{"id":8293,"date":"2026-03-03T06:43:00","date_gmt":"2026-03-02T21:43:00","guid":{"rendered":"https:\/\/devneko.jp\/wordpress\/?p=8293"},"modified":"2026-02-28T15:50:38","modified_gmt":"2026-02-28T06:50:38","slug":"dream-deep-research-evaluation-with-agentic-metrics","status":"publish","type":"post","link":"https:\/\/devneko.jp\/wordpress\/?p=8293","title":{"rendered":"DREAM: Deep Research Evaluation with Agentic Metrics\u00a0"},"content":{"rendered":"\n<ul class=\"wp-block-list\">\n<li><strong>DREAM: Deep Research Evaluation with Agentic Metrics\u00a0<\/strong>[21.6]<br>\u672c\u7a3f\u3067\u306f,DREAM(Deep Research Evaluation with Agentic Metrics)\u3092\u63d0\u6848\u3059\u308b\u3002 DREAM\u69cb\u9020\u8a55\u4fa1\u306f\u3001\u30af\u30a8\u30ea\u975e\u4f9d\u5b58\u306e\u30e1\u30c8\u30ea\u30af\u30b9\u3068\u30c4\u30fc\u30eb\u547c\u3073\u51fa\u3057\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u304c\u751f\u6210\u3059\u308b\u9069\u5fdc\u7684\u306a\u30e1\u30c8\u30ea\u30af\u30b9\u3092\u7d44\u307f\u5408\u308f\u305b\u305f\u8a55\u4fa1\u30d7\u30ed\u30c8\u30b3\u30eb\u3092\u7528\u3044\u3066\u884c\u308f\u308c\u308b\u3002 \u5236\u5fa1\u3055\u308c\u305f\u8a55\u4fa1\u306f\u3001DREAM\u304c\u65e2\u5b58\u306e\u30d9\u30f3\u30c1\u30de\u30fc\u30af\u3088\u308a\u3082\u4e8b\u5b9f\u3084\u6642\u9593\u7684\u52a3\u5316\u306b\u304b\u306a\u308a\u654f\u611f\u3067\u3042\u308b\u3053\u3068\u3092\u793a\u3057\u3066\u3044\u308b\u3002<br><a href=\"http:\/\/arxiv.org\/abs\/2602.18940v1\">\u8ad6\u6587<\/a>\u00a0\u00a0<a href=\"https:\/\/fugumt.com\/fugumt\/paper_check\/2602.18940v1\">\u53c2\u8003\u8a33\uff08\u30e1\u30bf\u30c7\u30fc\u30bf\uff09<\/a>\u00a0 \u00a0(Sat, 21 Feb 2026 19:14:31 GMT)<\/li>\n\n\n\n<li>\u300cDREAM structures assessment through an evaluation protocol combining query-agnostic metrics with adaptive metrics generated by a tool-calling agent, enabling temporally aware coverage, grounded verification, and systematic reasoning probes.\u300d\u3068\u30d5\u30a1\u30af\u30c8\u30c1\u30a7\u30c3\u30af\u3092\u601d\u3044\u51fa\u3059\u8a55\u4fa1\u30a2\u30d7\u30ed\u30fc\u30c1\u3002\u300cWe demonstrate that current LLM-as-a-judge and reference-based benchmarks are often blinded by surface-level fluency and citation alignment, failing to detect deep-seated defects in factual correctness, temporal validity, and logical reasoning.\u300d\u306f\u305d\u3046\u3060\u308d\u3046\u3068\u601d\u3046\u3002<\/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":[648],"class_list":["post-8293","post","type-post","status-publish","format-standard","hentry","category-arxiv","tag-648"],"_links":{"self":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/8293","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=8293"}],"version-history":[{"count":1,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/8293\/revisions"}],"predecessor-version":[{"id":8294,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/8293\/revisions\/8294"}],"wp:attachment":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=8293"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=8293"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=8293"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}