{"type":"ScholarlyArticle","title":"GeoMind: Explicit Spatial Reasoning via Dual-Reference Geometric Modeling","authors":[{"name":"Xing Wei","slug":"xing-wei","given":"Xing","family":"Wei","affiliation":""},{"name":"Aoxiang Tian","slug":"aoxiang-tian","given":"Aoxiang","family":"Tian","affiliation":""},{"name":"Shaofan Liu","slug":"shaofan-liu","given":"Shaofan","family":"Liu","affiliation":""},{"name":"Jiansheng Peng","slug":"jiansheng-peng","given":"Jiansheng","family":"Peng","affiliation":""},{"name":"Chong Zhao","slug":"chong-zhao","given":"Chong","family":"Zhao","affiliation":""},{"name":"Xiang Bi","slug":"xiang-bi","given":"Xiang","family":"Bi","affiliation":""},{"name":"Yang Lu","slug":"yang-lu","given":"Yang","family":"Lu","affiliation":""},{"name":"Benhong Zhang","slug":"benhong-zhang","given":"Benhong","family":"Zhang","affiliation":""},{"name":"Fan Yang","slug":"fan-yang","given":"Fan","family":"Yang","affiliation":""}],"abstract":"While Vision-Language Models (VLMs) excel at semantic understanding, they struggle to comprehend 3D spatial relationships from limited views. Their reliance on implicit geometric encoding often leads to severe hallucinations and inconsistencies in spatial reasoning tasks. To address this, we introduce GeoMind, a model-then-reason framework that employs a single LLM to autoregressively generate an explicit Geometric Description Language (GDL) map, serving as a grounded context to derive the final answer. This intermediate GDL map provides an explicit and queryable world representation. Leveraging this explicit representation, we enforce a strict referential constraint, compelling the model to ground reasoning solely on the instantiated entities to ensure referential integrity and auditability. Specifically, we lift multi-view observations into object-centric tokens using frozen geometric priors and instance masks. The LLM is trained via a two-stage curriculum with programmatic supervision to generate the GDL map as a prerequisite for answering. On five spatial understanding benchmarks in both image and video settings, GeoMind delivers average accuracy gains of +6.9% (2B) and +9.8% (8B) over Qwen3-VL baselines. Our results suggest that explicit geometric grounding enables robust spatial reasoning without human annotation, providing a scalable and practical route to stronger spatial intelligence in large VLMs.","doi":"10.24963/ijcai.2026/203","url":"https://ijcai.org/proceedings/2026/203","pdf":"https://ijcai.org/proceedings/2026/0203.pdf","markdown":"https://ijcai.org/proceedings/2026/203.md","bibtex":"https://ijcai.org/proceedings/2026/bibtex/203","volume":{"slug":"2026","title":"Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence","short_title":"IJCAI 2026","url":"https://ijcai.org/proceedings/2026/","year":2026,"location":"Bremen, Germany","dates":"15-21 August 2026","editors":"Diego Calvanese","isbn":"978-1-956792-09-6","publisher":"International Joint Conferences on Artificial Intelligence Organization","doi_prefix":"10.24963/ijcai.2026"},"section":"Computer Vision","track":"Main Track","number":203,"pages":{"first":1821,"last":1829},"keywords":["Humans and AI: Cognitive modeling","Computer Vision: Vision, language and reasoning"],"videos":[],"references":[],"license_note":"Copyright International Joint Conferences on Artificial Intelligence."}