning is a distinct capability requiring explicit training, not an emergent property of grounding.
This paper challenges the assumption that compositional reasoning emerges from symbol grounding in neuro-symbolic systems. Introducing the Iterative Logic Tensor Network (iLTN), they show that grounding alone fails to generalize, while joint training on grounding and reasoning achieves high zero-shot accuracy across novel entities, relations, and rule compositions.
Compositional generalization remains a foundational weakness of modern neural networks, limiting their robustness and applicability in domains requiring out-of-distribution reasoning. A central, yet unverified, assumption in neuro-symbolic AI is that compositional reasoning will emerge as a byproduct of successful symbol grounding. This work presents the first systematic empirical analysis to challenge this assumption by disentangling the contributions of grounding and reasoning. To operationalize this investigation, we introduce the Iterative Logic Tensor Network ($i$LTN), a fully differentiable architecture designed for multi-step deduction. Using a formal taxonomy of generalization -- probing for novel entities, unseen relations, and complex rule compositions -- we demonstrate that a model trained solely on a grounding objective fails to generalize. In contrast, our full $i$LTN, trained jointly on perceptual grounding and multi-step reasoning, achieves high zero-shot accuracy across all tasks. Our findings provide conclusive evidence that symbol grounding, while necessary, is insufficient for generalization, establishing that reasoning is not an emergent property but a distinct capability that requires an explicit learning objective.