Injin AI da yake aiki a cikin kowace tafiyar Uber
Karfe 5:00 na yamma ne a ranar Juma'a a tsakiyar birnin Sydney.
Ofisoshi na fanko a cikin mintuna. Abinci na cika. Mutane 40,000 sun buɗe manhajar Uber ɗinsu kusan a lokaci guda. Ruwan sama ya fara sauka.
Tambayar da Uber dole ta amsa yanzu tana da sauƙi sosai. Nishaɗin tafiya nawa za su iso a cikin mintuna 15 masu zuwa?
Idan ka yi kuskure, sakamakonsa nan take. Jira mai tsawo. Soke tafiye-tafiye. Abokan ciniki suna koma waɗanda ke fafatawa. A ma'auni na Uber, karamin kuskure a hasashe ba matsalar ƙidaya ba ne. Matsalar kuɗaɗen shiga ce da ake aunawa da miliyoyin kuɗi.
Mafi yawan mutane suna tunanin cewa Uber yana aiki ne kawai bayan an fara buƙata. Mai tafiya ya buɗe manhaja kuma tsarin ya fara gudu don nemo direba kusa. Ba haka yake aiki ba.
Tsarin AI na Uber yana gudanar da hasashen buƙata a koyaushe, awanni 24 a rana. Da lokacin da ka buɗe manhajar a yammacin Juma'a, Uber ta riga ta sani kana zuwa. Direban da zai ɗauke ka wataƙila an sake ajiye shi mintuna 20 da suka wuce bisa ga hasashen da ya annabci buƙata a ainihin wannan wurin a ainihin wannan lokacin.
Wannan ƙarfi ba a gina shi ne a kan hasashen zuciya ba. An gina shi ne a kan lissafi.
Ƙarin kashi 1 cikin ɗari a daidaiton hasashe a ma'auni na Uber yana juya zuwa ɗaruruwan miliyoyin daloli na ƙarin kuɗaɗen shiga na shekara-shekara. Ba daga wani babban canji guda ba, amma daga dubban ƙananan rashin inganci da aka cire a lokaci guda, suna taruwa kowace awa a kowane birni.
Wannan ita ce Kashi na 1 daga jerin labarai guda 5 game da yadda Uber ke annabta nan gaba kafin ya faru.
Kashi na 2 yana rufe gine-ginen ƙungiyar samfura uku. Kashi na 3 yana tafiya cikin kowace dabara da ainihin lambobi daga yammacin Juma'a a Sydney. Kashi na 4 yana gina cikakken aiwatarwa daga farko a cikin Python. Kashi na 5 yana nuna tsarin PyTorch na samarwa da yake aiki a babban ma'auni.
An tabbatar da kowace lissafi. Kowace layin lambar yana aiki.
Ci gaba da binmu. Lissafin ya fi sauƙi fiye da yadda mafi yawan mutane ke zato.
#MachineLearning #ArtificialIntelligence #AI #DataScience #Mathematics #Uber #TimeSeriesForecasting #DemandForecasting #AIEngineering #TechLeadership #TheMagicOfHiddenMathematicsInAI #PredictiveAnalytics #DeepLearning #MLOps #Python
Originally published on LinkedIn by Keshavan Thiagaraja. Read the original on LinkedIn.