Ngokwengxelo yeMarike ye-AI yoShishino kunye ne-AI ka-2021-2026 ekhutshwe kutshanje, izinga lokwamkelwa kwe-AI kwiindawo zoshishino linyuke ukusuka kwi-19 ekhulwini ukuya kwi-31 ekhulwini kwiminyaka emibini nje. Ukongeza kwi-31 ekhulwini yabaphenduli abaye baqalisa ngokupheleleyo okanye ngokuyinxenye i-AI kwimisebenzi yabo, abanye abangama-39 ekhulwini ngoku bavavanya okanye bavavanya ubuchwepheshe.
I-AI ivela njengeteknoloji ephambili kubavelisi nakwiinkampani zamandla kwihlabathi liphela, kwaye uhlalutyo lwe-IoT luqikelela ukuba imakethi yezisombululo ze-AI zoshishino iya kubonisa izinga lokukhula konyaka elihlanganisiweyo emva kobhubhane (CAGR) elinamandla lama-35% ukufikelela kwi-$102.17 yezigidigidi ngo-2026.
Ixesha ledijithali lizale i-Intanethi yezinto. Kuyabonakala ukuba ukuvela kobukrelekrele bokwenziwa kuye kwakhawulezisa isantya sophuhliso lwe-Intanethi yezinto.
Makhe sijonge ezinye zezinto ezibangela ukunyuka kwe-AI kunye ne-AIoT yemizi-mveliso.
Ingongoma 1: Izixhobo zesoftware ezingakumbi ze-AIoT yemizi-mveliso
Ngowama-2019, xa uhlalutyo lwe-Iot lwaqala ukugubungela i-AI yemizi-mveliso, bekukho iimveliso ezimbalwa zesoftware ye-AI ezizinikeleyo ezivela kubathengisi betekhnoloji yokusebenza (i-OT). Ukusukela ngoko, abathengisi abaninzi be-OT bangene kwimarike ye-AI ngokuphuhlisa nokubonelela ngezisombululo zesoftware ye-AI ngendlela yamaqonga e-AI kumgangatho wefektri.
Ngokwedatha, phantse abathengisi abangama-400 banikezela ngesoftware ye-AIoT. Inani labathengisi besoftware abajoyina imarike ye-AI yemizi-mveliso linyuke kakhulu kwiminyaka emibini edlulileyo. Ngexesha lophando, i-IoT Analytics ichonge abathengisi abangama-634 betekhnoloji ye-AI kubavelisi/abathengi bemizi-mveliso. Kwezi nkampani, ezingama-389 (61.4%) zibonelela ngesoftware ye-AI.
Iqonga lesoftware entsha ye-AI ligxile kwiindawo zoshishino. Ngaphaya kwe-Uptake, iBraincube, okanye i-C3 AI, inani elikhulayo labathengisi betekhnoloji yokusebenza (OT) banikezela ngamaqonga esoftware ye-AI azinikeleyo. Imizekelo ibandakanya i-ABB's Genix Industrial analytics kunye ne-AI suite, i-Rockwell Automation's FactoryTalk Innovation suite, iqonga lokubonisana ngokuvelisa leSchneider Electric, kwaye kutshanje, izongezo ezithile. Amanye ala maqonga ajolise kuluhlu olubanzi lweemeko zokusetyenziswa. Umzekelo, iqonga le-ABB's Genix libonelela ngohlalutyo oluphambili, kubandakanya izicelo ezakhelwe kwangaphambili kunye neenkonzo zolawulo lokusebenza, ukuthembeka kweeasethi, uzinzo kunye nokusebenza kakuhle kwekhonkco lokubonelela.
Iinkampani ezinkulu zibeka izixhobo zazo zesoftware ye-AI kwindawo yokuthenga.
Ukufumaneka kwezixhobo zesoftware ye-ai kukwaqhutywa zizixhobo zesoftware ezintsha ezisetyenziswayo eziphuhliswe yi-AWS, iinkampani ezinkulu ezifana neMicrosoft kunye neGoogle. Umzekelo, ngoDisemba ka-2020, i-AWS ikhuphe i-Amazon SageMaker JumpStart, uphawu lwe-Amazon SageMaker olubonelela ngezisombululo ezakhiwe kwangaphambili nezinokwenziwa ngokwezifiso kwiimeko eziqhelekileyo zokusetyenziswa kwemizi-mveliso, ezifana ne-PdM, umbono wekhompyutha, kunye nokuqhuba ngokuzimela, i-Deploy ngokucofa nje okumbalwa.
Izisombululo zesoftware ezisebenzisa umzekelo othile ziqhuba ukuphuculwa kokusebenziseka.
