Abin da ke haifar da hankali na wucin gadi: yadda yake aiki, fa'idodi da haɗari

Generative AI shine mafi kyawun batun tattaunawa na fasaha na 2023.

Menene hazaka na wucin gadi, ta yaya yake aiki kuma menene game da shi? Bari mu ga shi tare a cikin wannan labarin

Menene hazaka na wucin gadi?

Generative AI nau'in fasaha ne na fasaha na wucin gadi wanda ke bayyana tsarin koyo na inji wanda zai iya samar da rubutu, hotuna, lamba, ko wasu nau'ikan abun ciki.

Samfuran na Generative wucin gadi hankali ana ƙara haɗawa cikin kayan aikin kan layi da chatbot wanda ke ba masu amfani damar rubuta tambayoyi ko umarni a cikin filin shigarwa, wanda samfurin AI zai samar da amsa irin ta ɗan adam.

Ta yaya haɓakar basirar wucin gadi ke aiki?

Samfuran na Generative wucin gadi hankali suna amfani da tsarin kwamfuta mai rikitarwa da aka sani da deep learning don nazarin tsarin gama-gari da tsare-tsare a cikin manyan saitin bayanai sannan amfani da wannan bayanin don ƙirƙirar sabbin sakamako masu jan hankali. Samfuran suna yin hakan ne ta hanyar haɗa dabarun koyan na'ura da aka fi sani da cibiyoyin sadarwa na jijiyoyi, waɗanda aka yi su da sauƙi ta hanyar yadda kwakwalwar ɗan adam ke sarrafa bayanai da fassarar bayanai sannan ta koya daga gare ta cikin lokaci.

Don ba da misali, ciyar da samfurin Generative wucin gadi hankali tare da ƙididdiga masu yawa, bayan lokaci samfurin zai iya ganowa da sake sake fasalin abubuwan da ke cikin labari, kamar tsarin makirci, haruffa, jigogi, na'urorin labari, da sauransu.

Samfuran na Generative wucin gadi hankali sun zama mafi ƙwarewa yayin da bayanan da suke karɓa da kuma samar da karuwa, sake godiya ga fasaha na deep learning kuma daga hanyoyin sadarwa na jijiyoyi kasa. Sakamakon haka, ƙarin abun ciki da samfuri ke haifarwa Generative wucin gadi hankali, yadda sakamakonsa ya kasance mai gamsarwa da kuma kama da mutum.

Misalai na Generative AI

ShahararriyarGenerative wucin gadi hankali Fashe a cikin 2023, galibi godiya ga shirye-shirye Taɗi GPT e SLAB di BABI. Bugu da ƙari, saurin ci gaban fasaha wucin gadi, kamar sarrafa harshe na halitta, ya sanyaGenerative wucin gadi hankali m ga masu amfani da abun ciki masu ƙirƙira a sikelin.

Manyan kamfanonin fasaha sun yi saurin yin tsalle-tsalle, tare da Google, Microsoft, Amazon, Meta da sauransu duk sun jera nasu kayan aikin ci gaba. Generative wucin gadi hankali cikin 'yan watanni.

Akwai kayan aiki da yawa Generative wucin gadi hankali, ko da yake nassi da samfurin tsara hotuna tabbas sun fi sani. Samfuran na Generative wucin gadi hankali yawanci sun dogara ga mai amfani da ke ba da saƙon da ke jagorantar su wajen samar da abin da ake so, walau ta rubutu, hoto, bidiyo ko wani yanki na kiɗa, kodayake ba koyaushe haka lamarin yake ba.

Misalai na ƙirƙira samfuran basirar ɗan adam

• ChatGPT: samfurin yaren AI wanda OpenAI ya haɓaka wanda zai iya amsa tambayoyi da samar da martani irin na ɗan adam daga umarnin rubutu.
• DAGA-E 3: wani samfurin AI daga OpenAI wanda zai iya ƙirƙirar hotuna da zane-zane daga umarnin rubutu.
• Google Bard: Google's Generative AI chatbot kuma kishiya ga ChatGPT. An horar da shi akan samfurin babban harshe na PaLM kuma yana iya amsa tambayoyi da samar da rubutu daga faɗakarwa.
• Claude 2 : Anthropic na tushen San Francisco, wanda tsoffin masu binciken OpenAI suka kafa a cikin 2021, sun sanar da sabon sigar ƙirar Claude AI a cikin Nuwamba.
• Tafiya ta tsakiya : Cibiyar bincike ta tushen San Francisco Midjourney Inc. ta haɓaka, wannan ƙirar AI tana fassara umarnin rubutu don samar da hotuna da zane-zane, kama da DALL-E 2.
• GitHub Copilot : kayan aiki mai ƙarfi na AI wanda ke ba da shawarar kammala lambar a cikin Kayayyakin Kayayyakin Kayayyakin Kayayyakin Kayayyakin, Neovim, da JetBrains ci gaban mahallin.
• Lamma 2: Za a iya amfani da tushen buɗe tushen Meta babban samfurin harshe don ƙirƙirar ƙirar AI ta tattaunawa don masu taɗi da mataimakan kama-da-wane, kama da GPT-4.
• xAI: Bayan bayar da tallafin OpenAI, Elon Musk ya bar aikin a watan Yuli 2023 kuma ya sanar da wannan sabon kamfani na AI. Samfurin sa na farko, Grok mara mutunci, ya fito a watan Nuwamba.

