Overview
Musanganiswa-we-maAgents (MoA) inzira iyo mhando dzemitauro yakati wandei dzinonyora mhinduro uyezve aggregator modhi inosanganisa mazano avo epamusoro kuita mhinduro imwe yakavandudzwa. Inobvumira timu yemhando dzakavhurika kukwikwidza kana kurova imwe yepamusoro-tier modhi.
Musanganiswa-we-Agents Aggregation chikamu chemutauro-AI stack inoshandiswa kuverenga, kugadzira, kuronga, uye kushandura zvinyorwa uye kutaura pamwero.
Deep Dive
Yakaunzwa mubepa ra2024 kubva Together AI, Musanganiswa-we-Agents unoronga akawanda maLLM muzvikamu. Muchikamu chekutanga, akawanda 'proposer' modhi imwe neimwe yakazvimirira inopindura kukurumidza. Zvavakaburitsa zvinobva zvabatanidzwa uye zvinopihwa kune inotevera dhizaini, uko modhi zvakare inopindura, ikozvino yakarongedzwa pane ese apfuura madhizaini. Mushure menguva imwe chete kana akawanda akadai, modhi ye 'aggregator' inogadzirisa zvese kuita mhinduro imwechete. Muono mukuru, unonzi nevanyori 'kubatana kweLLMs', ndeyekuti mhando dzinoburitsa mhinduro dziri nani kana dzichiratidzwa mhinduro dzevezera, kunyangwe dzisina kukwana. Pabhenji reAlpacaEval 2.0, iyo MoA yakavakwa zvachose kubva kune yakavhurika-sosi modhi inonzi yakapfuura GPT-4 Omni chibodzwa, zvichiratidza kuti nekungwarira kuunganidzwa kweakasiyana, akachipa mamodheru anogona kurova imwe muganhu system.
Technical Insight
MoA inosiyana nekuvhota kwakapfava kwevazhinji: pane kutora mhinduro imwe, muunganidzi anoverenga mhinduro dzese dzevamiriri sezviri mukati uye inogadzira synthesis nyowani, kusanganisa masimba uye zvikanganiso zvekusefa. Kusiyana pakati pevanokurudzira kunobatsira, saka kusanganisa mhuri dzemhando dzakasiyana kwakakosha. Mamiriro acho akaturikidzana, senge network yakadzika apo imwe neimwe 'neurons' izere LLM inofona. Kutengeserana-kubvisa ndeye latency uye mutengo: imwe neimwe layer inowanza nhamba yemafoni ekufungidzira, saka MoA inoshandisa yakawanda compute kusimudza mhando.
Mastering Musanganiswa-we-Agents Aggregation
Musanganiswa-we-maAgents (MoA) inzira iyo mhando dzemitauro yakati wandei dzinonyora mhinduro uyezve aggregator modhi inosanganisa mazano avo epamusoro kuita mhinduro imwe yakavandudzwa. Inobvumira timu yemhando dzakavhurika kukwikwidza kana kurova imwe yepamusoro-tier modhi. Musanganiswa-we-Agents Aggregation chikamu chemutauro-AI stack inoshandiswa kuverenga, kugadzira, kuronga, uye kushandura zvinyorwa uye kutaura pamwero. Kuti uvake kunzwisisa kwakadzama, tora Musanganiswa-we-Agents Aggregation semuenzaniso wekushandisa, kwete chinhu chimwe chete: tsanangura zvinodiwa, kujekesa fungidziro, uye patsanura izvo system inogona kuita nekuvimbika kubva kune ichiri kuda kutonga kwenyanzvi.
Mukuita, zvikwata zvakasimba zvinoshandisa Mixture-of-Agents Aggregation dhizaini zvinokurudzira, kudzoreredza, uye kuongorora zvishwe seimwe yakabatanidzwa yekutaurirana system. Ivo vanonyora zvakajeka maitiro ebudiriro, bvunzo vachipokana ne data rechokwadi uye mafambiro ebasa, uye iterate zvichibva pane zvakacherechedzwa maitiro ekutadza kwete kuhwina-nguva imwe chete yebhenji. Apa ndipo apo kunzwisisa kwe theoretical kunoshanduka kuve kugona kwakasimba pane chigadzirwa, mutemo, uye mashandiro.