Ii-software ezisetyenziswa kwiimeko ezithile, ezifana nezo zigxile ekugcinweni kwangaphambili, ziya zixhaphaka. I-IoT Analytics iqaphele ukuba inani lababoneleli abasebenzisa izisombululo zesoftware yolawulo lwedatha yemveliso esekwe kwi-AI (PdM) lenyuke laya kuma-73 ekuqaleni kuka-2021 ngenxa yokwanda kwemithombo eyahlukeneyo yedatha kunye nokusetyenziswa kweemodeli zoqeqesho lwangaphambi koqeqesho, kunye nokwamkelwa ngokubanzi kobuchwepheshe bokuphucula idatha.
Ingongoma 2: Uphuhliso kunye nokugcinwa kwezisombululo ze-AI kwenziwa lula
Ukufunda koomatshini okwenziwa ngokuzenzekelayo (i-AutoML) kuya kuba yimveliso eqhelekileyo.
Ngenxa yobunzima bemisebenzi enxulumene nokufunda koomatshini (ML), ukukhula ngokukhawuleza kwezicelo zokufunda koomatshini kudale isidingo seendlela zokufunda koomatshini ezingasetyenziswa ngaphandle kobuchule. Intsimi yophando ephumayo, ukuziqhelanisa okuqhubekayo kokufunda koomatshini, ibizwa ngokuba yi-AutoML. Iinkampani ezahlukeneyo zisebenzisa obu buchwephesha njengenxalenye yeminikelo yazo ye-AI ukunceda abathengi baphuhlise iimodeli ze-ML kwaye basebenzise iimeko zokusetyenziswa kwemizi-mveliso ngokukhawuleza. NgoNovemba ka-2020, umzekelo, i-SKF yabhengeza imveliso esekwe kwi-automL edibanisa idatha yenkqubo yomatshini nedatha yokungcangcazela kunye nobushushu ukunciphisa iindleko kunye nokwenza iimodeli ezintsha zoshishino kubathengi.
Imisebenzi yokufunda koomatshini (i-ML Ops) yenza kube lula ulawulo kunye nokugcinwa kwemodeli.
Le ndlela intsha yokusebenza kokufunda koomatshini ijolise ekwenzeni lula ukugcinwa kweemodeli ze-AI kwiindawo zokuvelisa. Ukusebenza kwemodeli ye-AI kudla ngokuwohloka ngokuhamba kwexesha njengoko kuchatshazelwa zizinto ezininzi ngaphakathi kwisityalo (umzekelo, utshintsho ekusasazweni kwedatha kunye nemigangatho yomgangatho). Ngenxa yoko, ukugcinwa kweemodeli kunye nemisebenzi yokufunda koomatshini kuye kwaba yimfuneko ukuhlangabezana neemfuno zomgangatho ophezulu kwiindawo zoshishino (umzekelo, iimodeli ezinentsebenzo engaphantsi kwe-99% zisenokungakwazi ukuchonga indlela yokuziphatha ebeka ukhuseleko lwabasebenzi emngciphekweni).
Kwiminyaka yakutshanje, iinkampani ezininzi ezintsha ziye zajoyina indawo ye-ML Ops, kuquka iDataRobot, iGrid.AI, iPinecone/Zilliz, iSeldon, kunye neWeights & Biases. Iinkampani ezisele ziqalisiwe zongeze imisebenzi yokufunda koomatshini kwizibonelelo zazo zesoftware ye-AI ezikhoyo, kuquka iMicrosoft, eyazisa ukufunyanwa kwedatha kwi-Azure ML Studio. Olu phawu lutsha luvumela abasebenzisi ukuba babone utshintsho ekusasazweni kwedatha yokufaka eyonakalisa ukusebenza kwemodeli.
Ingongoma 3: Ubukrelekrele bokwenziwa obusetyenziswayo kwizicelo ezikhoyo kunye namatyala okusetyenziswa
Ababoneleli besoftware bemveli bongeza amandla e-AI.
Ukongeza kwizixhobo zesoftware ze-AI ezinkulu ezikhoyo ezifana ne-MS Azure ML, i-AWS SageMaker, kunye ne-Google Cloud Vertex AI, iisuti zesoftware zemveli ezifana neComputerized Maintenance Management Systems (CAMMS), iinkqubo zokwenziwa kweMveliso (MES) okanye ucwangciso lwezixhobo zeshishini (ERP) ngoku zinokuphucuka kakhulu ngokufaka amandla e-AI. Umzekelo, umboneleli we-ERP i-Epicor Software yongeza amandla e-AI kwiimveliso zayo ezikhoyo nge-Epicor Virtual Assistant yayo (EVA). Iiarhente ze-EVA ezikrelekrele zisetyenziselwa ukwenza iinkqubo ze-ERP ngokuzenzekelayo, ezinje ngokuhlela kwakhona imisebenzi yokuvelisa okanye ukwenza imibuzo elula (umzekelo, ukufumana iinkcukacha malunga namaxabiso emveliso okanye inani leendawo ezikhoyo).
Amatyala okusetyenziswa kwemizi-mveliso ayaphuculwa ngokusebenzisa i-AIoT.
