{"id":6920,"date":"2025-06-23T04:29:00","date_gmt":"2025-06-22T19:29:00","guid":{"rendered":"https:\/\/devneko.jp\/wordpress\/?p=6920"},"modified":"2025-06-14T21:32:07","modified_gmt":"2025-06-14T12:32:07","slug":"scientists-first-exam-probing-cognitive-abilities-of-mllm-via-perception-understanding-and-reasoning","status":"publish","type":"post","link":"https:\/\/devneko.jp\/wordpress\/?p=6920","title":{"rendered":"Scientists&#8217; First Exam: Probing Cognitive Abilities of MLLM via Perception, Understanding, and Reasoning\u00a0"},"content":{"rendered":"\n<ul class=\"wp-block-list\">\n<li><strong>Scientists&#8217; First Exam: Probing Cognitive Abilities of MLLM via Perception, Understanding, and Reasoning\u00a0<\/strong>[59.5]<br>\u6211\u3005\u306f,Multimodal Large Language Models (MLLM) \u306e\u79d1\u5b66\u7684\u8a8d\u77e5\u80fd\u529b\u3092\u8a55\u4fa1\u3059\u308b\u305f\u3081\u306b\u8a2d\u8a08\u3055\u308c\u305f,Scientists&#8217; First Exam (SFE) \u30d9\u30f3\u30c1\u30de\u30fc\u30af\u3092\u63d0\u793a\u3059\u308b\u3002 SFE\u306f3\u3064\u306e\u8cea\u554f\u30bf\u30a4\u30d7\u306b\u307e\u305f\u304c\u308b830\u306e\u30a8\u30ad\u30b9\u30d1\u30fc\u30c8\u691c\u8a3cVQA\u30da\u30a2\u3067\u69cb\u6210\u3055\u308c\u30015\u3064\u306e\u9ad8\u4fa1\u5024\u5206\u91ce\u306b\u307e\u305f\u304c\u308b66\u306e\u30de\u30eb\u30c1\u30e2\u30fc\u30c0\u30eb\u30bf\u30b9\u30af\u306b\u307e\u305f\u304c\u308b\u3002 \u5b9f\u9a13\u306e\u7d50\u679c\u3001\u73fe\u5728\u6700\u5148\u7aef\u306eGPT-o3\u3068InternVL-3\u306fSFE\u3067\u308f\u305a\u304b34.08%\u306826.52%\u3057\u304b\u9054\u6210\u3067\u304d\u305a\u3001MLLM\u304c\u79d1\u5b66\u9818\u57df\u3067\u6539\u5584\u3059\u308b\u4f59\u5730\u304c\u3042\u308b\u3053\u3068\u304c\u660e\u3089\u304b\u306b\u306a\u3063\u305f\u3002<br><a href=\"http:\/\/arxiv.org\/abs\/2506.10521v1\">\u8ad6\u6587<\/a>\u00a0\u00a0<a href=\"https:\/\/fugumt.com\/fugumt\/paper_check\/2506.10521v1\">\u53c2\u8003\u8a33\uff08\u30e1\u30bf\u30c7\u30fc\u30bf\uff09<\/a>\u00a0 \u00a0(Thu, 12 Jun 2025 09:29:16 GMT)<\/li>\n\n\n\n<li>\u300cwe introduce the Scientists\u2019 First Exam (SFE) benchmark, designed to comprehensively evaluate the scientific cognitive capabilities of MLLMs through three cognitive levels (cog-levels): Scientific Signal Perception (L1) characterizes the capacity to discern critical components within visualizations of scientific raw data; Scientific Attribute Understanding (L2) demonstrates the ability to interpret domain-expert knowledge; Scientific Comparative Reasoning (L3) manifests the ability to derive phenomenological insights through structured comparison of multiple scientific visual sources. SFE encompasses 66 expert-curated, high-value multimodal tasks across five disciplines: Astronomy, Chemistry, Earth, Life, and Materials Sciences (Fig. 1b).\u300d\u3068\u3044\u3046\u30d9\u30f3\u30c1\u30de\u30fc\u30af\u3002MLLM\u5411\u3051\u3001VQA\u3068\u3057\u3066\u69cb\u6210\u3055\u308c\u3066\u3044\u308b\u3002<\/li>\n\n\n\n<li>\u30ea\u30dd\u30b8\u30c8\u30ea\u306f<a href=\"https:\/\/huggingface.co\/datasets\/PrismaX\/SFE\">PrismaX\/SFE \u00b7 Datasets at Hugging Face<\/a>\u3001\u30d7\u30ed\u30b8\u30a7\u30af\u30c8\u30b5\u30a4\u30c8\u306f<a href=\"https:\/\/prismax.opencompass.org.cn\/\">PrismaX<\/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":[251,441,517,623],"class_list":["post-6920","post","type-post","status-publish","format-standard","hentry","category-arxiv","tag-mllm","tag-vqa","tag-517","tag-623"],"_links":{"self":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/6920","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=6920"}],"version-history":[{"count":1,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/6920\/revisions"}],"predecessor-version":[{"id":6921,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/6920\/revisions\/6921"}],"wp:attachment":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=6920"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=6920"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=6920"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}