{"id":7553,"date":"2025-10-09T05:50:00","date_gmt":"2025-10-08T20:50:00","guid":{"rendered":"https:\/\/devneko.jp\/wordpress\/?p=7553"},"modified":"2025-10-04T21:53:28","modified_gmt":"2025-10-04T12:53:28","slug":"more-than-one-teacher-adaptive-multi-guidance-policy-optimization-for-diverse-exploration","status":"publish","type":"post","link":"https:\/\/devneko.jp\/wordpress\/?p=7553","title":{"rendered":"More Than One Teacher: Adaptive Multi-Guidance Policy Optimization for Diverse Exploration"},"content":{"rendered":"\n<ul class=\"wp-block-list\">\n<li><strong>More Than One Teacher: Adaptive Multi-Guidance Policy Optimization for Diverse Exploration\u00a0<\/strong>[103.2]<br>\u30ac\u30a4\u30c0\u30f3\u30b9\u30fb\u30aa\u30f3\u30fb\u30c7\u30de\u30f3\u30c9\u300d\u30a2\u30d7\u30ed\u30fc\u30c1\u306f\u3001\u81ea\u5df1\u767a\u898b\u306e\u4fa1\u5024\u3092\u4fdd\u3061\u306a\u304c\u3089\u63a2\u7a76\u3092\u5e83\u3052\u308b\u3002 \u5b9f\u9a13\u306e\u7d50\u679c\u3001AMPO\u306f\u5f37\u3044\u30d9\u30fc\u30b9\u30e9\u30a4\u30f3\u3092\u5927\u5e45\u306b\u4e0a\u56de\u308b\u3053\u3068\u304c\u793a\u3055\u308c\u305f\u3002 \u30d4\u30a2\u30b5\u30a4\u30ba\u306e4\u4eba\u306e\u6559\u5e2b\u3092\u7528\u3044\u3066\u3001\u3088\u308a\u5f37\u529b\u306a1\u4eba\u306e\u6559\u5e2b\u3092\u6d3b\u7528\u3067\u304d\u308b\u624b\u6cd5\u306b\u5339\u6575\u3059\u308b\u7d50\u679c\u304c\u5f97\u3089\u308c\u308b\u3002<br><a href=\"http:\/\/arxiv.org\/abs\/2510.02227v1\">\u8ad6\u6587<\/a>\u00a0\u00a0<a href=\"https:\/\/fugumt.com\/fugumt\/paper_check\/2510.02227v1\">\u53c2\u8003\u8a33\uff08\u30e1\u30bf\u30c7\u30fc\u30bf\uff09<\/a>\u00a0 \u00a0(Thu, 02 Oct 2025 17:14:00 GMT)<\/li>\n\n\n\n<li>\u300cwe introduce Adaptive Multi-Guidance Policy Optimization (AMPO), a novel Mixed-Policy RL framework. Instead of relying on a single stronger teacher (e g , GPT4o or DeepSeek-R1), AMPO leverages the collective intelligence of multiple peer models. It operates on a \u201cguidance-on-demand\u201d principle: external guidance from diverse teachers replaces on-policy failures only when the student model is unable to solve a problem, thus maximizing the value of self- exploration. Furthermore, AMPO employs a comprehension-based guidance selection mechanism.\u300d\u3068\u3044\u3046\u30d5\u30ec\u30fc\u30e0\u30ef\u30fc\u30af\u306e\u63d0\u6848\u3002\u6559\u5e2b\u5074\u304c\u5f37\u529b\u306a1\u30e2\u30c7\u30eb\u3067\u306f\u306a\u304f\u3001\u8907\u6570\u306e\u5c0f\u578b\u30e2\u30c7\u30eb\u3067\u826f\u3044\u3068\u3044\u3046\u306f\u9762\u767d\u3044\u3002<\/li>\n\n\n\n<li>\u30ea\u30dd\u30b8\u30c8\u30ea\u306f<a href=\"https:\/\/github.com\/SII-Enigma\/AMPO\">GitHub &#8211; SII-Enigma\/AMPO: Official Repository of &#8220;More Than One Teacher: Adaptive Multi-Guidance Policy Optimization for Diverse Exploration&#8221;<\/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":[576],"class_list":["post-7553","post","type-post","status-publish","format-standard","hentry","category-arxiv","tag-576"],"_links":{"self":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/7553","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=7553"}],"version-history":[{"count":1,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/7553\/revisions"}],"predecessor-version":[{"id":7554,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/7553\/revisions\/7554"}],"wp:attachment":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=7553"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=7553"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=7553"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}