Komawa ga jigon, nasarar AIGC a cikin singularity hade da abubuwa uku:

1. GPT kwafi ne na neurons na ɗan adam

GPT AI wanda NLP ke wakilta shine algorithm neural network na kwamfuta, wanda asalinsa shine a kwaikwayi hanyoyin sadarwa na jijiyoyi a cikin kwakwalwar mutum.

Gudanarwa da ƙwararrun tunanin harshe, kiɗa, hotuna, har ma da ɗanɗano bayanai duk ayyukan ne da ɗan adam ya tara.

kwakwalwa a matsayin "kwamfutar furotin" a lokacin juyin halitta na dogon lokaci.

Don haka, a zahiri GPT shine mafi dacewa kwaikwaiyo don sarrafa irin wannan bayanai, wato, harshe mara tsari, kiɗa, da hotuna.

Hanyar sarrafa ta ba fahimtar ma'ana ba ce, a'a tsari ne na tacewa, ganowa, da haɗin gwiwa.Wannan shi ne sosai

abu mai ban mamaki.

Algorithms na fahimtar ma'anar magana ta farko sun kafa ƙirar nahawu da ma'ajin bayanai, sannan aka tsara jawabin zuwa ƙamus,

sa'an nan kuma sanya ƙamus a cikin ma'ajin nahawu don fahimtar ma'anar ƙamus, kuma a ƙarshe ya sami sakamakon ganewa.

Fahimtar ingancin wannan “hanyar ma’ana” tushen fahimtar ma’anar syntax yana ta shawagi kusan kashi 70%, kamar fitinar ViaVoice.

Algorithm wanda IBM ya gabatar a cikin 1990s.

AIGC ba game da wasa kamar wannan ba.Mahimmancinsa ba shine kula da nahawu ba, amma don kafa tsarin hanyar sadarwa na jijiyoyi wanda ke ba da damar

kwamfuta don ƙididdige haɗin kai mai yiwuwa tsakanin kalmomi daban-daban, waɗanda haɗin gwiwar jijiyoyi ne, ba haɗin haɗin kai ba.

Kamar koyan harshen mu na asali sa’ad da muke ƙuruciya, a zahiri mun koyi shi, maimakon koyan “batu, predicate, abu, fi’ili, ƙarin,”

sannan kuma fahimtar sakin layi.

Wannan shine tsarin tunani na AI, wanda shine ganewa, ba fahimta ba.

Wannan kuma shine mahimmin ɓarna na AI ga duk samfuran kayan aikin gargajiya - kwamfutoci ba sa buƙatar fahimtar wannan lamarin a matakin ma'ana,

sai dai gano da kuma gane alaƙar da ke tsakanin bayanan ciki, sannan ku san su.

Misali, yanayin kwararar wutar lantarki da tsinkayar grid masu wutar lantarki sun dogara ne akan simintin wutar lantarki na gargajiya, inda tsarin lissafi na

an kafa tsarin sannan kuma a haɗa shi ta amfani da matrix algorithm.A nan gaba, bazai zama dole ba.AI za ta gano kai tsaye da tsinkaya a

wani tsari na tsari dangane da matsayin kowane kumburi.

Yawancin nodes akwai, ƙarancin mashahurin matrix algorithm na gargajiya shine, saboda rikitarwa na algorithm yana ƙaruwa tare da adadin.

nodes da ci gaban geometric yana ƙaruwa.Duk da haka, AI ya fi son samun babban ma'auni mai mahimmanci, saboda AI yana da kyau wajen ganowa da kuma

tsinkaya hanyoyin sadarwar da suka fi dacewa.

Ko dai tsinkaya na gaba na Go (AlphaGO na iya hasashen matakai da yawa na gaba, tare da yuwuwar ƙididdigewa ga kowane mataki) ko hasashen yanayin.

na tsarin yanayi mai rikitarwa, daidaiton AI yana da girma fiye da na ƙirar injina.