Nau'in samfuran AI masu haɓakawa

Akwai nau'ikan nau'ikan nau'ikan ƙirar AI, kowanne an tsara shi don takamaiman ƙalubale da ayyuka. Ana iya rarraba waɗannan gabaɗaya zuwa nau'ikan masu zuwa.

Transformer-based models

An horar da samfuran tushen canji akan manyan bayanan bayanai don fahimtar alaƙa tsakanin bayanan jeri, kamar kalmomi da jimloli. Goyan bayan deep learning, Waɗannan samfuran AI suna da masaniyar NLP da fahimtar tsari da yanayin harshe, suna sa su dace da ayyukan samar da rubutu. ChatGPT-3 da Google Bard misalai ne na samfuran AI masu haɓakawa na tushen wuta.

Generative adversarial networks

GANs sun ƙunshi hanyoyin sadarwa na jijiyoyi guda biyu waɗanda aka sani da janareta da wariya, waɗanda ke aiki da juna da gaske don ƙirƙirar ingantattun bayanai. Kamar yadda sunan ke nunawa, aikin janareta shine samar da tabbataccen fitarwa kamar hoto dangane da shawara, yayin da mai wariya ke aiki don tantance sahihancin hoton da aka faɗa. A tsawon lokaci, kowane bangare yana inganta a cikin ayyukansu, yana samun ƙarin sakamako masu gamsarwa. Dukansu DALL-E da Midjourney misalai ne na samfuran AI na tushen GAN.

Variational autoencoders

VAEs suna amfani da cibiyoyin sadarwa guda biyu don fassarawa da samar da bayanai: a wannan yanayin shi ne mai ɓoyewa da mai ƙididdigewa. Mai rikodin rikodin yana ɗaukar bayanan shigarwa kuma yana matsa su cikin sauƙi mai sauƙi. Mai rikodin sai ya ɗauki wannan matsewar bayanai ya sake gina shi zuwa wani sabon abu wanda yayi kama da ainihin bayanan, amma ba iri ɗaya bane.

Misali zai kasance koyar da tsarin kwamfuta don samar da fuskokin mutane ta amfani da hotuna azaman bayanan horo. A tsawon lokaci, shirin yana koyon sauƙaƙa hotunan fuskokin mutane ta hanyar rage su zuwa wasu abubuwa masu mahimmanci, kamar girma da siffar idanu, hanci, baki, kunnuwa da sauransu, sannan a yi amfani da su don ƙirƙirar sabbin fuskoki.

Multimodal models

Samfuran Multimodal na iya fahimta da sarrafa nau'ikan bayanai da yawa a lokaci ɗaya, kamar rubutu, hotuna, da sauti, yana ba su damar ƙirƙirar abubuwan da suka fi dacewa. Misali zai zama samfurin AI wanda zai iya samar da hoto dangane da saƙon rubutu, da kuma bayanin rubutu na saurin hoto. DALL-E 2 e GPT-4 ta OpenAI misalai ne na samfuran multimodal.

Fa'idodin haɓakar hankali na wucin gadi

Ga 'yan kasuwa, haɓakawa shine mafi girman fa'ida ta AI mai haɓakawa saboda yana iya baiwa 'yan kasuwa damar sarrafa takamaiman ayyuka da mayar da hankali kan lokaci, kuzari da albarkatu akan mahimman dabarun dabarun. Wannan na iya haifar da raguwar farashin aiki, haɓaka ingantaccen aiki da sabbin fahimta kan ko wasu hanyoyin kasuwanci suna aiki ko a'a.

