{"id":6734,"date":"2025-05-12T02:54:00","date_gmt":"2025-05-11T17:54:00","guid":{"rendered":"https:\/\/devneko.jp\/wordpress\/?p=6734"},"modified":"2025-05-11T08:05:51","modified_gmt":"2025-05-10T23:05:51","slug":"on-path-to-multimodal-generalist-general-level-and-general-bench","status":"publish","type":"post","link":"https:\/\/devneko.jp\/wordpress\/?p=6734","title":{"rendered":"On Path to Multimodal Generalist: General-Level and General-Bench"},"content":{"rendered":"\n<ul class=\"wp-block-list\">\n<li><strong>On Path to Multimodal Generalist: General-Level and General-Bench&nbsp;<\/strong>[154.0]<br>\u672c\u7a3f\u3067\u306f,MLLM\u306e\u6027\u80fd\u3068\u6c4e\u7528\u6027\u30925\u6bb5\u968e\u306b\u5b9a\u7fa9\u3057\u305f\u8a55\u4fa1\u30d5\u30ec\u30fc\u30e0\u30ef\u30fc\u30af\u3067\u3042\u308bGeneral-Level\u3092\u7d39\u4ecb\u3059\u308b\u3002 \u30d5\u30ec\u30fc\u30e0\u30ef\u30fc\u30af\u306e\u4e2d\u6838\u306fSynergy\u306e\u6982\u5ff5\u3067\u3042\u308a\u3001\u30e2\u30c7\u30eb\u304c\u7406\u89e3\u3068\u751f\u6210\u3092\u307e\u305f\u3044\u3060\u4e00\u8cab\u6027\u306e\u3042\u308b\u6a5f\u80fd\u3092\u7dad\u6301\u3059\u308b\u304b\u3069\u3046\u304b\u3092\u6e2c\u5b9a\u3059\u308b\u3002 \u65e2\u5b58\u306e100\u4ee5\u4e0a\u306eMLLM\u3092\u542b\u3080\u8a55\u4fa1\u7d50\u679c\u306f\u3001\u30b8\u30a7\u30cd\u30e9\u30ea\u30b9\u30c8\u306e\u80fd\u529b\u30e9\u30f3\u30ad\u30f3\u30b0\u3092\u660e\u3089\u304b\u306b\u3059\u308b\u3002<br><a href=\"http:\/\/arxiv.org\/abs\/2505.04620v1\">\u8ad6\u6587<\/a>&nbsp;&nbsp;<a href=\"https:\/\/fugumt.com\/fugumt\/paper_check\/2505.04620v1\">\u53c2\u8003\u8a33\uff08\u30e1\u30bf\u30c7\u30fc\u30bf\uff09<\/a>&nbsp; &nbsp;(Wed, 07 May 2025 17:59:32 GMT)<\/li>\n\n\n\n<li>\u300cThis leads to a critical question: Can we simply assume that higher performance across tasks indicates a stronger MLLM capability, bringing us closer to human-level AI?\u300d\u306b\u5bfe\u3059\u308b\u8a55\u4fa1\u30d5\u30ec\u30fc\u30e0\u30ef\u30fc\u30af\u3002\u81ea\u52d5\u904b\u8ee2\u306e\u3088\u3046\u306a\u5927\u304d\u304f5\u6bb5\u968e\u306e\u30ec\u30d9\u30eb\u8a2d\u5b9a\u3092\u884c\u3063\u3066\u3044\u308b\u3002\u73fe\u6642\u70b9\u3067\u306f\u300cOur evaluation of over 100 existing top-performing LLM\/MLLM systems has uncovered critical insights into their capabilities and rankings as multimodal generalists. The most notable finding is that most MLLMs lack the cross-task or cross-modal synergy ability required for higher-level classifications, with even advanced models like GPT-4V and GPT-4o not achieving top ranks.