Dalilin da yasa grid ɗin wutar lantarki a halin yanzu baya buƙatar AI shine adadin nodes a cikin 220 kV da sama da hanyoyin sadarwar wutar lantarki da lardin ke gudanarwa.

aikawa ba ta da girma, kuma an saita sharuɗɗa da yawa don daidaitawa da ɓata matrix ɗin, yana rage ƙwaƙƙwaran ƙididdiga na

tsarin tsarin.

Koyaya, a matakin rarrabuwar wutar lantarki na cibiyar sadarwa, yana fuskantar dubun dubatan ko ɗaruruwan dubban nodes na wutar lantarki, nodes ɗin lodi, da na gargajiya.

Algorithms na matrix a cikin babbar hanyar sadarwar rarraba ba ta da ƙarfi.

Na yi imani cewa ƙirar ƙirar AI a matakin cibiyar sadarwar rarraba zai yiwu a nan gaba.

2. Tari, horo, da kuma samar da bayanan da ba a tsara su ba

Dalili na biyu da ya sa AIGC ta yi nasara shine tarin bayanai.Daga canjin A/D na magana (microphone+PCM

Samfura) zuwa canza A/D na hotuna (CMOS + taswirar sararin samaniya), mutane sun tattara bayanan holographic a cikin gani da sauraro.

filaye a cikin hanyoyi masu rahusa sosai a cikin ƴan shekarun da suka gabata.

Musamman yadda manyan kyamarori da wayoyin komai da ruwanka ke yaduwa, da tarin bayanan da ba a tsara su ba a fagen sauti na gani ga mutane.

a kusan farashin sifili, da kuma tarin abubuwan fashewa na bayanan rubutu akan Intanet shine mabuɗin horon AIGC - saitin bayanan horo ba su da tsada.

Hoton da ke sama yana nuna ci gaban bayanai na duniya, wanda ke ba da fa'ida a fili.

Wannan ci gaban da ba na layi ba na tarin bayanai shine ginshiƙi don ci gaban da ba na layi ba na iyawar AIGC.

AMMA, galibin waɗannan bayanan bayanai ne marasa tsari na sauti da gani, waɗanda aka tara akan sifiri.

A fagen wutar lantarki, ba za a iya cimma hakan ba.Da fari dai, yawancin masana'antar wutar lantarki an tsara su ne da kuma bayanan da aka tsara, kamar su

irin ƙarfin lantarki da halin yanzu, waxanda su ne ma'aunin bayanai na jerin lokaci da tsararru kaɗan.

Saitin bayanan tsarin yana buƙatar fahimtar kwamfutoci kuma suna buƙatar “jeri”, kamar daidaitawar na'ura - ƙarfin lantarki, na yanzu, da bayanan wuta

na maɓalli yana buƙatar daidaitawa zuwa wannan kumburi.

Mafi damuwa shine daidaitawar lokaci, wanda ke buƙatar daidaita wutar lantarki, halin yanzu, da aiki da ƙarfin amsawa dangane da sikelin lokaci, don haka

ana iya yin ganewa na gaba.Hakanan akwai kwatance gaba da baya, waɗanda ke daidaita sararin samaniya a cikin huɗun huɗu.

Ba kamar bayanan rubutu ba, wanda baya buƙatar daidaitawa, ana jefa sakin layi kawai zuwa kwamfutar, wanda ke gano yuwuwar ƙungiyoyin bayanai.

a kanta.

Don daidaita wannan batu, kamar daidaitawar kayan aiki na bayanan rarraba kasuwanci, ana buƙatar daidaitawa akai-akai, saboda matsakaici da matsakaici.

ƙananan hanyar sadarwar rarraba wutar lantarki tana ƙarawa, sharewa, da gyara kayan aiki da layukan yau da kullun, kuma kamfanonin grid suna kashe makudan kuɗin aiki.

Kamar "bayanin bayanai," kwamfutoci ba za su iya yin wannan ba.

Abu na biyu, kudin da ake kashewa wajen siyan bayanai a bangaren wutar lantarki ya yi yawa, kuma ana bukatar na’urori masu auna firikwensin maimakon wayar salula don yin magana da daukar hotuna.”

