{"id":4208,"date":"2023-12-26T05:37:00","date_gmt":"2023-12-25T20:37:00","guid":{"rendered":"https:\/\/devneko.jp\/wordpress\/?p=4208"},"modified":"2023-12-26T05:37:00","modified_gmt":"2023-12-25T20:37:00","slug":"gemini%e3%81%ae%e8%a9%95%e4%be%a1","status":"publish","type":"post","link":"https:\/\/devneko.jp\/wordpress\/?p=4208","title":{"rendered":"Gemini\u306e\u8a55\u4fa1"},"content":{"rendered":"\n<p>Gemini\u306e\u8a55\u4fa1\u306b\u95a2\u3059\u308b\u8ad6\u6587\u304c\u51fa\u3066\u3044\u308b\u3002<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>An In-depth Look at Gemini&#8217;s Language Abilities\u00a0<\/strong>[49.9]<br>OpenAI GPT\u3068Google Gemini\u30e2\u30c7\u30eb\u306e\u80fd\u529b\u3092\u6bd4\u8f03\u3059\u308b\u3002 \u3053\u306e\u5206\u6790\u306f\u3001\u3055\u307e\u3056\u307e\u306a\u8a00\u8a9e\u80fd\u529b\u3092\u30c6\u30b9\u30c8\u3059\u308b10\u306e\u30c7\u30fc\u30bf\u30bb\u30c3\u30c8\u306b\u5bfe\u3057\u3066\u5b9f\u65bd\u3057\u307e\u3059\u3002 Gemini Pro \u306f GPT 3.5 Turbo \u3088\u308a\u3082\u8fd1\u3044\u304c\u308f\u305a\u304b\u306b\u52a3\u308b\u7cbe\u5ea6\u3092\u5b9f\u73fe\u3057\u3066\u3044\u308b\u3002<br><a href=\"http:\/\/arxiv.org\/abs\/2312.11444v1\">\u8ad6\u6587<\/a>\u00a0\u00a0<a href=\"https:\/\/fugumt.com\/fugumt\/paper_check\/2312.11444v1\">\u53c2\u8003\u8a33\uff08\u30e1\u30bf\u30c7\u30fc\u30bf\uff09<\/a>\u00a0 \u00a0(Mon, 18 Dec 2023 18:47:42 GMT)<\/li>\n\n\n\n<li>Gemini Pro\u306b\u5bfe\u3059\u308b\u4e3b\u3068\u3057\u3066\u8a00\u8a9e\u80fd\u529b\u306e\u8a55\u4fa1\u3002\u300cwe find that Gemini Pro achieves accuracy that is close but slightly inferior to the corresponding GPT 3.5 Turbo on all tasks that we benchmarked.\u300d\u3068\u306e\u3053\u3068\u3002Gemini Pro\u306fGPT-3.5\u3068\u7af6\u5408\u7684\u3001GPT-4\u3068\u6bd4\u3079\u3089\u308c\u3066\u3044\u305f\u306e\u306f\u4e3b\u306bGemini Ultra\u306a\u306e\u3067\u7d50\u679c\u306b\u9055\u548c\u611f\u306f\u306a\u3044\u3002<\/li>\n\n\n\n<li>\u30ea\u30dd\u30b8\u30c8\u30ea\u306f<a href=\"https:\/\/github.com\/neulab\/gemini-benchmark\">GitHub &#8211; neulab\/gemini-benchmark<\/a><\/li>\n<\/ul>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>A Challenger to GPT-4V? Early Explorations of Gemini in Visual Expertise\u00a0<\/strong>[78.5]<br>Gemini\u306fGoogle\u306e\u6700\u65b0\u304b\u3064\u6700\u3082\u6709\u80fd\u306aMLLM\u3067\u3001\u30de\u30eb\u30c1\u30e2\u30c0\u30ea\u30c6\u30a3\u306e\u305f\u3081\u306b\u30bc\u30ed\u304b\u3089\u69cb\u7bc9\u3055\u308c\u3066\u3044\u307e\u3059\u3002 Gemini\u306f\u30de\u30eb\u30c1\u30e2\u30fc\u30c0\u30eb\u5b66\u7fd2\u306b\u304a\u3051\u308bGPT-4V\u306e\u30ea\u30fc\u30c9\u30dd\u30b8\u30b7\u30e7\u30f3\u306b\u6311\u6226\u3067\u304d\u308b\u304b? Gemini Pro\u3068\u6700\u5148\u7aef\u306eGPT-4V\u3092\u6bd4\u8f03\u3057\u3066\u3001\u6700\u65b0\u306e\u30aa\u30fc\u30d7\u30f3\u30bd\u30fc\u30b9MLLM\u3067\u3042\u308bSphinx\u3068\u3068\u3082\u306b\u3001\u305d\u306e\u4e0a\u9650\u3092\u8a55\u4fa1\u3059\u308b\u3002<br><a href=\"http:\/\/arxiv.org\/abs\/2312.12436v2\">\u8ad6\u6587<\/a>\u00a0\u00a0<a href=\"https:\/\/fugumt.com\/fugumt\/paper_check\/2312.12436v2\">\u53c2\u8003\u8a33\uff08\u30e1\u30bf\u30c7\u30fc\u30bf\uff09<\/a>\u00a0 \u00a0(Wed, 20 Dec 2023 12:40:47 GMT)<\/li>\n\n\n\n<li>\u3053\u3061\u3089\u306f\u30de\u30eb\u30c1\u30e2\u30fc\u30c0\u30eb\u3067\u306e\u8a55\u4fa1\u3002\u6bd4\u8f03\u5bfe\u8c61\u306f\u4e0a\u8a18\u3068\u540c\u3058\u3067Gemini Pro\u3060\u3067\u3042\u308b\u3053\u3068\u306b\u8981\u6ce8\u610f\u3002\u300cThe qualitative results indicate that Gemini is indeed a strong challenger to GPT-4V, given its superior multi-modal reasoning capacity.\u300d\u3068\u8a55\u4fa1<\/li>\n\n\n\n<li>\u30ea\u30dd\u30b8\u30c8\u30ea\u306f<a href=\"https:\/\/github.com\/BradyFU\/Awesome-Multimodal-Large-Language-Models\">GitHub &#8211; BradyFU\/Awesome-Multimodal-Large-Language-Models: :sparkles::sparkles:Latest Papers and Datasets on Multimodal Large Language Models, and Their Evaluation.<\/a><\/li>\n<\/ul>\n","protected":false},"excerpt":{"rendered":"<p>Gemini\u306e\u8a55\u4fa1\u306b\u95a2\u3059\u308b\u8ad6\u6587\u304c\u51fa\u3066\u3044\u308b\u3002<\/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":[166,176],"class_list":["post-4208","post","type-post","status-publish","format-standard","hentry","category-arxiv","tag-gemini","tag-gpt-4v"],"_links":{"self":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/4208","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=4208"}],"version-history":[{"count":0,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/4208\/revisions"}],"wp:attachment":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=4208"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=4208"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=4208"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}