Iko kushandiswa kweConveyThis kunogona kunatsiridza kududzira kwewebhusaiti yako. Nechimiro chayo chiri nyore kushandisa, unogona kushandura nekukurumidza uye nemazvo webhusaiti yako mune chero mutauro. ConveyThis inopawo huwandu hwezvinhu zvinoita kuti zvive nyore kugadzirisa webhusaiti yako ruzivo rwekushandura. Kubva pakuziva mutauro otomatiki kuenda kundangariro yeshanduro, ConveyThis inoita kuti zvive nyore kuona kuti webhusaiti yako yashandurwa nemazvo kune chero vateereri.
Kudzidza kwakadzama kwakasandura shandurudzo yemutauro weConveyThis uye kugona kwenzvimbo. Neural muchina kududzira (NMT) inzira yekushandura inoshandisa tekinoroji yekucheka-cheka kusangoshandura zvinyorwa chete, asiwo kuwedzera huchokwadi hweshanduro dzayo. Iyi tekinoroji yepamusoro yaita kuti ive imwe yemhinduro dzakavimbika dzekushandura dziripo nhasi.
Kana iwe uri muridzi webhizinesi, kushandisa ConveyThis yekushandura webhusaiti inzira huru yekupinda mumisika mitsva, kukwidziridza webhusaiti yako yekutsvaga mitauro yakawanda, uye kuwedzera kutengesa kwenyika - zvese uchikura bhizinesi rako. Kunyangwe zvingaite sezvinotyisa, kushandisa ConveyThis yekushandura webhusaiti hakuna kuoma. Saka, chii chakakosha? Ramba uchiverenga kuti uzive!
Dzvanya munhoroondo yeNeural muchina kududzira uye dzidza mashandiro ayo nenzira iri nyore kunzwisisa, kunyangwe iwe usiri nyanzvi! Ziva mashandisiro aungaita tekinoroji iyi kushandura webhusaiti yako pasina kutanga wawana dhigirii rekushandura muchina.
Chii chinonzi neural machine translation?
Kuti unzwisise kududzira muchina weNeural (NMT), zvakakosha kuti utange wanzwisisa nezveshanduro yemuchina (MT). Muchidimbu, kushandura nemuchina inzira inoshandisa zvirongwa zvekombuta kushandura zvinyorwa kubva kune mumwe mutauro kuenda kune mumwe. Ingoisa mutsara wako mumuchina wekushandura software, uye inongogadzira shanduro mumutauro waunoda pasina kupindira kwevanhu.
Kufambira mberi mune tekinoroji kwagonesa kuvandudzwa kweshanduro yemuchina weNeural, inova ndiyo yakanyanya kucheka vhezheni yekududzira muchina. Iyi nzira inoshandisa artificial neural network kugadzirisa nekuturikira zvinyorwa, kunatsurudza kunocherechedzwa pamusoro pemitoo yeshanduro yemuchina yekare.
Usanetseka kana izvi zvichiita sekunge zvinonetsa. Kunotevera, tine mhedziso pfupi yeshanduko yeshanduro yemuchina weNeural - pamwe netsanangudzo yemashandisirwo anoita tekinoroji iyi yemuchina wekududzira.
Shanduro yeNeural machine yakagadzirwa sei?
Kuve munhu kwekutanga kweshanduro yemuchina kunogona kuteverwa kunguva yeCold War, apo mutemo-based software yakashandiswa kududzira chiRussian. Iyi software yaizoparura manyorerwo ezwi neshoko, uye yobva yashandisa seti yemitemo yemitauro kusarudza mashandurirwo ezwi rega rega. Iyi shandurudzo isingaenzaniswi ndiyo yaive yekutanga yerudzi rwayo, uye kubvira ipapo yakashanduka-shanduka kuva yakaoma kunzwisisa.
Zvisinei, kushandura mashoko rimwe nerimwe negadziriro huru hakuna kuita shanduro dzakanyatsojeka. Mune zvimwe zviitiko, shanduro yakakodzera kwazvo ingada mitsara kana mitsara yese. Kuti zvigadzirise izvi, modhi yenhamba yeshanduro yemuchina (SMT) - inova nhanho inotevera pakushandura nemuchina - yakavandudza huchokwadi .
