{"id":6405,"date":"2025-03-10T03:53:00","date_gmt":"2025-03-09T18:53:00","guid":{"rendered":"https:\/\/devneko.jp\/wordpress\/?p=6405"},"modified":"2025-03-10T03:53:00","modified_gmt":"2025-03-09T18:53:00","slug":"predictive-data-selection-the-data-that-predicts-is-the-data-that-teaches","status":"publish","type":"post","link":"https:\/\/devneko.jp\/wordpress\/?p=6405","title":{"rendered":"Predictive Data Selection: The Data That Predicts Is the Data That Teaches"},"content":{"rendered":"\n<ul class=\"wp-block-list\">\n<li><strong>Predictive Data Selection: The Data That Predicts Is the Data That Teaches\u00a0<\/strong>[19.0]<br>\u4e88\u6e2c\u30c7\u30fc\u30bf\u9078\u629e(PreSelect)\u306f,\u9ad8\u901f\u30c6\u30ad\u30b9\u30c8\u30d9\u30fc\u30b9\u306e\u30b9\u30b3\u30a2\u30e9\u306e\u307f\u306e\u30c8\u30ec\u30fc\u30cb\u30f3\u30b0\u3068\u30c7\u30d7\u30ed\u30a4\u3092\u5fc5\u8981\u3068\u3059\u308b\u8efd\u91cf\u3067\u52b9\u7387\u7684\u306a\u30c7\u30fc\u30bf\u9078\u629e\u624b\u6cd5\u3067\u3042\u308b\u3002 \u6211\u3005\u306f\u3001PreSelect\u3067\u9078\u629e\u3055\u308c\u305f30B\u30c8\u30fc\u30af\u30f3\u3067\u30c8\u30ec\u30fc\u30cb\u30f3\u30b0\u3055\u308c\u305f\u30e2\u30c7\u30eb\u304c300B\u30c8\u30fc\u30af\u30f3\u3067\u30c8\u30ec\u30fc\u30cb\u30f3\u30b0\u3055\u308c\u305f\u30d0\u30cb\u30e9\u30d9\u30fc\u30b9\u30e9\u30a4\u30f3\u306e\u30d1\u30d5\u30a9\u30fc\u30de\u30f3\u30b9\u3092\u4e0a\u56de\u308b\u3053\u3068\u3092\u793a\u3057\u305f\u3002<br><a href=\"http:\/\/arxiv.org\/abs\/2503.00808v2\">\u8ad6\u6587<\/a>\u00a0\u00a0<a href=\"https:\/\/fugumt.com\/fugumt\/paper_check\/2503.00808v2\">\u53c2\u8003\u8a33\uff08\u30e1\u30bf\u30c7\u30fc\u30bf\uff09<\/a>\u00a0 \u00a0(Tue, 04 Mar 2025 06:15:27 GMT)<\/li>\n\n\n\n<li>\u300cBuilding on this observation, we hypothesize that data on which model losses are predictive of downstream abilities also contribute effectively to learning.\u300d\u3068\u3044\u3046\u4eee\u5b9a\u306e\u4e0b\u8a2d\u8a08\u3055\u308c\u305f\u30c7\u30fc\u30bf\u9078\u629e\u624b\u6cd5PRESELECT\u306e\u63d0\u6848\u3002\u300cPRESELECT demonstrates remarkable performance, with an average absolute improvement of 2.8% over the random selection and 20% gains in Math and Code raw text BPC, which shows a promising trend.\u300d\u3068\u52b9\u679c\u3092\u4e3b\u5f35\u3002<\/li>\n\n\n\n<li>\u30ea\u30dd\u30b8\u30c8\u30ea\u306f<a href=\"https:\/\/github.com\/hkust-nlp\/PreSelect\">GitHub &#8211; hkust-nlp\/PreSelect<\/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":[101,102],"class_list":["post-6405","post","type-post","status-publish","format-standard","hentry","category-arxiv","tag-dataset-condensation","tag-dataset-distillation"],"_links":{"self":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/6405","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=6405"}],"version-history":[{"count":0,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/6405\/revisions"}],"wp:attachment":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=6405"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=6405"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=6405"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}