{"id":5780,"date":"2024-11-18T06:07:00","date_gmt":"2024-11-17T21:07:00","guid":{"rendered":"https:\/\/devneko.jp\/wordpress\/?p=5780"},"modified":"2026-01-11T07:54:39","modified_gmt":"2026-01-10T22:54:39","slug":"spartan-sparse-transformer-world-model","status":"publish","type":"post","link":"https:\/\/devneko.jp\/wordpress\/?p=5780","title":{"rendered":"SPARTAN: SPARse TrANsformer World model"},"content":{"rendered":"\n<ul class=\"wp-block-list\">\n<li><strong>SPARTAN: A Sparse Transformer Learning Local Causation\u00a0<\/strong>[63.3]<br>\u56e0\u679c\u69cb\u9020\u306f\u3001\u74b0\u5883\u306e\u5909\u5316\u306b\u67d4\u8edf\u306b\u9069\u5fdc\u3059\u308b\u4e16\u754c\u30e2\u30c7\u30eb\u306b\u304a\u3044\u3066\u4e2d\u5fc3\u7684\u306a\u5f79\u5272\u3092\u679c\u305f\u3059\u3002 \u672c\u7814\u7a76\u3067\u306f,SPARse TrANsformer World Model(SPARTAN)\u3092\u63d0\u6848\u3059\u308b\u3002 \u30aa\u30d6\u30b8\u30a7\u30af\u30c8\u6307\u5411\u30c8\u30fc\u30af\u30f3\u9593\u306e\u6ce8\u610f\u30d1\u30bf\u30fc\u30f3\u306b\u7a7a\u9593\u898f\u5247\u3092\u9069\u7528\u3059\u308b\u3053\u3068\u3067\u3001SPARTAN\u306f\u3001\u5c06\u6765\u306e\u30aa\u30d6\u30b8\u30a7\u30af\u30c8\u72b6\u614b\u3092\u6b63\u78ba\u306b\u4e88\u6e2c\u3059\u308b\u30b9\u30d1\u30fc\u30b9\u5c40\u6240\u56e0\u679c\u30e2\u30c7\u30eb\u3092\u7279\u5b9a\u3059\u308b\u3002<br><a href=\"http:\/\/arxiv.org\/abs\/2411.06890v1\">\u8ad6\u6587<\/a>\u00a0\u00a0<a href=\"https:\/\/fugumt.com\/fugumt\/paper_check\/2411.06890v1\">\u53c2\u8003\u8a33\uff08\u30e1\u30bf\u30c7\u30fc\u30bf\uff09<\/a>\u00a0 \u00a0(Mon, 11 Nov 2024 11:42:48 GMT)<\/li>\n\n\n\n<li>\u300cConceptually, we argue that in order to perform efficient adaptation, world models should be structured to reflect the underlying sparse causal structure of the observed dynamics, and that these structures should be local.\u300d\u306e\u3082\u3068\u3001\u300cwe propose SPARTAN, a structured world model that jointly performs dynamics model learning and causal discovery.\u300d\u3068\u306e\u3053\u3068\u3002<\/li>\n<\/ul>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>Language Models as Causal Effect Generators\u00a0<\/strong>[44.8]<br>\u5236\u5fa1\u53ef\u80fd\u306a\u56e0\u679c\u69cb\u9020\u3092\u6301\u3064\u5927\u898f\u6a21\u8a00\u8a9e\u30e2\u30c7\u30eb(LLM)\u306b\u57fa\u3065\u304f\u30c7\u30fc\u30bf\u751f\u6210\u306e\u305f\u3081\u306e\u30d5\u30ec\u30fc\u30e0\u30ef\u30fc\u30af\u3092\u63d0\u6848\u3059\u308b\u3002 \u6211\u3005\u306f\u3001\u4efb\u610f\u306e\u8a00\u8a9e\u30e2\u30c7\u30eb\u3068\u6709\u5411\u975e\u5de1\u56de\u30b0\u30e9\u30d5(DAG)\u3092\u30b7\u30fc\u30b1\u30f3\u30b9\u99c6\u52d5\u69cb\u9020\u56e0\u679c\u30e2\u30c7\u30eb(SD-SCM)\u306b\u5909\u63db\u3059\u308b\u624b\u9806\u3092\u5b9a\u7fa9\u3059\u308b\u3002<br><a href=\"http:\/\/arxiv.org\/abs\/2411.08019v1\">\u8ad6\u6587<\/a>\u00a0\u00a0<a href=\"https:\/\/fugumt.com\/fugumt\/paper_check\/2411.08019v1\">\u53c2\u8003\u8a33\uff08\u30e1\u30bf\u30c7\u30fc\u30bf\uff09<\/a>\u00a0 \u00a0(Tue, 12 Nov 2024 18:50:35 GMT)<\/li>\n\n\n\n<li>\u3053\u3061\u3089\u306fLLM\uff0bDAG\u3067sequence-driven structural causal model\u3092\u4f5c\u308b\u30a2\u30d7\u30ed\u30fc\u30c1<\/li>\n<\/ul>\n\n\n\n<p>\u56e0\u679c\u30b0\u30e9\u30d5\uff0bLLM\u3068\u3044\u3046\u8a71\u306f\u3068\u3066\u3082\u8208\u5473\u6df1\u3044\u3002<\/p>\n","protected":false},"excerpt":{"rendered":"<p>\u56e0\u679c\u30b0\u30e9\u30d5\uff0bLLM\u3068\u3044\u3046\u8a71\u306f\u3068\u3066\u3082\u8208\u5473\u6df1\u3044\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":[712,223,563,564],"class_list":["post-5780","post","type-post","status-publish","format-standard","hentry","category-arxiv","tag-foresight","tag-llm","tag-563","tag-564"],"_links":{"self":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/5780","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=5780"}],"version-history":[{"count":1,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/5780\/revisions"}],"predecessor-version":[{"id":8062,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/5780\/revisions\/8062"}],"wp:attachment":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=5780"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=5780"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=5780"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}