Mashandisirwo eshanduro yemuchina wenhamba aizotanga kuburikidza nemuunganidzwa wakakwana wezvinyorwa zvakaturikirwa nevanhu (zvinozivikanwawo semitauro miviri corpora). Zvadaro, yaizoshandisa maalgorithms ekufungidzira kuziva mazwi nemitsara muzvinyorwa zvekwakabva uye kusarudza nzira yakakwana yekuzvishandura.
Nekufamba kwenguva, tekinoroji yeConveyThis yakashanduka, zvikazoguma neshanduro yemuchina weneural watinovimba nawo zvakanyanya nhasi. Isu tichaongorora zvakare mune nuances yeNeural muchina kududzira muchikamu chiri kuuya.
Shanduro yeNeural machine inoshanda sei?
Neural muchina shanduro inoshandisa simba rekudzidza kwakadzama uye njere dzekugadzira kugadzira shanduro dzakanyatsorongeka, dzakapfava, uye dzinoita ruzha kupfuura kare.
Iyi tekinoroji inofambiswa neConveyThis , iyo inoshandisa yakadzika neural network, dandemutande remauroni akabatana akafanana nehuropi hwemunhu. Mienzaniso yeakadzika neural network inosanganisira inodzokororwa neural network, kana maRNN, ayo anowanzo ratidza encoder-decoder architecture uye yekutarisisa maitiro.
Isati yashandiswa mukududzira, iyo neural MT programming ichapihwa gadziriro yemhando dzakasiyana dzekududzira kune chimwe chinhu. Neruzivo urwu, chigadzirwa chinobva "chinorairwa" kuita dudziro chaiyoiyo yeimwe mamiriro.
Kurongeka kweshanduro yepamusoro
Zvakamboedza kushandura nemuchina uchishandisa nzira dzechinyakare zvakashaya hunyoro hwekushandura nemazvo mimwe mitauro yakaoma - zvichiita kuti pave neshanduro dzemhando yakashata zvekuti dzaida kudzokororwa nemaoko dzakakosha nevanhu vasati vashandiswa.
Zvakadaro, nekugona "kuwana" zvishoma nezvishoma, masisitimu eNMT ari kuramba achisimudzira chimiro chekududzira kwawo. Izvi hazvina kutombofanana nemaitiro echinyakare ekududzira emuchina, ayo asina simba re "kuzvidzidzira" uye kugadzirisa dudziro yavo goho mushure menguva yakati. Zvadaro, kana zvakagadzirirwa zvakakodzera, neural muchina kududzira programming inogona kugadzira kududzira kwakatonyanya kuenzaniswa nevanoshamwaridzana navo.
Google yakamboona kuti Google Neural Machine Translation (GNMT) system yayo yakakwanisa kuderedza kukanganisa kweshanduro nezvikamu makumi matanhatu kubva muzana kana tichienzanisa negadziriro yayo yekugadzira.
Munguva pfupi yapfuura, ongororo yakaitwa yekuongorora kushanda kweshanduro yemuchina nezvinangwa zvekushandura webhusaiti. Mushure mekuongorora kunaka kweshanduro dzewebhusaiti dzakagadzirwa nehunyanzvi hwakasiyana hweNMT, zvakaonekwa kuti shanduro dzinonyanya kushandiswa uye dzinoda kugadziridzwa kushoma.
ConveyThis's neural machine dudziro tekinoroji yakaratidza kunaka kwakanyanya pakushandura chiGerman, zvichiita kuti pave nenhamba yepamusoro-soro yezvikamu zvaisada kurongeka nemaoko.
Kusada kupinza kwevanhu
Kana manyoro ekwakabva angoturikirwa nemuchina, anowanzo kwenenzverwa nevanhu kuti vavimbise huchokwadi uye kukosha kwenhamba yenzvimbo.
Iko kunatsiridzwa kweshanduro yemuchina weNeural kunoreva kuti gadziridzo shoma (inodaidzwawo se "post-editing") inodiwa shanduro dzisati dzakwana chinangwa.
