{"id":962712,"date":"2023-08-16T23:58:56","date_gmt":"2023-08-17T06:58:56","guid":{"rendered":"https:\/\/www.microsoft.com\/en-us\/research\/?post_type=msr-research-item&#038;p=962712"},"modified":"2024-05-13T11:11:10","modified_gmt":"2024-05-13T18:11:10","slug":"autogen-enabling-next-gen-llm-applications-via-multi-agent-conversation-framework","status":"publish","type":"msr-research-item","link":"https:\/\/www.microsoft.com\/en-us\/research\/publication\/autogen-enabling-next-gen-llm-applications-via-multi-agent-conversation-framework\/","title":{"rendered":"AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversation"},"content":{"rendered":"<p><span dir=\"ltr\" role=\"presentation\">We present<\/span> <span dir=\"ltr\" role=\"presentation\">AutoGen<\/span><span dir=\"ltr\" role=\"presentation\">, an open-source framework that allows developers to build <\/span><span dir=\"ltr\" role=\"presentation\">LLM applications via multiple<\/span> <span dir=\"ltr\" role=\"presentation\">agents<\/span> <span dir=\"ltr\" role=\"presentation\">that can converse with each other to ac<\/span><span dir=\"ltr\" role=\"presentation\">complish tasks.<\/span> <span dir=\"ltr\" role=\"presentation\">AutoGen<\/span> <span dir=\"ltr\" role=\"presentation\">agents are customizable,<\/span> <span dir=\"ltr\" role=\"presentation\">conversable<\/span><span dir=\"ltr\" role=\"presentation\">, and can operate <\/span><span dir=\"ltr\" role=\"presentation\">in various modes that employ combinations of LLMs, human inputs, and tools.<\/span><\/p>\n<p><span dir=\"ltr\" role=\"presentation\">Using<\/span> <span dir=\"ltr\" role=\"presentation\">AutoGen<\/span><span dir=\"ltr\" role=\"presentation\">, developers can also flexibly define agent interaction behaviors. <\/span><span dir=\"ltr\" role=\"presentation\">Both natural language and computer code can be used to program flexible conver<\/span><span dir=\"ltr\" role=\"presentation\">sation patterns for different applications.<\/span> <span dir=\"ltr\" role=\"presentation\">AutoGen<\/span> <span dir=\"ltr\" role=\"presentation\">serves as a generic infrastruc<\/span><span dir=\"ltr\" role=\"presentation\">ture to build diverse applications of various complexities and LLM capacities. We <\/span><span dir=\"ltr\" role=\"presentation\">provide many examples to build effective applications for domains ranging from <\/span><span dir=\"ltr\" role=\"presentation\">mathematics, coding, question answering, operations research, online decision-<\/span><span dir=\"ltr\" role=\"presentation\">making, entertainment, etc.<\/span><\/p>\n<span id=\"label-external-link\" class=\"sr-only\" aria-hidden=\"true\">Opens in a new tab<\/span>","protected":false},"excerpt":{"rendered":"<p>We present AutoGen, an open-source framework that allows developers to build LLM applications via multiple agents that can converse with each other to accomplish tasks. AutoGen agents are customizable, conversable, and can operate in various modes that employ combinations of LLMs, human inputs, and tools. Using AutoGen, developers can also flexibly define agent interaction behaviors. [&hellip;]<\/p>\n","protected":false},"featured_media":0,"template":"","meta":{"msr-url-field":"","msr-podcast-episode":"","msrModifiedDate":"","msrModifiedDateEnabled":false,"ep_exclude_from_search":false,"footnotes":""},"msr-content-type":[3],"msr-research-highlight":[246574],"research-area":[13556,13554],"msr-publication-type":[193718],"msr-product-type":[],"msr-focus-area":[],"msr-platform":[],"msr-download-source":[],"msr-field-of-study":[246694,248485],"msr-conference":[],"msr-journal":[],"msr-impact-theme":[264846],"msr-pillar":[],"msr_publishername":"","msr_edition":"","msr_affiliation":"","msr_published_date":"2023-8-16","msr_host":"","msr_duration":"","msr_version":"","msr_speaker":"","msr_other_contributors":"","msr_booktitle":"","msr_pages_string":"","msr_chapter":"","msr_isbn":"","msr_journal":"","msr_volume":"","msr_number":"MSR-TR-2023-33","msr_editors":"","msr_series":"","msr_issue":"","msr_organization":"Microsoft","msr_how_published":"","msr_notes":"","msr_highlight_text":"Best