{"ID":2890226,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2507.20021","arxiv_id":"2507.20021","title":"When Engineering Outruns Intelligence: Rethinking Instruction-Guided Navigation","abstract":"Recent ObjectNav systems credit large language models (LLMs) for sizable zero-shot gains, yet it remains unclear how much comes from language versus geometry. We revisit this question by re-evaluating an instruction-guided pipeline, InstructNav, under a detector-controlled setting and introducing two training-free variants that only alter the action value map: a geometry-only Frontier Proximity Explorer (FPE) and a lightweight Semantic-Heuristic Frontier (SHF) that polls the LLM with simple frontier votes. Across HM3D and MP3D, FPE matches or exceeds the detector-controlled instruction follower while using no API calls and running faster; SHF attains comparable accuracy with a smaller, localized language prior. These results suggest that carefully engineered frontier geometry accounts for much of the reported progress, and that language is most reliable as a light heuristic rather than an end-to-end planner. Code available at: https://github.com/matinaghaei/instructnav-scrutinized","short_abstract":"Recent ObjectNav systems credit large language models (LLMs) for sizable zero-shot gains, yet it remains unclear how much comes from language versus geometry. We revisit this question by re-evaluating an instruction-guided pipeline, InstructNav, under a detector-controlled setting and introducing two training-free vari...","url_abs":"https://arxiv.org/abs/2507.20021","url_pdf":"https://arxiv.org/pdf/2507.20021v3","authors":"[\"Matin Aghaei\",\"Lingfeng Zhang\",\"Mohammad Ali Alomrani\",\"Mahdi Biparva\",\"Yingxue Zhang\"]","published":"2025-07-26T17:37:15Z","proceeding":"cs.RO","tasks":"[\"cs.RO\",\"cs.AI\",\"cs.LG\"]","methods":"[\"Large Language Model\",\"Language Model\"]","has_code":false,"code_links":[{"ID":611756,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_id":2890226,"paper_url":"https://arxiv.org/abs/2507.20021","paper_title":"When Engineering Outruns Intelligence: Rethinking Instruction-Guided Navigation","repo_url":"https://github.com/matinaghaei/instructnav-scrutinized","is_official":false,"mentioned_in_paper":false,"mentioned_in_github":true,"github_stars":0}]}