{"id":5398,"date":"2024-09-09T02:35:00","date_gmt":"2024-09-08T17:35:00","guid":{"rendered":"https:\/\/devneko.jp\/wordpress\/?p=5398"},"modified":"2024-09-09T02:35:00","modified_gmt":"2024-09-08T17:35:00","slug":"xlam-a-family-of-large-action-models-to-empower-ai-agent-systems","status":"publish","type":"post","link":"https:\/\/devneko.jp\/wordpress\/?p=5398","title":{"rendered":"xLAM: A Family of Large Action Models to Empower AI Agent Systems \/ ToolACE: Winning the Points of LLM Function Calling"},"content":{"rendered":"\n<ul class=\"wp-block-list\">\n<li><strong>xLAM: A Family of Large Action Models to Empower AI Agent Systems&nbsp;<\/strong>[111.6]<br>AI\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u30bf\u30b9\u30af\u7528\u306b\u8a2d\u8a08\u3055\u308c\u305f\u5927\u898f\u6a21\u306a\u30a2\u30af\u30b7\u30e7\u30f3\u30e2\u30c7\u30eb\u3067\u3042\u308bxLAM\u3092\u30ea\u30ea\u30fc\u30b9\u3059\u308b\u3002 xLAM\u306f\u3001\u8907\u6570\u306e\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u80fd\u529b\u30d9\u30f3\u30c1\u30de\u30fc\u30af\u3067\u4f8b\u5916\u7684\u306a\u30d1\u30d5\u30a9\u30fc\u30de\u30f3\u30b9\u3092\u63d0\u4f9b\u3059\u308b\u3002<br><a href=\"http:\/\/arxiv.org\/abs\/2409.03215v1\">\u8ad6\u6587<\/a>&nbsp;&nbsp;<a href=\"https:\/\/fugumt.com\/fugumt\/paper_check\/2409.03215v1\">\u53c2\u8003\u8a33\uff08\u30e1\u30bf\u30c7\u30fc\u30bf\uff09<\/a>&nbsp; &nbsp;(Thu, 05 Sep 2024 03:22:22 GMT)<\/li>\n\n\n\n<li>Salesforce AI Research\u306b\u3088\u308b\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u52d5\u4f5c\u306b\u9069\u3057\u305f\u30e2\u30c7\u30eb\u306e\u63d0\u6848\u3002\u30c7\u30fc\u30bf\u30bb\u30c3\u30c8\u306e\u7d71\u5408\u30fb\u62e1\u5f35\u3067\u5408\u6210\u30c7\u30fc\u30bf\u95a2\u9023\u306e\u624b\u6cd5\u3092\u3046\u307e\u304f\u6d3b\u7528\u3057\u3066\u3044\u308b\u3002\u30bd\u30fc\u30b9\u30b3\u30fc\u30c9\u306fApache-2\u30e9\u30a4\u30bb\u30f3\u30b9\u3002\u30e2\u30c7\u30eb\u306f\u516c\u958b\u3055\u308c\u3066\u3044\u308b\u304c\u5546\u7528\u5229\u7528\u4e0d\u53ef\u306eCC-BY-NC\u3002\u6027\u80fd\u306f\u300cOur experimental results demonstrate that xLAM consistently delivers exceptional performance across multiple agent ability benchmarks, notably securing the 1st position on the Berkeley Function-Calling Leaderboard, outperforming GPT-4, Claude-3, and many other models in terms of tool use.\u300d\u3068\u306e\u3053\u3068\u3002\u300cThe insights we learned from training these models highlight the importance of rigorous data processing and the potential of data synthesis in developing capable AI agents.\u300d\u3068\u3044\u3046\u8a18\u8f09\u3082\u3042\u308a\u3001\u5408\u6210\u30c7\u30fc\u30bf\u306e\u6d3b\u7528\u306b\u3064\u3044\u3066\u91cd\u8981\u6027\u304c\u4e0a\u304c\u3063\u3066\u3044\u308b\u3088\u3046\u306b\u898b\u3048\u308b\u3002<\/li>\n\n\n\n<li>\u30ea\u30dd\u30b8\u30c8\u30ea\u306f<a href=\"https:\/\/github.com\/SalesforceAIResearch\/xLAM\">GitHub &#8211; SalesforceAIResearch\/xLAM<\/a>\u3001<a href=\"https:\/\/huggingface.co\/collections\/Salesforce\/xlam-models-65f00e2a0a63bbcd1c2dade4\">xLAM models &#8211; a Salesforce