Mutauro workflows inogona kufamba nekukurumidza pasina kupira kuenderana. Panguva imwecheteyo, chokwadi cheHallucified chinogona kupinda chinyararire mishumo, kuyerera kwetsigiro, kana kutsvagisa zvinobuda. Nzira yakatsiga ndeyekubatanidza kukurumidza kuyedza nekutonga: mhanyisa vatyairi vendege, tora humbowo, buritsa matanda esarudzo, uye urambe uchivandudza chengetedzo semaitiro emuenzaniso, zvinotarisirwa nemushandisi, uye zvinodikanwa zvekutonga.
Strategic Impact
Mutauro workflows inogona kufamba nekukurumidza pasina kupira kuenderana.
Mutauro workflows inogona kufamba nekukurumidza pasina kupira kuenderana. Mukutumirwa kwemhando yepamusoro, izvi zvinoshandurirwa kuita mitemo inoyerwa yekushanda, miganhu yevaridzi, uye tsika dzekudzokorora dzinodzokororwa kuitira kuti zvikwata zvikwire kuvimba pane kukwidza kusajeka.
Inopamhidzira kupinda mumitauro yese nemataera ekutaurirana.
Inopamhidzira kupinda mumitauro yese nemataera ekutaurirana. Mukutumirwa kwemhando yepamusoro, izvi zvinoshandurirwa kuita mitemo inoyerwa yekushanda, miganhu yevaridzi, uye tsika dzekudzokorora dzinodzokororwa kuitira kuti zvikwata zvikwire kuvimba pane kukwidza kusajeka.
Zvikwata zvinogona kupedza nguva yakawanda pakutonga uku otomatiki ichibata kudzokorora.
Zvikwata zvinogona kupedza nguva yakawanda pakutonga uku otomatiki ichibata kudzokorora. Mukutumirwa kwemhando yepamusoro, izvi zvinoshandurirwa kuita mitemo inoyerwa yekushanda, miganhu yevaridzi, uye tsika dzekudzokorora dzinodzokororwa kuitira kuti zvikwata zvikwire kuvimba pane kukwidza kusajeka.
Real-World Implementation
Kubatanidza matatu akasiyana eakavhurika machat modhi sevamiriri, wozoshandisa yakasimba aggregator kuburitsa imwe yakakwenenzverwa-tsigiro yemutengi.
Kuwedzera kuraira-kutevera zvibodzwa paAlpacaEval-maitiro mabhenji uchishandisa chete akavhurika-sosi mhando.
Kusanganisa akasiyana kodhi mazano kubva kune akati wandei mamodheru kuita imwechete, yakasimba kwazvo kuita basa.
Kumhanyisa pombi yehuremu yakavhurika iyo inosvika kumuganhu wemhando yekuvanzika-inonzwa kutumirwa uko data risingakwanise kusiya maseva ekambani.
Maitiro Ekuita
Musanganiswa-we-Agents Aggregation mukuita
Kubatanidza matatu akasiyana eakavhurika machat modhi sevamiriri, wozoshandisa yakasimba aggregator kuburitsa imwe yakakwenenzverwa-tsigiro yemutengi.
Kubatanidza matatu akasiyana akavhurika ekutaura modhi sevamiriri, wobva washandisa aggregator yakasimba kugadzira imwe yakakwenenzverwa-tsigiro yemutengi Matimu anowanzo kuwana mhedzisiro iri nani kana achinge atsanangura emhando yepamusoro kumberi, chengetedza nzira yekukwira kwevanhu yemakesi ekumucheto, uye kuteedzera zvese zvakawanikwa zvechigadzirwa uye kukanganisa mutengo nekufamba kwenguva.