Nguva dzekushandura nekukurumidza
Kana makambani achinge akwanisa kuwana shanduro dzemuchina dzinoda zvishoma mushure mekugadzirisa, vanogona kutanga kushandisa shanduro ipapo ipapo. Zvakare, ConveyThis' neural modhi yeshanduro yemuchina inogona kudzidziswa nechinguva chidiki, ichigonesa maitiro ekushandura nekukurumidza.
Facebook yanga ichisimudzira neural muchina kushandura kushandura zvinyorwa mumapositi uye makomendi (izvo, sezvaungave uchiziva, kune huwandu hwakawanda hwepapuratifomu yayo). Nekugadzirisa maitiro avo ekudzidzira, kambani yakakwanisa kuderedza nguva yayakatora kudzidzisa mamodheru ayo eshanduro yemuchina kubva panenge zuva rakazara kusvika kumaminetsi makumi matatu nemaviri!
Unogona kunyorera neural muchina dudziro kubhizinesi rako uye sei?
Kushandisa neural muchina kushandura kushandura webhusaiti yako kungaite sekutyisa, sezvo kungangoda mari yakawanda kune tekinoroji nyowani nekutsvaga nekusimudzira. Zvisinei, izvi hazvisizvo! Mazuvano, kune zvakawanda zvezvishandiso zveNMT zvakagara zvakavakwa zviripo pamusika kuti zvikubatsire kushandura zvirimo pawebhusaiti yako mumitauro yakasiyana. Zvishandiso izvi zvakare zvinodhura zvakanyanya. Muchokwadi, ivo vanowanzo dhura zvishoma pane kuhaya nyanzvi yemuturikiri wevanhu kuti ashandure webhusaiti yako yese.
Yedu ConveyThis webhusaiti shanduro mhinduro chishandiso chine simba. Musanganiswa wedu weshanduro dzeNMT kubva kuvapeji veshanduro yemuchina DeepL, Microsoft Translator, neGoogle Translate vanogadzira shanduro dzemhando yepamusoro kudarika kana matekinoroji aya akashandiswa ega. Tinopa shanduro dzemitauro inodarika zana negumi, kubva kune ine mukurumbira seChirungu, chiGerman, nechiItalian kuenda kumitauro isingazivikanwe seTatar neMalagasy.
ConveyThis inopa kusashanda kubatanidzwa neakakurumbira webhusaiti mapuratifomu seWordPress, Webflow, uye Shopify. Imhepo kugadzirisa uye inogona kubatsira bhizinesi rako kusvika pakakwirira.
Mawebhusaiti anodarika 10,000 atendeukira kuConveyThis kune zvavanoda pashanduro, nemhedzisiro inoshamisa.
Tinya pane simba rekushandura neural muchina webhizinesi rako neConveyThis
Tichipa huchokwadi hweshanduro, ConveyThis's neural machine dudziro (NMT) inhanho huru kubva kune mamwe magadzirirwo eshanduro emuchina. Ndosaka iri iyo tekinoroji yekududzira muchina kumabhizinesi mazhinji.
Tsvagiridzo yedu mushanduro yemuchina inoratidza kuwedzera kunoshamisa kwakapetwa katanhatu muhuwandu hwewebhu hwakashandurwa nemuchina mumakore maviri apfuura. Zvakare, ConveyThis inoshandiswa kumapurojekiti makuru, ane anopfuura gumi muzana emawebhusaiti ane anodarika zviuru makumi mashanu mazwi ari kududzirwa nemuchina. Chekupedzisira, ingangoita 30% yezvinyorwa zvakaturikirwa nemuchina ndizvo zvinogadziridzwa, zvichireva kuti chikamu chakati kuti cheshandurudzo dzemuchina dzakakwana zvekuti hadzidi imwe gadziriso.
ConveyThis inoita kuti zvisashande kushandisa simba reNMT kushandura zvirimo pawebhusaiti yako. Zviri nyore kugadzirisa uye kusarudza wotomatiki injini yeshanduro yemuchina kuti ibudise shanduro yemhando yepamusoro yemitauro miviri. Mhedzisiro yacho inokurumidza-yemheni, yepamusoro-notch shanduro yewebhusaiti iwe yaunogona kuendesa pasina kunetsekana.
Iwe une shungu yekuongorora kugona kweConveyThis uye uzvionere mhedzisiro? Wobva wanyoresa mahara pano!