Paper Award, ICLR 2024","msr_release_tracker_id":"","msr_original_fields_of_study":"","msr_download_urls":"","msr_external_url":"","msr_secondary_video_url":"","msr_longbiography":"","msr_microsoftintellectualproperty":1,"msr_main_download":"","msr_publicationurl":"","msr_doi":"","msr_publication_uploader":[{"type":"url","viewUrl":"false","id":"false","title":"https:\/\/arxiv.org\/pdf\/2308.08155.pdf","label_id":"243109","label":0}],"msr_related_uploader":[{"type":"url","viewUrl":"false","id":"false","title":"https:\/\/microsoft.github.io\/autogen\/","label_id":"264520","label":0}],"msr_attachments":[],"msr-author-ordering":[{"type":"text","value":"Qingyun Wu","user_id":0,"rest_url":false},{"type":"user_nicename","value":"Gagan Bansal","user_id":41707,"rest_url":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=Gagan Bansal"},{"type":"text","value":"Jieyu Zhang","user_id":0,"rest_url":false},{"type":"text","value":"Yiran Wu","user_id":0,"rest_url":false},{"type":"text","value":"Shaokun Zhang","user_id":0,"rest_url":false},{"type":"user_nicename","value":"Erkang (Eric) Zhu","user_id":38718,"rest_url":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=Erkang (Eric) Zhu"},{"type":"user_nicename","value":"Beibin Li","user_id":41835,"rest_url":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=Beibin Li"},{"type":"text","value":"Li Jiang","user_id":0,"rest_url":false},{"type":"text","value":"Xiaoyun Zhang","user_id":0,"rest_url":false},{"type":"user_nicename","value":"Chi Wang","user_id":31406,"rest_url":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=Chi Wang"}],"msr_impact_theme":["Computing foundations"],"msr_research_lab":[992148],"msr_event":[1014303],"msr_group":[781564],"msr_project":[973047],"publication":[],"video":[],"download":[970209],"msr_publication_type":"techreport","_links":{"self":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/962712"}],"collection":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item"}],"about":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/types\/msr-research-item"}],"version-history":[{"count":4,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/962712\/revisions"}],"predecessor-version":[{"id":1033590,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/962712\/revisions\/1033590"}],"wp:attachment":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/media?parent=962712"}],"wp:term":[{"taxonomy":"msr-content-type","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-content-type?post=962712"},{"taxonomy":"msr-research-highlight","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-highlight?post=962712"},{"taxonomy":"msr-research-area","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/research-area?post=962712"},{"taxonomy":"msr-publication-type","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-publication-type?post=962712"},{"taxonomy":"msr-product-type","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-product-type?post=962712"},{"taxonomy":"msr-focus-area","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-focus-area?post=962712"},{"taxonomy":"msr-platform","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-platform?post=962712"},{"taxonomy":"msr-download-source","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-download-source?post=962712"},{"taxonomy":"msr-field-of-study","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-field-of-study?post=962712"},{"taxonomy":"msr-conference","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-conference?post=962712"},{"taxonomy":"msr-journal","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-journal?post=962712"},{"taxonomy":"msr-impact-theme","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-impact-theme?post=962712"},{"taxonomy":"msr-pillar","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-pillar?post=962712"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}