Musanganiswa-we-Agents Aggregation mukuita
Kuwedzera kuraira-kutevera zvibodzwa paAlpacaEval-maitiro mabhenji uchishandisa chete akavhurika-sosi mhando.
Kuwedzera kuraira-kutevera zvibodzwa paAlpacaEval-maitiro mabhenji anoshandisa akavhurika-sosi modhi Matimu anowanzo kuwana mhedzisiro iri nani kana achinge atsanangura emhando yepamusoro kumberi, chengetedza nzira yekukwira kwevanhu yemakesi emupendero, uye kuteedzera zvese zvakawanikwa zvechigadzirwa uye mutengo wekukanganisa nekufamba kwenguva.
Musanganiswa-we-Agents Aggregation mukuita
Kusanganisa akasiyana kodhi mazano kubva kune akati wandei mamodheru kuita imwechete, yakasimba kwazvo kuita basa.
Kusanganisa mazano akasiyana ekodhi kubva kune akati wandei mamodheru kuita imwechete, yakasimba kuita basa Matimu anowanzo kuwana mhedzisiro iri nani kana achinge atsanangura emhando yepamusoro kumberi, chengetedza nzira yekukwira kwevanhu yemakesi ekumucheto, uye kuteedzera zvese zvakawanikwa zvechigadzirwa uye mutengo wekukanganisa nekufamba kwenguva.
Musanganiswa-we-Agents Aggregation mukuita
Kumhanyisa pombi yehuremu yakavhurika iyo inosvika kumuganhu wemhando yekuvanzika-inonzwa kutumirwa uko data risingakwanise kusiya maseva ekambani.
Kumhanyisa pombi yakavhurika huremu iyo inosvika kumuganhu wemhando yekuvanzika-inonzwa kutumirwa uko data risingakwanise kusiya maseva ekambani Matimu anowanzo kuwana mibairo iri nani kana achinge atsanangura zvikumbaridzo zvemhando yepamusoro kumberi, chengetedza nzira yekukwira kwevanhu yemakesi ekumucheto, uye kuteedzera zvese zvakawanikwa zvechigadzirwa uye kukanganisa mutengo nekufamba kwenguva.
Njodzi & Guardrails
Chokwadi chehuroyi chinogona kupinda chinyararire mishumo, kuyerera kwetsigiro, kana tsvakiridzo.
Kunzwa nekukasira kunogona kugadzira mhedzisiro isingaenderane pane zvikumbiro zvakafanana.
Sensitive text data inogona kuburitswa kana zvidhiraivho zvisina kusimba.
Implementation Roadmap
Tsanangura chimiro chekubuda, toni, uye mhando zviyero usati waburitsa.
Tsanangura chimiro chekubuda, toni, uye mhando zviyero usati waburitsa. Bata nhanho yega yega segedhi rehumbowo: kana maitiro asina kusangana, imbomira kuburitsa, vhara gaka, uye wobva wawedzera kushandiswa.
Mhinduro dzepasi neakavimbika masosi pese pazvine basa.
Mhinduro dzepasi neakavimbika masosi pese pazvine basa. Bata nhanho yega yega segedhi rehumbowo: kana maitiro asina kusangana, imbomira kuburitsa, vhara gaka, uye wobva wawedzera kushandiswa.
Chengetedza ongororo yekuongorora yemunhu kune yakakwira-stake zvinobuda.
Chengetedza ongororo yekuongorora yemunhu kune yakakwira-stake zvinobuda. Bata nhanho yega yega segedhi rehumbowo: kana maitiro asina kusangana, imbomira kuburitsa, vhara gaka, uye wobva wawedzera kushandiswa.
Tevera maitiro ekutadza uye dzidzisazve kukurudzira kana mafambiro ebasa nguva nenguva.
Tevera maitiro ekutadza uye dzidzisazve kukurudzira kana mafambiro ebasa nguva nenguva. Bata nhanho yega yega segedhi rehumbowo: kana maitiro asina kusangana, imbomira kuburitsa, vhara gaka, uye wobva wawedzera kushandiswa.