\u300d\u3068\u306e\u3053\u3068\u3060\u304c\u2026<\/li>\n\n\n\n<li>\u30d7\u30ed\u30b8\u30a7\u30af\u30c8\u30b5\u30a4\u30c8\u306f<a href=\"https:\/\/generalist.top\/\">Path to Multimodal Generalist<\/a>\u3001\u30ea\u30fc\u30c0\u30fc\u30dc\u30fc\u30c9\u306f<a href=\"https:\/\/generalist.top\/leaderboard\">Path to Multimodal Generalist<\/a><\/li>\n<\/ul>\n\n\n\n<p>\u4e0b\u8a18\u30b5\u30fc\u30d9\u30a4\u3082\u6ce8\u76ee<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>Perception, Reason, Think, and Plan: A Survey on Large Multimodal Reasoning Models\u00a0<\/strong>[79.5]<br>\u63a8\u8ad6\u306f\u77e5\u6027\u306e\u4e2d\u5fc3\u306b\u3042\u308a\u3001\u6c7a\u5b9a\u3057\u3001\u7d50\u8ad6\u3092\u5c0e\u304d\u3001\u30c9\u30e1\u30a4\u30f3\u3092\u307e\u305f\u3044\u3067\u4e00\u822c\u5316\u3059\u308b\u80fd\u529b\u3092\u5f62\u6210\u3059\u308b\u3002 \u4eba\u5de5\u77e5\u80fd\u306b\u304a\u3044\u3066\u3001\u30b7\u30b9\u30c6\u30e0\u304c\u30aa\u30fc\u30d7\u30f3\u3067\u4e0d\u78ba\u5b9f\u3067\u30de\u30eb\u30c1\u30e2\u30fc\u30c0\u30eb\u306a\u74b0\u5883\u3067\u307e\u3059\u307e\u3059\u6a5f\u80fd\u3059\u308b\u306b\u3064\u308c\u3066\u3001\u63a8\u8ad6\u306f\u5805\u7262\u3067\u9069\u5fdc\u7684\u306a\u884c\u52d5\u3092\u53ef\u80fd\u306b\u3059\u308b\u305f\u3081\u306b\u4e0d\u53ef\u6b20\u3068\u306a\u308b\u3002 \u5927\u898f\u6a21\u30de\u30eb\u30c1\u30e2\u30fc\u30c0\u30eb\u63a8\u8ad6\u30e2\u30c7\u30eb(LMRM)\u306f\u3001\u30c6\u30ad\u30b9\u30c8\u3001\u753b\u50cf\u3001\u30aa\u30fc\u30c7\u30a3\u30aa\u3001\u30d3\u30c7\u30aa\u306a\u3069\u306e\u30e2\u30c0\u30ea\u30c6\u30a3\u3092\u7d71\u5408\u3057\u3001\u8907\u96d1\u306a\u63a8\u8ad6\u6a5f\u80fd\u3092\u30b5\u30dd\u30fc\u30c8\u3059\u308b\u3001\u6709\u671b\u306a\u30d1\u30e9\u30c0\u30a4\u30e0\u3068\u3057\u3066\u767b\u5834\u3057\u305f\u3002<br><a href=\"http:\/\/arxiv.org\/abs\/2505.04921v1\">\u8ad6\u6587<\/a>\u00a0\u00a0<a href=\"https:\/\/fugumt.com\/fugumt\/paper_check\/2505.04921v1\">\u53c2\u8003\u8a33\uff08\u30e1\u30bf\u30c7\u30fc\u30bf\uff09<\/a>\u00a0 \u00a0(Thu, 08 May 2025 03:35:23 GMT)<\/li>\n\n\n\n<li>\u30ea\u30dd\u30b8\u30c8\u30ea\u306f<a href=\"https:\/\/github.com\/HITsz-TMG\/Awesome-Large-Multimodal-Reasoning-Models\">GitHub &#8211; HITsz-TMG\/Awesome-Large-Multimodal-Reasoning-Models: The development and future prospects of multimodal reasoning models.<\/a><\/li>\n<\/ul>\n","protected":false},"excerpt":{"rendered":"<p>\u4e0b\u8a18\u30b5\u30fc\u30d9\u30a4\u3082\u6ce8\u76ee<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[2],"tags":[42,232,251,517],"class_list":["post-6734","post","type-post","status-publish","format-standard","hentry","category-arxiv","tag-autonomous-agent","tag-lrm","tag-mllm","tag-517"],"_links":{"self":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/6734","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=6734"}],"version-history":[{"count":2,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/6734\/revisions"}],"predecessor-version":[{"id":6739,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/6734\/revisions\/6739"}],"wp:attachment":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=6734"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=6734"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=6734"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}