Preprint
Jul 2026
RSF-GLLM: Bridging the Semantic Gap in Multi-Hop Knowledge Graph QA via Recurrent Soft-Flow and Decoupled LLM Generation
This work proposes RSF-GLLM, a framework decoupling differentiable graph reasoning from answer generation, and introduces flow sparsity regularization to theoretically guarantee convergence from soft probabilities to discrete reasoning paths.
Sambaran Bandyopadhyay, Ananth Muppidi
· 1 citation