Isingeniso Somkhiqizo
I-DeepDragon ™, Iwuhlelo oluhlakaniphile oluhamba phambili emhlabeni wonke futhi oluhamba phambili emhlabeni jikelele olungasiza ngokushesha ekwakhiweni nasekusebenzeni kahle kwezinkampani zangaphandle ezindaweni ezikhethiwe. Ukuxazulula ngokushesha nangempumelelo izidingo zokuthatha izinqumo zokutshala izimali eziphuthumayo zemboni yokukhuculula indle ezindaweni zasemakhaya ukuze kuklanywe umshini ozishintshayo, uhlahlomali lwezindleko zokutshala izimali, kanye nokusebenza okudidiyelwe kwezitshalo zendle namapayipi amasha eziteshi. Uma kuqhathaniswa nomklamo wendabuko, ukusebenza kahle kungathuthukiswa ngamaphesenti angaphezu kuka-50, futhi izinga lokusebenza ngempumelelo kwempahla lingafinyelela ku-100%, kuzuzwe ukusebenza okuhlakaniphile okungu-24/7 kwenethiwekhi yefekthri, ngamathemba abanzi okusebenza kwezobuchwepheshe.
Ukuhlanganisa amandla ne-Institute of Automation ye-Chinese Academy of Sciences ukusungula i-China Academy of Sciences i-Suzhou Research Institute Liding Environmental Protection Joint Laboratory.
Umholi wokuqala emhlabeni jikelele
Umbiko Wosesho Olusha Lobuchwepheshe Bamazwe Ngamazwe
50+onjiniyela bokuthuthukisaIzinsuku eziyi-1000 zocwaningo lobuchwepheshe nokuthuthukiswa Kwamalungelo obunikazi ayisisekelo angu-4
I-Core Technology
Ukuze kuxazululwe inkinga yokuthola idatha eyisisekelo endaweni eqokiwe, isistimu isebenzisa indlela yokufanisa yasemoyeni esheshayo isebenzisa izimoto zasemoyeni ezingenamuntu ngaphezu kwezindlela zokuthola ezivamile zokuzuza ukutholwa kwedatha. Ama-algorithms asuselwe ekufundeni okujulile angabona ngokuzenzakalelayo izici ezifana nemigwaqo, izindlu, namasistimu amanzi. Bona ukuhlaziya okuzenzakalelayo kanye nokuhlonza imigomo ngokushesha.
Njengamanje, uhlelo seluphothulile ukuqeqeshwa okuyimodeli nokufunda kwedatha yokuqoqwa kwezithombe zasemoyeni ezingaphezu kwezingu-5000 zobukhulu bezigcawu ezihlukene, ngokunemba kwezichasiselo okuzenzakalelayo okungu-90%. Ingakwazi ukufinyeza kakhulu noma ithathele indawo umsebenzi wedatha oyisisekelo njengendawo yasemaphandleni kanye nokuhlola isimo somhlaba nesichasiselo, ithuthukise kakhulu ukusebenza kahle komsebenzi, futhi yonge isikhathi nezindleko zabasebenzi.
Ekwakheni amapayipi, ngokusekelwe emiphumeleni yokuqashelwa kwesici, ukuklanywa kwesikimu kuyenziwa, kucatshangelwa ubunzima bezindlu zasemaphandleni kanye nemigwaqo. Ekuqaleni kwedizayini, samukela ubuchwepheshe obuhlukahlukene kanye ne-algorithms efana nokunwetshwa kwenethiwekhi yomgwaqo ukuze kukhishwe futhi kuhlanzwe uhlaka lwamathambo enethiwekhi yomgwaqo, ukususa noma ukunciphisa amagatsha amanga kunethiwekhi yomgwaqo, futhi sithene futhi sihlanze inethiwekhi yomgwaqo ukuze kwenziwe inethiwekhi yomgwaqo ekhishiwe. uhlaka lwamathambo luhambisana kakhulu nesakhiwo sangempela senethiwekhi yomgwaqo.
Usebenzisa ithiyori yegrafu, phakamisa ngendlela entsha imodeli yemephu yendawo yasemaphandleni, ukhiphe futhi uguqule izici ezihlonziwe, futhi wakhe imodeli yemephu yendawo yasemaphandleni eyenza idijithali futhi ibone ngeso lengqondo ubudlelwano phakathi kwezindlu, imigwaqo, namabanga azo kanye nokuphakama okuhlobene endaweni.
Ekugcineni, idatha yesithombe nolwazi lokuphakama kufakwa ku-algorithm yokukhiqiza inethiwekhi yamapayipi. Ngokuphindaphinda okuningi, i-algorithm yendlela emfushane yomthombo owodwa ibizwa ukuhlanganisa ukuzizwisa kwedizayini eqoqwe amayunithi amaningi edizayini ibe yimithetho yokuklama.
