Web 3.0: Fasahar “Artificial Intelligence” (1)

Tsarin Web 3.0

Kamar yadda bayani ya gabata a makon jiya, marhalar “Web 2.0” na kan mana bankwana ne, domin alamomin sabuwar marhala sun fara bayyana tun shekarusama da biyar da suka gabata.  Daga wannan mako zuwa makonnin dake tafe, za mu fara sharhi ne kan siffofin wannan marhala dake tafe.  Wadannan siffofi sun samo asali ne daga manyan fasaha da tsare-tsaren sadarwa guda uku: fasahar Blockchain, wanda ke taimakawa wajen ta’ammali da bayanai ta hanya mai cike da tsaro, sai fasahar “Artificial Intelligence” ko “AI”, wanda ke samar da siffofi da dabi’un dan adam ga kayayyaki da hanyoyin sadarwa na duniya, sai kuma na karshe wato fasahar “Edge Computing”, wanda ya kunshi tsarin kusantar da bayanai ga na’urorin sadarwa, don samunsu a duk sadda ake bukata.

Makalarmu ta yau da na makon dake tafe, za su yi takaitaccen sharhi ne kan fasahar “Artificial Intelligence” (AI), da misalai kan hakan, don taimaka wa masu karatu fahimtar inda duniyar sadarwa ta dosa, saboda samun karuwar wayewa.

Fasahar “AI” – Wata Sabuwar Duniya

A cewar Mai girma Kachallan Kano, Malam Magaji Galadima, “Duniya kafar mota ce; kullum juyawa take yi.”  Wannan magana haka take.  Dalili kuwa, a cewarsa: “Duk sadda kaga wani abin da ya birgeka na ci gaba, sai ka dauka an zo karshe kenan.  Amma da zarar wani ya yi nutso cikin kogin ilimi, sai kaga ya fito da wani abu sabo.”  Sadda aka samu kwamfutoci a duniya, yawancin mutane sun dauka duniyar ci gaba ta zo karshe.  Domin a tunaninsu, babu wani abu da dan adam ya taba kerawa a fannin sadarwa da sarrafa bayanai, irin kwamfuta.  Har ta kai daya daga cikin manyan masana ilimin kwamfuta dake karantarwa a wata babbar jami’a daga cikin manyan jami’o’in kasar Amurka mai suna: David Eck, ya rubuta littafi na musamman mai suna: “The Most Complex Machine – A Survey of Computers and Computing.”  Wanda ke nuna lallai na’urar kwamfuta fa, ita ce karshe; babu wata na’ura da aka taba kerawa mai sarkakiya irinta.

Bayyanar Intanet kuma ya sa duk inda kaga kwamfuta, muddin ba a iya amfani da Intanet a kanta, to, gani ake kamar holoko ce.  Ana haka kuma sai bincike ya ingiza wancan ci gaban zuwa gaba, inda a halin yanzu da kwamfutar, da kuma fasahar Intanet din, muddin babu wani dandano na “dabi’un” dan adam wajen gudanuwarsu, to, sun zama kamar hoto.  Fannin da ya kawo wannan ci gaba kuma shi ake kira: “Artificial Intelligence”, ko “AI” a gajarce.

Ma’ana da Asalin Fasahar “AI”

Wannan fanni na “AI”, fanni ne dake karantar da yadda za a iya tsofa wa kowace irin na’ura da hanyoyin sadarwa na zamani dabi’u da halayya irin ta dan adam.  Manufar fasahar “AI” ita ce, koya wa kwamfuta da manhajojinta, da na’urorin sadarwa (irin su wayar salula da nau’ukanta), wasu daga cikin tsarin tunani da dabi’un dan adam, don ba su damar aiwatar da ayyuka a kintse, a natse, a cike, a lokaci da yanayin da ake son su gabatar.

Fasahar “AI” fanni ne mai fadin gaske.  Duk da cewa yanzu ne ake fara cin gajiyarta a aikace, sai dai ta samo asali ne tun shekarun 1950s, sadda bincike ya fara gudana har zuwa shekarar 1977.  Daga nan kuma komai ya tsaya.  Domin ta la’akari da ka’idojin wannan fanni, ana bukatar tarin bayanai masu dimbin yawa don amfani da su wajen koya wa kwamfuta, tsari da kintsin tunanin da ake son tayi.  Wanda a lokacin da wancan bincike ya faro, babu irin wannan bayanai da ake bukata; domin ko fasahar Intanet ma bata bunkasa ba.  To amma yanzu a dukkan dakika, mutane da ma’aikatu da masana’antu na samar da bayanai masu dimbin yawa da ya wuce tunanin dan adam; la’alla ta hanyar Intanet ne, ko kuma wajen gudanar da ayyukansu ne.  Wannan dimbin bayanai shi ake kira: “Big Data”, kuma dasu ake amfani wajen aiwatar da wannan tsari na fasahar “AI”.

A karkashin fannin “AI” akwai fannoni guda biyu mahimmai, wadanda su ne ginshikin fannin baki daya.  Bangaren farko shi ake kira: “Machine Learning” ko “ML” a gajarce.  Wannan fanni na dogaro ne da hanyoyin koya wa kwamfuta ko wata manhaja iya  fahimtar abubuwa – rubutu ne, ko sauti, ko alamomi, ko siffofi na musamman, ko hotuna – don sabawa da wadannan siffofi da iya gane su nan gaba.  Galibi akan yi amfani ne da ka’idojin ilimin kididdiga, wato: “Statistical Models”, don tabbatar da dabi’un da ake son kwamfuta ko wata manhajar kwamfuta na musamman ta saba dasu.  Karkashin wannan fanni, ana dabi’antar da kwamfuta ko manhajar kwamfuta ne da abubuwan da aka tsara mata na musamman.  Wannan yanayi na koyar da kwamfuta hanyoyin gane abubuwa shi ake kira: “Supervised Learning.”

Fanni na biyu kuma shi ne fannin: “Deep Learning”, ko “DL” a gajarce.  Shi kuma wannan fannin dashi ake amfani wajen dabi’antar da wata na’ura ta musamman, kamar na’urorin da ake amfani dasu a masana’antu don kere-kere na zamani, wajen tantance abubuwa, ko aiwatar da wasu ayyuka, ko tantance ingancin wani tsari da dai sauransu.  Karkashin wannan tsari ko fanni, ana girka wa na’ura ne irin dabi’un da kwakwalwar dan adam ke dauke dasu, sai a bar na’urar tayi tunani sannan ta ba da sakamakon abin da ta fahimta daga abubuwan da aka gabatar mata.  Wannan fanni na amfani ne da wani tsari mai suna: “Neural Network”.  Wato amfani da irin tsarin da kwakwalwar dan adam ke bi wajen tunani da zartarwa.  Kuma wannan tsarin karantarwa shi ake kira: “Unsupervised Learning.”