Title: BiNSGPS: Solving Geometry Problems Through Bidirectional Neuro-Symbolic Interaction

Original: arXiv:2606.04648v1 Announce Type: new Abstract: Geometry problem solving poses distinct challenges in artificial intelligence. Existing approaches typically fall into two paradigms: symbolic methods, which exhibit limited adaptability, and neural methods, which are prone to hallucinations. Recent neuro-symbolic hybrids predominantly rely on a unidirectional pipeline where neural outputs are fed into solvers without feedback, making system brittle to early-stage errors. To break this unidirectional bottleneck, we propose BiNSGPS, a framework that establishes Bidirectional Neuro-Symbolic Interaction (BiNS) between a MLLM Adviser and a Symbolic Solver. MLLM Adviser actively incorporates feedback from the symbolic solver to dynamically rectify inconsistent formal representations or propose auxiliary hypotheses, resolving symbolic conflicts and facilitating complex deductions.

Rewrite: Addressing geometry problems presents unique hurdles for artificial intelligence. Current methodologies generally adhere to one of two frameworks: symbolic techniques, which struggle with adaptability, or neural approaches, which are susceptible to generating hallucinations. While recent hybrid neuro-symbolic models have emerged, they mostly depend on a one-way workflow in which neural predictions are passed to solvers without any return feedback. This lack of interaction renders the system fragile, particularly when early errors occur. To overcome this limitation, we introduce BiNSGPS, a framework designed to enable Bidirectional Neuro-Symbolic Interaction (BiNS) between a Symbolic Solver and an MLLM Adviser. Within this structure, the MLLM Adviser leverages feedback from the symbolic solver to dynamically correct inconsistent formal representations or suggest auxiliary hypotheses. This bidirectional exchange helps resolve symbolic conflicts and supports intricate logical deductions.

Source: arXiv Generated at: 2026-06-04 00:00:00 UTC