Kuhlanganiswe ne-algorithm, ukufunda okujulile kwemithetho yokuklama amapayipi kungalingisa ngokunembile ukugeleza kwendle kanye nezinhlelo zamapayipi zangaphambi kokubeka. Yehlisa izinga lephutha lokuklama okuzenzakalelayo kwepayipi libe phakathi kuka-10%.
Uhlelo lungakwazi ukukhiqiza ngokuzenzakalelayo uhlu oluningiliziwe lwebhajethi yotshalomali olususelwe kudatha yenani lobunjiniyela kanye nemithetho yezindleko enikezwe esikimini sokuklama amapayipi okuzenzakalelayo. Qaphela ukuhlelwa kwesabelomali okufika ngesikhathi nokunengqondo.
Enkambweni yokuklama ipayipi, ukukhethwa kwemishini kungafinyelelwa, futhi uhlelo lungakhetha imikhiqizo ehlukene yemishini nezinqubo ngokusekelwe kudatha yesifunda yokusetshenziswa kwamanzi ehlinzekwa yi-algorithm, ukuqinisa ukuphathwa okuhlanganisiwe kwefekthri kanye nenethiwekhi.
Inkundla ihlinzeka ngokuhlelwa kwesikhathi sangempela kwezici zesici nomklamo wepayipi, futhi ama-algorithms angamunca futhi afunde ulwazi lokusebenza komsebenzisi, athuthukise amakhono okusebenzisana phakathi kwabantu namakhompyutha futhi athuthukise izinga lobuhlakani bomkhiqizo.
Inkundla futhi ihlinzeka ngesixazululo sokumisa okukodwa kubasebenzisi ukuze benze ngendlela oyifisayo amafomethi emephu ukuze kuthuthukiswe ukubonwa kwe-WebGIS, amafomethi asekelayo afana namafayela efomethi yefomethi ye-AutoCAD DWG evame ukusetshenziswa, i-GeoJSON namanye amafomethi efayela e-GIS asetshenziswa kakhulu, Isebenzisa amabhlogo e-WebGL vector kanye nezitayela zangokwezifiso ukwethula. amamephu asebenzisanayo, ahlinzeka ngokubonwa kwedatha enkulu entsha kanye nemisebenzi yokubuka idatha yokusakaza-bukhoma ngesikhathi sangempela. Ngalo mkhiqizo, amafayela okwakhiwa kwamapayipi angalayishwa futhi aboniswe kumamephu we-GIS. Qaphela ukuphathwa okudidiyelwe kwedatha yenethiwekhi yefekthri kuwo wonke umjikelezo wokuphila.
I-Liding Environmental Protection's Isiteji sokusebenza kokukhucululwa kwendle esihlukaniselwe abantu abaningi sidale imodeli entsha yokusebenza kwendle. Ingakwazi ukufeza ukuzimela kwenqubo egcwele nokulawula, ngokuthunyelwa okusheshayo, kulungele ukusebenzisa ukufinyelela kwedatha, ukucushwa okuguquguqukayo kokusebenza, nezindleko eziphansi zokwakha. Mayelana nokwehliswa kwezindleko, ukuthuthukiswa kokusebenza kahle, ukugwema ubungozi kanye nokuthotshelwa kwemithetho, sihlose ukukhulisa izinga lokusebenza kwemishini, ukunciphisa izindleko zokusetshenziswa kwamandla, ukwehlisa isikhathi sokuhlola abasebenzi, kanye nokuthuthukisa ukuqapha nokusebenza kahle kokuthatha izinqumo. Ukwenza ngcono ngokuphelele izinga, ukwenziwa ngobungcweti, ukumisa, kanye nezinga lobuhlakani bokusebenza nokulungiswa kwamayunithi, kanye nokuhlinzeka kangcono ukusebenza nokulawulwa kokulahlwa kwendle.
Uhlelo lungasetha ngokukhululekile imithetho ye-alamu ukuze kubonakale i-alamu yesikhathi sangempela kanye nokuphusha kwe-akhawunti esemthethweni ye-WeChat, ukusatshalaliswa komsebenzi okukhaliphile, ukungena nokuphuma kwezisebenzi kanye nemoto, okokufaka nokucutshungulwa kolwazi lokusebenza, kanye nokukhiqizwa okuzenzakalelayo kombiko wokusebenza.
Inganciphisa isikhathi sokuhlola okwenziwa ngesandla ngama-40%, ithuthukise ukuqapha kanye nokusebenza kahle kokuthatha izinqumo ngama-20%, ithuthukise ngokuphelele izinga, ukwenziwa ngobungcweti, ukumisa izinga, nezinga lezobunhloli zokusebenza nokugcinwa kwezindawo zokunakekela, futhi inikeze kangcono ukusebenza nokuphathwa kwendle. ukwelashwa.
Inqubo ye-R&D
