Amamodeli Wokushayela we-Wayve kanye ne-End-to-End
I-Wayve yinkampani yase-UK eyakha amasistimu okuzishayela enenethiwekhi eyodwa efundiwe ye-neural ebeka amaphikseli ekhamera ngokuqondile kuzilawuli zokushayela — ayikho imithetho enekhodi ngesandla noma amamephu e-HD.
Uhlolojikelele
It matters because this end-to-end approach promises cars that generalize to new cities without expensive remapping.
I-Deep Dive
Yasungulwa e-Cambridge ngo-2017, i-Wayve yenqaba iresiphi evamile yokuzishayela yamamojula ahlukene okubona, ukubikezela, nokuhlela ahlanganiswe ndawonye ngekhodi ebhalwe ngesandla. Esikhundleni salokho, iqeqesha inethiwekhi enkulu ye-neural ekupheleni-kuya-ekupheleni: ividiyo evela kumakhamera angabizi iyangena, isiteringi nokusheshisa kuyaphuma, okufundwe emibonisweni yokushayela kwabantu. U-Wayve ugwema i-LiDAR ebizayo kanye namamephu e-HD akhiwe ngaphambilini, ukubheja ukuthi ukufunda kuveza indlela abashayeli abangabantu. I-GAIA-1 yayo kanye ne-GAIA-2 kamuva amamodeli omhlaba akhiqizayo alingisa ividiyo yokushayela engokoqobo ukuze aqeqeshe futhi ahlole inqubomgomo. Ngo-2024 i-Wayve iqoqe ngaphezu kwesigidigidi esingu-$1 eholwa yi-SoftBank, Nvidia, kanye ne-Microsoft, futhi ihlole izimoto emadolobheni amaningi ase-UK futhi yaqala ukunwetshwa e-US nase-Japan.
I-Technical Insight
Ukufunda kusuka ekupheleni kuya ekupheleni kungena esikhundleni samapayipi ajwayelekile ngenethiwekhi ehlukanisekayo eqeqeshwe ngokulingisa ukufunda ekushayeleni komuntu, okuvamise ukucolisiswa ngokufunda okuqinisiwe. Amamodeli omhlaba ka-Wayve afana ne-GAIA-2 angamamodeli evidiyo akhiqizayo abikezela ozimele besikhathi esizayo abasesimweni sezenzo, okuvumela iqembu ukuthi likhiqize izimo ezingavamile (ama-jaywalkers, inkungu) ngokushibhile ngokulingisa. Uhlangothi oluphendukile luwukutolika: inqubomgomo yebhokisi elimnyama elilodwa inzima ukuyisusa nokuqinisekisa kunepayipi lapho okukhiphayo kwemojuli ngayinye kungahlolwa.
I-Strategic Impact
Isu lomthengisi
Imephu yemigwaqo yabathengisi ithonya ukuthi yiziphi izici iqembu lakho elingazakha ngokulandelayo.
Izindleko kanye nesabelomali
Imigomo yezohwebo nezinketho zokuthunyelwa zithinta izindleko zesikhathi eside nobungozi.
Ingozi nokuphepha
Izinxephezelo zenkampani zibumba okuzenzakalelayo komkhiqizo, ukuma kokuphepha, nokuvuleleka.
Ikusasa Lendlela kanye Namamodeli Okushayela Asekugcineni
U-Wayve unikeza ilayisensi 'ye-AI equkethwe' njengesofthiwe kubakhiqizi bezimoto esikhundleni sokwakha i-robotaxis yayo, ehlose ukuthumela usizo lwabashayeli futhi ekugcineni ibe nokuzimela kuzo zonke izinhlobo zezimoto. Lindela ukuhlanganiswa okuqinile ngamasu emodeli yesisekelo, amamodeli omhlaba amakhulu e-multimodal, kanye nokucindezela kokufakazela ukuthi amasistimu angenawo imephu angakwazi ukufanisa izimbangi ezinzima zemephu kwezokuphepha. Ukwemukelwa ngokomthetho kwezinhlelo ezifundiwe, ezingachazeki kancane kuseyisithiyo esibalulekile.
