{"id":4235,"date":"2024-01-02T04:18:00","date_gmt":"2024-01-01T19:18:00","guid":{"rendered":"https:\/\/devneko.jp\/wordpress\/?p=4235"},"modified":"2024-01-02T04:18:00","modified_gmt":"2024-01-01T19:18:00","slug":"nomiracl-knowing-when-you-dont-know-for-robust-multilingual-retrieval-augmented-generation","status":"publish","type":"post","link":"https:\/\/devneko.jp\/wordpress\/?p=4235","title":{"rendered":"NoMIRACL: Knowing When You Don&#8217;t Know for Robust Multilingual Retrieval-Augmented Generation"},"content":{"rendered":"\n<ul class=\"wp-block-list\">\n<li><strong>NoMIRACL: Knowing When You Don&#8217;t Know for Robust Multilingual Retrieval-Augmented Generation&nbsp;<\/strong>[92.5]<br>Retrieval-augmented Generation (RAG) \u306f\u3001\u5916\u90e8\u306e\u77e5\u8b58\u30bd\u30fc\u30b9\u3092\u6d3b\u7528\u3057\u3066\u3001\u4e8b\u5b9f\u306e\u5e7b\u899a\u3092\u6e1b\u3089\u3059\u3053\u3068\u3067\u3001\u5927\u304d\u306a\u8a00\u8a9e\u30e2\u30c7\u30eb(LLM)\u3092\u51fa\u529b\u3059\u308b\u3002 NoMIRACL\u306f18\u8a00\u8a9e\u306b\u307e\u305f\u304c\u308bRAG\u306b\u304a\u3051\u308bLDM\u5805\u7262\u6027\u3092\u8a55\u4fa1\u3059\u308b\u305f\u3081\u306e\u4eba\u70ba\u7684\u30a2\u30ce\u30c6\u30fc\u30b7\u30e7\u30f3\u4ed8\u304d\u30c7\u30fc\u30bf\u30bb\u30c3\u30c8\u3067\u3042\u308b\u3002 \u8a55\u4fa1\u306e\u7d50\u679c,GPT-4\u306f\u30d5\u30e9\u30f3\u30b9\u8a9e\u3084\u82f1\u8a9e\u306a\u3069\u306e\u9ad8\u30ea\u30bd\u30fc\u30b9\u8a00\u8a9e\u3067\u983b\u7e41\u306b\u5e7b\u899a\u3092\u5448\u3059\u308b\u3053\u3068\u304c\u308f\u304b\u3063\u305f\u3002<br><a href=\"http:\/\/arxiv.org\/abs\/2312.11361v1\">\u8ad6\u6587<\/a>&nbsp;&nbsp;<a href=\"https:\/\/fugumt.com\/fugumt\/paper_check\/2312.11361v1\">\u53c2\u8003\u8a33\uff08\u30e1\u30bf\u30c7\u30fc\u30bf\uff09<\/a>&nbsp; &nbsp;(Mon, 18 Dec 2023 17:18:04 GMT)<\/li>\n\n\n\n<li>RAG\u306b\u304a\u3051\u308b\u9811\u5065\u6027\u306e\u30de\u30eb\u30c1\u30ea\u30f3\u30ac\u30eb\u306a\u30d9\u30f3\u30c1\u30de\u30fc\u30af\u3002\u00a0hallucination rate\u3068error rate\u304c\u30e1\u30c8\u30ea\u30af\u30b9\u3002GPT-4\u306ebase line\u304c\u3042\u308b\u304c\u300cGPT-4 achieves a high 33.2% hallucination rate on the non-relevant subset and 14.9% error rate on the relevant NoMIRACL split, highlighting that GPT-4 finds it challenging to dismiss non-relevant passages over relevant passages in first-stage retrieved information.\u300d\u3068\u5341\u5206\u3068\u306f\u8a00\u3048\u306a\u3055\u305d\u3046\u306a\u7d50\u679c\u3002\u65e5\u672c\u8a9e\u306fhallucination rate\u304c\u9ad8\u304ferror rate\u304c\u4f4e\u3044\u7d50\u679c\u3068\u306a\u3063\u3066\u3044\u308b\u3002<\/li>\n\n\n\n<li>\u30ea\u30dd\u30b8\u30c8\u30ea\u306f<a href=\"https:\/\/github.com\/project-miracl\/nomiracl\">project-miracl\/nomiracl: A multilingual dataset to evaluate LLM robustness in RAG setup against first-stage retrieval errors on 18 languages. (github.com)<\/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":[267,324,344,517],"class_list":["post-4235","post","type-post","status-publish","format-standard","hentry","category-arxiv","tag-multilingual","tag-rag","tag-robustness","tag-517"],"_links":{"self":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/4235","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=4235"}],"version-history":[{"count":0,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/4235\/revisions"}],"wp:attachment":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=4235"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=4235"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=4235"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}