Ga masu sana'a da masu ƙirƙirar abun ciki, kayan aikin AI masu haɓakawa zasu iya taimakawa tare da tsara ra'ayi, tsarawa da tsarawa, haɓaka injin bincike, tallace-tallace, haɗin gwiwar masu sauraro, bincike da gyarawa, da yiwuwar ƙari. Bugu da ƙari, babban fa'idar da aka ba da shawarar ita ce inganci saboda kayan aikin AI masu haɓakawa na iya taimaka wa masu amfani su rage lokacin da suke ciyarwa akan wasu ayyuka don su iya saka hannun jarin su a wasu wurare. Wancan ya ce, kulawa da hannu da sarrafa samfuran AI masu ƙima suna da mahimmanci.

Generative AI amfani lokuta

Generative AI ya sami gindin zama a cikin sassan masana'antu da yawa kuma yana haɓaka cikin sauri zuwa kasuwannin kasuwanci da kasuwanni. McKinsey kimanta cewa, nan da 2030, ayyukan da a halin yanzu ke da kusan kashi 30% na sa'o'in aiki a Amurka za a iya sarrafa su ta atomatik, godiya ga haɓakar haɓakar basirar ɗan adam.

A cikin sabis na abokin ciniki, AI-powered chatbots da kama-da-wane mataimaka na taimaka wa kamfanoni rage lokacin amsawa da sauri magance tambayoyin abokin ciniki gama gari, rage nauyi akan ma'aikata. A cikin haɓaka software, kayan aikin AI na haɓaka suna taimakawa masu haɓaka ƙididdige ƙididdigewa da inganci ta hanyar yin bitar lamba, nuna kwaro, da ba da shawarar yuwuwar mafita kafin su zama manyan matsaloli. A halin yanzu, marubuta za su iya amfani da kayan aikin AI na haɓakawa don tsarawa, tsarawa, da kuma sake fasalin kasidu, labarai, da sauran ayyukan da aka rubuta, kodayake galibi tare da sakamako masu gauraya.

Sassan aikace-aikace

Amfani da Generative AI ya bambanta daga masana'antu zuwa masana'antu kuma an fi kafa shi a wasu fiye da wasu. Abubuwan amfani na yanzu da shawarwarin amfani sun haɗa da masu zuwa:

• Lafiya: Generative AI ana binciko a matsayin kayan aiki don hanzarta gano miyagun ƙwayoyi, yayin da kayan aiki irin su AWS HealthScribe suna ƙyale likitoci su rubuta shawarwarin haƙuri kuma su loda mahimman bayanai a cikin rikodin likitan su na lantarki.
• Tallan Dijital: masu tallace-tallace, masu tallace-tallace da ƙungiyoyin kasuwanci na iya amfani da AI mai ƙima don ƙirƙirar keɓaɓɓen kamfen da keɓance abun ciki zuwa abubuwan da mabukaci ke so, musamman idan aka haɗa tare da bayanan gudanarwar abokin ciniki.
• Umarni: Wasu kayan aikin ilimi sun fara haɗawa da AI don haɓaka keɓaɓɓen kayan koyo waɗanda ke dacewa da salon koyo na ɗaiɗaikun ɗalibai.
• Kudi: Generative AI yana ɗaya daga cikin kayan aikin da yawa a cikin tsarin hadaddun kuɗi don nazarin tsarin kasuwa da hasashen yanayin kasuwannin hannun jari, kuma ana amfani dashi tare da sauran hanyoyin hasashen don taimakawa manazarta kuɗi.
• yanayi: a cikin kimiyar muhalli, masu bincike suna amfani da ƙirar ƙima ta wucin gadi don hasashen yanayin yanayi da daidaita tasirin canjin yanayi.

Hatsari da iyakoki na haɓakar basirar wucin gadi

Babban damuwa game da amfani da kayan aikin AI masu haɓakawa - musamman waɗanda ke da damar jama'a - shine yuwuwar su don yada rashin fahimta da abun ciki mai cutarwa. Tasirin hakan na iya zama mai fadi da tsanani, tun daga dawwamar ra'ayi, kalaman kyama da akidu masu cutarwa zuwa lalacewar mutunci da mutunci da kuma barazanar illar shari'a da kudi. Har ma an ba da shawarar cewa rashin amfani ko rashin amfani da janareta AI na iya jefa tsaron ƙasa cikin haɗari.

Wadannan hadarin ba su tsira daga 'yan siyasa ba. A cikin Afrilu 2023, Tarayyar Turai ta ba da shawarar sabbin dokokin haƙƙin mallaka don haɓaka AI wanda zai buƙaci kamfanoni su bayyana duk wani abu mai haƙƙin mallaka da aka yi amfani da shi don haɓaka kayan aikin fasaha na wucin gadi. An amince da waɗannan ka'idoji a cikin daftarin dokar da Majalisar Tarayyar Turai ta kada kuri'a a watan Yuni, wanda kuma ya hada da tsauraran takunkumi kan amfani da bayanan sirri na wucin gadi a cikin kasashe membobin EU, gami da shirin hana fasahar tantance fuska ta zahiri a sararin samaniya.