Duk lokacin da ƙarfin lantarki ya ragu da matakin ɗaya (ko dangantakar rarraba wutar lantarki ta ragu da matakin ɗaya), saka hannun jari na firikwensin da ake buƙata yana ƙaruwa

ta aƙalla tsari ɗaya na girma.Don cimma ma'anar ɗaukar nauyi (ƙarshen capillary) hankali, ya ma fi babban saka hannun jari na dijital. ".

Idan ya zama dole don gano yanayin wucin gadi na grid na wutar lantarki, ana buƙatar babban madaidaicin ƙima mai girma, kuma farashin ya fi girma.

Saboda tsananin tsadar siyar da bayanai da daidaita bayanai, a halin yanzu grid ɗin wutar ba ta iya tara isassun isassun da ba na layi ba.

haɓakar bayanan bayanai don horar da algorithm don isa ga maƙasudin AI.

Ba tare da ambaton buɗewar bayanai ba, ba zai yiwu ba don farawar AI mai ƙarfi don samun waɗannan bayanan.

Sabili da haka, kafin AI, ya zama dole don magance matsalar bayanan bayanan, in ba haka ba ba za a iya horar da lambar AI na gaba ɗaya don samar da AI mai kyau ba.

3. Ci gaba a cikin ikon lissafi

Baya ga algorithms da bayanai, ci gaban singularity na AIGC shima ci gaba ne a cikin ikon lissafi.CPUs na gargajiya ba

dace da manyan-sikelin lokaci guda neuronal kwamfuta.Daidai aikace-aikacen GPUs a cikin wasanni na 3D da fina-finai ne ke yin daidai da manyan sikelin

iyo-point+computing mai yuwuwa.Dokar Moore ta ƙara rage farashin lissafin kowace naúrar ikon lissafi.

Wutar wutar lantarki AI, yanayin da babu makawa a nan gaba

Tare da haɗin kai da yawa na rarraba photovoltaic da kuma rarraba tsarin ajiyar makamashi, da kuma bukatun aikace-aikacen.

Load gefen rumbun wutar lantarki, da gaske ya zama dole don gudanar da tushen da kisa kisa don tsarin cibiyar sadarwar jama'a da mai amfani.

Rarraba (micro) tsarin grid, da kuma ingantaccen ƙarfin wutar lantarki na lokaci-lokaci don rarraba tsarin grid (micro).

Ƙayyadaddun ƙididdiga na gefen cibiyar sadarwar rarraba ya fi girma fiye da tsarin tsarin sadarwar watsawa.Ko don kasuwanci

hadaddun, za a iya samun dubun dubatar na'urori masu ɗaukar nauyi da ɗaruruwan masu sauyawa, da kuma buƙatar aikin cibiyar sadarwa na tushen micro grid/rarrabuwa.

sarrafawa zai tashi.

Tare da ƙarancin farashi na na'urori masu auna firikwensin da kuma yaɗuwar amfani da na'urorin lantarki masu ƙarfi kamar su na'urorin wutar lantarki masu ƙarfi, na'urori masu ƙarfi, da inverters (masu juyawa),

hadewar ji, lissafi, da sarrafawa a gefen grid ɗin wutar lantarki shima ya zama sabon salo.

Saboda haka, AIGC na wutar lantarki shine gaba.Koyaya, abin da ake buƙata a yau ba shine kai tsaye ɗaukar AI algorithm don samun kuɗi ba,

Madadin haka, da farko magance batutuwan gine-ginen bayanai da AI ke buƙata

A cikin haɓakar AIGC, akwai buƙatar samun isasshen tunani game da matakin aikace-aikacen da makomar ikon AI.

A halin yanzu, mahimmancin ikon AI ba shi da mahimmanci: alal misali, algorithm photovoltaic tare da tsinkayar tsinkaya na 90% an sanya shi a cikin kasuwar tabo.

tare da madaidaicin madaidaicin ciniki na 5%, kuma karkatar da algorithm zai shafe duk ribar ciniki.

Bayanan ruwa ne, kuma ikon lissafin algorithm shine tashar.Kamar yadda ya faru, zai kasance.