{"id":6826,"date":"2025-05-29T06:03:00","date_gmt":"2025-05-28T21:03:00","guid":{"rendered":"https:\/\/devneko.jp\/wordpress\/?p=6826"},"modified":"2025-05-25T10:18:03","modified_gmt":"2025-05-25T01:18:03","slug":"visual-planning-lets-think-only-with-images","status":"publish","type":"post","link":"https:\/\/devneko.jp\/wordpress\/?p=6826","title":{"rendered":"Visual Planning: Let&#8217;s Think Only with Images"},"content":{"rendered":"\n<ul class=\"wp-block-list\">\n<li><strong>Visual Planning: Let&#8217;s Think Only with Images&nbsp;<\/strong>[30.7]<br>\u6211\u3005\u306f\u3001\u7279\u306b\u7a7a\u9593\u7684\u60c5\u5831\u3084\u5e7e\u4f55\u5b66\u7684\u60c5\u5831\u3092\u542b\u3080\u30bf\u30b9\u30af\u306b\u304a\u3044\u3066\u3001\u8a00\u8a9e\u304c\u63a8\u8ad6\u306b\u304a\u3044\u3066\u6700\u3082\u81ea\u7136\u3067\u52b9\u679c\u7684\u306a\u30e2\u30c0\u30ea\u30c6\u30a3\u3067\u3042\u308b\u3068\u306f\u9650\u3089\u306a\u3044\u3068\u8ad6\u3058\u308b\u3002 \u305d\u3053\u3067\u672c\u7814\u7a76\u3067\u306f,\u30c6\u30ad\u30b9\u30c8\u304b\u3089\u72ec\u7acb\u3057\u3066,\u7d14\u7c8b\u306b\u8996\u899a\u7684\u306a\u8868\u73fe\u306b\u3088\u308b\u30d7\u30e9\u30f3\u30cb\u30f3\u30b0\u3092\u53ef\u80fd\u306b\u3059\u308b,\u30d3\u30b8\u30e5\u30a2\u30eb\u30d7\u30e9\u30f3\u30cb\u30f3\u30b0\u3068\u3044\u3046\u65b0\u305f\u306a\u30d1\u30e9\u30c0\u30a4\u30e0\u3092\u63d0\u6848\u3059\u308b\u3002 \u3053\u306e\u30d1\u30e9\u30c0\u30a4\u30e0\u3067\u306f\u3001\u8a08\u753b\u306f\u8996\u899a\u9818\u57df\u306b\u304a\u3051\u308b\u30b9\u30c6\u30c3\u30d7\u30d0\u30a4\u30b9\u30c6\u30c3\u30d7\u306e\u63a8\u8ad6\u3092\u7b26\u53f7\u5316\u3059\u308b\u4e00\u9023\u306e\u753b\u50cf\u3092\u901a\u3057\u3066\u5b9f\u884c\u3055\u308c\u308b\u3002<br><a href=\"http:\/\/arxiv.org\/abs\/2505.11409v1\">\u8ad6\u6587<\/a>&nbsp;&nbsp;<a href=\"https:\/\/fugumt.com\/fugumt\/paper_check\/2505.11409v1\">\u53c2\u8003\u8a33\uff08\u30e1\u30bf\u30c7\u30fc\u30bf\uff09<\/a>&nbsp; &nbsp;(Fri, 16 May 2025 16:17:22 GMT)<\/li>\n\n\n\n<li>\u300cBy enabling models to operate entirely through visual state transitions without textual mediation, we demonstrate that purely visual representations can lead to more effective and intuitive planning,\u300d\u3068\u306e\u3053\u3068\u3002\u30c6\u30ad\u30b9\u30c8\u306f\u5f37\u529b\u3060\u304c\u4e07\u80fd\u3068\u3044\u3046\u308f\u3051\u3067\u306f\u306a\u304f\u30bf\u30b9\u30af\u306b\u3088\u3063\u3066\u306f\u8a08\u753b\u30ec\u30d9\u30eb\u3067\u753b\u50cf\u304c\u6709\u52b9\u306a\u3053\u3068\u304c\u3042\u308b\u306e\u306f\u7d0d\u5f97\u611f\u304c\u3042\u308b\u3002\u3068\u3066\u3082\u9762\u767d\u3044\u3002GRIT\u3067\u3082\u601d\u3063\u305f\u304c\u753b\u50cf\u306e\u529b\u3092\u4f7f\u3063\u3066\u3044\u304f\u30a2\u30d7\u30ed\u30fc\u30c1\u306f\u3068\u3066\u3082\u6709\u671b\u306b\u601d\u3048\u308b\u3002<\/li>\n\n\n\n<li>\u30ea\u30dd\u30b8\u30c8\u30ea\u306f<a href=\"https:\/\/github.com\/yix8\/VisualPlanning\">GitHub &#8211; yix8\/VisualPlanning: Visual Planning: Let&#8217;s Think Only with Images<\/a><\/li>\n<\/ul>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>GRIT: Teaching MLLMs to Think with Images\u00a0<\/strong>[22.7]<br>Grounded Reasoning with Images and Texts (GRIT) \u306fMLLM\u3092\u753b\u50cf\u3067\u8003\u3048\u308b\u305f\u3081\u306e\u65b0\u3057\u3044\u624b\u6cd5\u3067\u3042\u308b\u3002 GRIT\u306f\u81ea\u7136\u8a00\u8a9e\u3068\u660e\u793a\u7684\u306a\u5883\u754c\u30dc\u30c3\u30af\u30b9\u5ea7\u6a19\u3092\u30a4\u30f3\u30bf\u30fc\u30ea\u30fc\u30d6\u3059\u308b\u63a8\u8ad6\u9023\u9396\u3092\u751f\u6210\u3059\u308b\u3002 GRIT\u306f\u4f8b\u5916\u7684\u306a\u30c7\u30fc\u30bf\u52b9\u7387\u3092\u5b9f\u73fe\u3057\u3001\u65e2\u5b58\u306e\u30c7\u30fc\u30bf\u30bb\u30c3\u30c8\u304b\u308920\u306e\u30a4\u30e1\u30fc\u30b8\u30af\u30a8\u30b9\u30c8\u30fb\u30a2\u30f3\u30b5\u30fc\u30c8\u30ec\u30c3\u30c8\u3092\u5fc5\u8981\u3068\u3059\u308b\u3002<br><a href=\"http:\/\/arxiv.org\/abs\/2505.15879v1\">\u8ad6\u6587<\/a>\u00a0\u00a0<a href=\"https:\/\/fugumt.com\/fugumt\/paper_check\/2505.15879v1\">\u53c2\u8003\u8a33\uff08\u30e1\u30bf\u30c7\u30fc\u30bf\uff09<\/a>\u00a0 \u00a0(Wed, 21 May 2025 17:54:49 GMT)<\/li>\n\n\n\n<li>\u30d7\u30ed\u30b8\u30a7\u30af\u30c8\u30b5\u30a4\u30c8\u306f<a href=\"https:\/\/grounded-reasoning.github.io\/\">GRIT: Teaching MLLMs to Think with Images<\/a><\/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":[636],"class_list":["post-6826","post","type-post","status-publish","format-standard","hentry","category-arxiv","tag-636"],"_links":{"self":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/6826","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=6826"}],"version-history":[{"count":2,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/6826\/revisions"}],"predecessor-version":[{"id":6831,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/6826\/revisions\/6831"}],"wp:attachment":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=6826"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=6826"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=6826"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}