{"id":7981,"date":"2026-01-01T05:07:00","date_gmt":"2025-12-31T20:07:00","guid":{"rendered":"https:\/\/devneko.jp\/wordpress\/?p=7981"},"modified":"2025-12-27T14:11:08","modified_gmt":"2025-12-27T05:11:08","slug":"spatialtree-how-spatial-abilities-branch-out-in-mllms","status":"publish","type":"post","link":"https:\/\/devneko.jp\/wordpress\/?p=7981","title":{"rendered":"SpatialTree: How Spatial Abilities Branch Out in MLLMs\u00a0"},"content":{"rendered":"\n<ul class=\"wp-block-list\">\n<li><strong>SpatialTree: How Spatial Abilities Branch Out in MLLMs\u00a0<\/strong>[109.3]<br>\u4f4e\u30ec\u30d9\u30eb\u77e5\u899a(L1)\u3001\u30e1\u30f3\u30bf\u30eb\u30de\u30c3\u30d4\u30f3\u30b0(L2)\u3001\u30b7\u30df\u30e5\u30ec\u30fc\u30b7\u30e7\u30f3(L3)\u3001\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u80fd\u529b(L4)\u306e4\u3064\u306e\u30ec\u30d9\u30eb\u306b\u7a7a\u9593\u80fd\u529b\u3092\u6574\u7406\u3059\u308b\u8a8d\u77e5\u79d1\u5b66\u306b\u7740\u60f3\u3092\u5f97\u305f\u968e\u5c64\u3092\u5c0e\u5165\u3059\u308b\u3002 \u8907\u96d1\u306a\u63a8\u8ad6\u306b\u306f\u5f79\u7acb\u3061\u307e\u3059\u304c\u3001\u76f4\u611f\u7684\u306a\u77e5\u899a\u3092\u640d\u306a\u3044\u307e\u3059\u3002 \u672c\u7a3f\u3067\u306f,\u4e0d\u5fc5\u8981\u306a\u719f\u8003\u3092\u6291\u5236\u3059\u308b\u30b7\u30f3\u30d7\u30eb\u306a\u81ea\u5df1\u601d\u8003\u6226\u7565\u3092\u63d0\u6848\u3059\u308b\u3002<br><a href=\"http:\/\/arxiv.org\/abs\/2512.20617v1\">\u8ad6\u6587<\/a>\u00a0\u00a0<a href=\"https:\/\/fugumt.com\/fugumt\/paper_check\/2512.20617v1\">\u53c2\u8003\u8a33\uff08\u30e1\u30bf\u30c7\u30fc\u30bf\uff09<\/a>\u00a0 \u00a0(Tue, 23 Dec 2025 18:59:46 GMT)<\/li>\n\n\n\n<li>\u300cSpatial abilities refer to the capacity to perceive, understand, reason about, and interact with 2D and 3D space, a long-standing topic in cognitive science [13, 45, 48]. In multimodal large language models (MLLMs), these abilities form the cornerstone of Spatial Intelligence (SI), yet remain challenging to study systematically due to their inherent complexity and broad scope [31, 63].\u300d\u3068\u306e\u3053\u3068\u3067Spatial abilities\u3092\u6e2c\u308b\u30d9\u30f3\u30c1\u30de\u30fc\u30af\u3092\u69cb\u7bc9\u3057\u3066\u3044\u308b\u30024\u30ec\u30d9\u30eb\u306f\u4e0b\u8a18\u306e\u901a\u308a\u3002\n<ul class=\"wp-block-list\">\n<li>L1 Perception: This level focuses on native perception of space, capturing raw geometric and physical attributes such as size, distance, and motion, without relying on language or symbolic reasoning. <\/li>\n\n\n\n<li>L2 Mental Mapping: This level maps spatial perception to language, grounding spatial concepts in linguistic semantics and forming language-structured spatial memory.<\/li>\n\n\n\n<li>L3 Mental Simulation: This level supports internal reasoning about space, enabling mental simulation, including causal reasoning about dynamics, relational and geometric problem solving, and sequential planning for actions and navigation.<\/li>\n\n\n\n<li>L4 Spatial Agent: This level executes actions in space, integrating perception, language, and reasoning to interact with the environment, interpret feedback, and complete long-horizon spatial tasks.<\/li>\n<\/ul>\n<\/li>\n\n\n\n<li>\u30ea\u30dd\u30b8\u30c8\u30ea\u306f<a href=\"https:\/\/spatialtree.github.io\/\">SpatialTree \u2013 How Spatial Abilities Branch Out in MLLMs<\/a>\u3001\u30ea\u30fc\u30c0\u30fc\u30dc\u30fc\u30c9\u306e\u30e2\u30c7\u30eb\u7fa4\u304c\u82e5\u5e72\u53e4\u3044\u3002\u3002<\/li>\n<\/ul>\n","protected":false},"excerpt":{"rendered":"","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[2],"tags":[251,517],"class_list":["post-7981","post","type-post","status-publish","format-standard","hentry","category-arxiv","tag-mllm","tag-517"],"_links":{"self":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/7981","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=7981"}],"version-history":[{"count":1,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/7981\/revisions"}],"predecessor-version":[{"id":7982,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/7981\/revisions\/7982"}],"wp:attachment":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=7981"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=7981"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=7981"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}