Collection (huggingface.co)<\/a> <\/li>\n<\/ul>\n\n\n\n<p>Berkeley Function-Calling Leaderboard\u306b\u3064\u3044\u3066\u306f\u4e0b\u8a18\u8ad6\u6587\u3082\u767a\u8868\u3055\u308c\u3066\u3044\u308b\u3002\u3053\u3061\u3089\u3082\u5408\u6210\u30c7\u30fc\u30bf\u3092\u7528\u3044\u308b\u30a2\u30d7\u30ed\u30fc\u30c1<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>ToolACE: Winning the Points of LLM Function Calling\u00a0<\/strong>[139.1]<br>ToolACE\u306f\u3001\u6b63\u78ba\u3067\u8907\u96d1\u3067\u591a\u69d8\u306a\u30c4\u30fc\u30eb\u5b66\u7fd2\u30c7\u30fc\u30bf\u3092\u751f\u6210\u3059\u308b\u3088\u3046\u306b\u8a2d\u8a08\u3055\u308c\u305f\u81ea\u52d5\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u30d1\u30a4\u30d7\u30e9\u30a4\u30f3\u3067\u3042\u308b\u3002 \u6211\u3005\u306f\u3001\u5408\u6210\u30c7\u30fc\u30bf\u306b\u57fa\u3065\u3044\u3066\u30c8\u30ec\u30fc\u30cb\u30f3\u30b0\u3055\u308c\u305f\u30e2\u30c7\u30eb\u304c\u30018B\u30d1\u30e9\u30e1\u30fc\u30bf\u3060\u3051\u3067\u3001\u30d0\u30fc\u30af\u30ec\u30fc\u30fb\u30d5\u30a1\u30f3\u30af\u30b7\u30e7\u30f3\u30fb\u30ab\u30ea\u30f3\u30b0\u30fb\u30ea\u30fc\u30c0\u30fc\u30dc\u30fc\u30c9\u4e0a\u3067\u6700\u5148\u7aef\u306e\u30d1\u30d5\u30a9\u30fc\u30de\u30f3\u30b9\u3092\u9054\u6210\u3059\u308b\u3053\u3068\u3092\u5b9f\u8a3c\u3057\u305f\u3002<br><a href=\"http:\/\/arxiv.org\/abs\/2409.00920v1\">\u8ad6\u6587<\/a>\u00a0\u00a0<a href=\"https:\/\/fugumt.com\/fugumt\/paper_check\/2409.00920v1\">\u53c2\u8003\u8a33\uff08\u30e1\u30bf\u30c7\u30fc\u30bf\uff09<\/a>\u00a0 \u00a0(Mon, 02 Sep 2024 03:19:56 GMT)<\/li>\n\n\n\n<li>the Berkeley Function-Calling Leaderboard\u3078\u306e\u300cTool Self-evolution Synthesis (TSS), Multi-Agent Interactive Dialog Generation (MAI), and Dual-Layer Validation Process (DLV).\u300d\u304b\u3089\u306a\u308b\u30d1\u30a4\u30d7\u30e9\u30a4\u30f3\u69cb\u6210\uff08Agentic\u306a\uff09\u30c7\u30fc\u30bf\u5408\u6210\u306b\u3088\u308b\u5bfe\u5fdc<\/li>\n\n\n\n<li>\u30ea\u30dd\u30b8\u30c8\u30ea\u306f<a href=\"https:\/\/huggingface.co\/Team-ACE\">Team-ACE (Team-ACE) (huggingface.co)<\/a><\/li>\n<\/ul>\n","protected":false},"excerpt":{"rendered":"<p>Berkeley Function-Calling Leaderboard\u306b\u3064\u3044\u3066\u306f\u4e0b\u8a18\u8ad6\u6587\u3082\u767a\u8868\u3055\u308c\u3066\u3044\u308b\u3002\u3053\u3061\u3089\u3082\u5408\u6210\u30c7\u30fc\u30bf\u3092\u7528\u3044\u308b\u30a2\u30d7\u30ed\u30fc\u30c1<\/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":[42,390],"class_list":["post-5398","post","type-post","status-publish","format-standard","hentry","category-arxiv","tag-autonomous-agent","tag-synthetic-data"],"_links":{"self":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/5398","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=5398"}],"version-history":[{"count":0,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/5398\/revisions"}],"wp:attachment":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=5398"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=5398"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=5398"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}