Ukuqaliswa Komhlaba Wangempela
Ukushayela kwedolobha ngaphandle kwemephu emadolobheni angajwayelekile ase-UK kusetshenziswa okokufaka kwekhamera kuphela kanye nenqubomgomo efundiwe
Imodeli yomhlaba ye-GAIA-2 ekhiqiza ividiyo ye-synthetic edge-case (abagibeli bamabhayisikili, isimo sezulu) ukuze ihlole ingcindezi inethiwekhi yokushayela
Ukunikeza ilayisense isofthiwe ye-AV2.0 kubakhi bezimoto ukuze amasudi ekhamera ezimoto athole ukushayela okusizwa okuthuthukile
Ukufunda nge-Fleet lapho idatha evela ezimotweni eziningi ezishayelwa abantu ithuthukisa imodeli eyodwa yokushayela kwe-neural okwabelwana ngayo
Izingozi & Guardrails
Izimemezelo zokwethula zingase zeqe ukuzinza ekugelezeni komsebenzi wangempela wokukhiqiza.
Izintengo ze-API noma izinguquko zenqubomgomo zingaphula ukucabanga ngobusuku obubodwa.
Ukuncika komthengisi oyedwa kukhulisa izindleko zokukhiya nokufuduka.
Ukuqalisa Umhlahlandlela
Linganisa abahlinzeki usebenzisa eyakho imisebenzi namasethi edatha.
Buyekeza ubumfihlo, ukuphepha, nemibandela yomthetho ngaphambi kokuhlanganiswa.
Gcina uhlelo lokubuyela emuva kuwo wonke amamodeli noma abathengisi.
Gada amanothi okukhululwa ukuze izinguquko zemephu yomgwaqo zingamangazi amaqembu.
Qhubeka Uhlole
Free newsletter
Get the daily AI briefing
Three verified AI stories every weekday morning, written in plain English. Free forever, no ads.
One email each weekday. Unsubscribe in one click. We never sell or share your address.
Test yourself
Take the Wayve and End-to-End Driving Models quiz
Instant feedback on every answer, and a shareable certificate with a verifiable ID once you pass a course.
Support free AI education. AI Understanding is a 501(c)(3) nonprofit — no ads, no paywall, ever. Make a donation
Umhlahlandlela olandelayo
I-Wayve LINGO Driving Language Models
Imibuzo evame ukubuzwa
What is Wayve and End-to-End Driving Models?
I-Wayve yinkampani yase-UK eyakha amasistimu okuzishayela enenethiwekhi eyodwa efundiwe ye-neural ebeka amaphikseli ekhamera ngokuqondile kuzilawuli zokushayela — ayikho imithetho enekhodi ngesandla noma amamephu e-HD. Kubalulekile ngoba le ndlela yokuphela konyaka ithembisa izimoto ezifika emadolobheni amasha ngaphandle kokulungiswa kabusha okumba eqolo.
Isiphi isici esichazayo sendlela ka-Wayve yokushayela esuka ekugcineni?
Ukuphela ukuya ekupheleni kusho ukuthi inethiwekhi eyodwa efundiwe isuka kokokufaka kwekhamera eluhlaza iye esiteringini nasekusheshiseni, kunokuhlanganisa amamojula akhiwe ngesandla ngamaketango.
Iliphi isu lezinzwa u-Wayve aziwa ngokuligcizelela?
I-Wayve ibhejela kumakhamera ashibhile futhi ufunde ukwenza okuvamile, igwema ngamabomu i-LiDAR ebizayo kanye namamephu e-HD akhiwe kusengaphambili.
Iyini i-GAIA-2 kusitaki sika-Wayve?
I-GAIA-1 ne-GAIA-2 amamodeli omhlaba akhiqizayo akhiqiza ividiyo yokushayela engokoqobo ukuqeqesha nokuhlola inqubomgomo yokushayela.
Ngabe u-Wayve uhlela kanjani ngokuyinhloko ukuthengisa ubuchwepheshe bakhe?
I-Wayve izibeka njengomhlinzeki wesofthiwe 'ehlanganisiwe ye-AI', enikeza amalayisense obuhlakani bayo bokushayela kubakhiqizi bezimoto.
Iyiphi i-drawback eyinhloko yenethiwekhi eyodwa yokushayela esuka ekupheleni?
Inqubomgomo yokufunda ye-monolithic iyibhokisi elimnyama, okwenza kube nzima ukuhlola, ukulungisa iphutha, kanye nokuqinisekisa kunepayipi elinemiphumela yemojuli ehlolekayo.