Ayyuka ta atomatik ta hanyar AI mai haɓakawa kuma yana haifar da damuwa game da ƙarfin aiki da ƙauracewa aiki, kamar yadda McKinsey ya haskaka. A cewar ƙungiyar masu ba da shawara, sarrafa kansa na iya haifar da sauye-sauyen sana'a miliyan 12 tsakanin yanzu zuwa 2030, tare da asarar ayyukan yi a cikin tallafin ofis, sabis na abokin ciniki da sabis na abinci. Rahoton ya kiyasta cewa bukatar ma'aikatan ofis na iya "…

Generative AI da janar AI

Generative AI da AI na gaba ɗaya suna wakiltar bangarori daban-daban na tsabar kudin guda. Dukansu biyu sun shafi fannin hankali na wucin gadi, amma tsohon nau'in nau'in na ƙarshe ne.

Generative AI yana amfani da dabaru daban-daban na koyon inji, kamar GAN, VAE, ko LLM, don samar da sabon abun ciki daga samfuran da aka koya daga bayanan horo. Waɗannan abubuwan fitarwa na iya zama rubutu, hotuna, kiɗa, ko wani abu wanda za'a iya wakilta ta lambobi.

Hankali na wucin gadi, wanda kuma aka sani da hankali na wucin gadi, gabaɗaya yana nufin manufar tsarin kwamfuta da na'ura mai ba da hanya tsakanin hanyoyin sadarwa waɗanda ke da hankali da yancin kai irin na ɗan adam. Wannan har yanzu kayan almara na kimiyya ne: yi tunanin Disney Pixar's WALL-E, Sonny daga 2004's I, Robot, ko HAL 9000, rashin hankali na wucin gadi daga Stanley Kubrick's 2001: A Space Odyssey. Yawancin tsarin AI na yanzu misalai ne na "kunkuntar AI", kamar yadda aka tsara su don takamaiman ayyuka.

Generative AI da koyo na inji

Kamar yadda aka bayyana a sama, Generative AI wani yanki ne na hankali na wucin gadi. Samfuran AI na Generative suna amfani da dabarun koyon injin don sarrafawa da samar da bayanai. Gabaɗaya, hankali na wucin gadi yana nufin ra'ayin kwamfutoci masu ikon yin ayyuka waɗanda in ba haka ba zasu buƙaci hankalin ɗan adam, kamar yanke shawara da NLP.

Koyon na'ura shine ainihin abin da ke tattare da hankali na wucin gadi kuma yana nufin aiwatar da algorithms na kwamfuta zuwa bayanai don manufar koyawa kwamfuta yin takamaiman aiki. Koyon na'ura shine tsari wanda ke ba da damar tsarin basirar wucin gadi don yanke shawara ko tsinkaya bisa tsarin koyo.

Shin haɓakar basirar wucin gadi na gaba?

Haɓaka haɓakar haɓakar haɓakar AI ba ta nuna alamun raguwa ba, kuma yayin da kamfanoni da yawa ke karɓar dijital da sarrafa kansa, AI mai haɓakawa ya yi kama da taka muhimmiyar rawa a nan gaba na masana'antar. Ƙwararrun haɓakar AI sun riga sun tabbatar da mahimmanci a cikin masana'antu irin su ƙirƙirar abun ciki, haɓaka software, da magani, kuma yayin da fasahar ke ci gaba da haɓakawa, aikace-aikacen sa da amfani da lokuta za su fadada.

Wannan ya ce, tasirin AI mai haɓakawa akan kasuwanci, daidaikun mutane da al'umma gabaɗaya ya dogara da yadda muke magance haɗarin da yake bayarwa. Tabbatar da cewa ana amfani da hankali na wucin gadi na da'a rage son zuciya, inganta gaskiya da rikon amana da tallafawa shugabanci na bayanai za su kasance masu mahimmanci, yayin da tabbatar da cewa ƙa'ida ta ci gaba da tafiya tare da saurin haɓakar fasaha ya riga ya zama kalubale. Hakanan, gano ma'auni tsakanin aiki da kai da sa hannun ɗan adam zai zama mahimmanci idan muna fatan yin amfani da cikakkiyar damar haɓakar AI yayin rage kowane mummunan sakamako.

Ercole Palmeri