Amatyala amaninzi okusetyenziswa kwemizi-mveliso ayaphuculwa ngokongeza amandla e-AI kwiziseko zehardware/software ezikhoyo. Umzekelo ocacileyo kukubona komatshini kwizicelo zolawulo lomgangatho. Iinkqubo zokubona koomatshini zemveli zicubungula imifanekiso ngeekhompyutha ezidityanisiweyo okanye ezizimeleyo ezixhotyiswe ngesoftware ekhethekileyo evavanya iiparameter kunye nemida emiselweyo (umz., umahluko ophezulu) ukumisela ukuba izinto zibonisa iziphene na. Kwiimeko ezininzi (umzekelo, izinto ze-elektroniki ezineemilo ezahlukeneyo zocingo), inani leeposi ezingezizo liphezulu kakhulu.
Nangona kunjalo, ezi nkqubo zivuselelwa ngobukrelekrele bokwenziwa. Umzekelo, umboneleli we-industrial machine Vision uCognex ukhuphe isixhobo esitsha se-Deep Learning (Vision Pro Deep Learning 2.0) ngoJulayi 2021. Ezi zixhobo zintsha zidibana neenkqubo ze-visual zemveli, zivumela abasebenzisi bokugqibela ukuba badibanise i-deep learning nezixhobo ze-visual zemveli kwisicelo esifanayo ukuhlangabezana neendawo zonyango neze-elektroniki ezifuna ukulinganiswa ngokuchanekileyo kwemikrwelo, ungcoliseko kunye nezinye iziphene.
Ingongoma 4: Izixhobo ze-AIoT zemizi-mveliso ziyaphuculwa
Iitships ze-AI ziyaphucuka ngokukhawuleza.
Iitships ze-AI zehardware ezifakiweyo zikhula ngokukhawuleza, kunye neendlela ezahlukeneyo ezikhoyo zokuxhasa uphuhliso kunye nokusasazwa kweemodeli ze-AI. Imizekelo ibandakanya iiyunithi zamva nje ze-NVIDIA zokucubungula imizobo (i-Gpus), i-A30 kunye ne-A10, ezaziswa ngoMatshi 2021 kwaye zifanelekile kwiimeko zokusetyenziswa kwe-AI ezifana neenkqubo zengcebiso kunye neenkqubo zokubona ikhompyutha. Omnye umzekelo yi-Google's fourth-generation Tensors Processing Units (TPus), eziziisekethe ezidibeneyo zeenjongo ezikhethekileyo (ii-ASics) ezinamandla ezinokufikelela kwi-1,000 ngokusebenza kakuhle kunye nesantya ekuphuhlisweni kwemodeli kunye nokusasazwa kwemithwalo ethile ye-AI (umz., ukufunyanwa kwezinto, ukuhlelwa komfanekiso, kunye nemilinganiselo yengcebiso). Ukusebenzisa i-hardware ye-AI ezinikeleyo kunciphisa ixesha lokubala imodeli ukusuka kwiintsuku ukuya kwimizuzu, kwaye kubonakalise ukuba kutshintsha umdlalo kwiimeko ezininzi.
Izixhobo ze-AI ezinamandla zifumaneka ngoko nangoko ngemodeli yokuhlawula ngokusetyenziswa.
Amashishini eSuperscale ahlala ephucula iiseva zawo ukuze enze izixhobo zekhompyutha zifumaneke kwilifu ukuze abasebenzisi bokugqibela bakwazi ukuphumeza usetyenziso lwe-AI yemizi-mveliso. NgoNovemba ka-2021, umzekelo, i-AWS yabhengeza ukukhutshwa ngokusemthethweni kweziganeko zayo zamva nje ezisekelwe kwiGPU, i-Amazon EC2 G5, eqhutywa yi-NVIDIA A10G Tensor Core GPU, kwiintlobo ngeentlobo zezicelo ze-ML, kubandakanya iinjini zombono wekhompyutha kunye neengcebiso. Umzekelo, umboneleli weenkqubo zokufumanisa iNanotronics usebenzisa imizekelo ye-Amazon EC2 yesisombululo sayo solawulo lomgangatho esisekelwe kwi-AI ukukhawulezisa imizamo yokucubungula kunye nokufezekisa amazinga okufumanisa achanekileyo ngakumbi ekuvelisweni kwee-microchips kunye nee-nanotubes.
Isiphelo kunye nethemba
I-AI iphuma kumzi-mveliso, kwaye iya kuba yindawo yonke kwizicelo ezintsha, ezifana ne-AI-based PdM, kwaye njengophuculo kwisoftware ekhoyo kunye neemeko zokusetyenziswa. Amashishini amakhulu aqalisa iimeko ezininzi zokusetyenziswa kwe-AI kwaye axela impumelelo, kwaye uninzi lweeprojekthi lunembuyekezo ephezulu kutyalo-mali. Lilonke, ukunyuka kwelifu, amaqonga e-iot kunye neetships ze-AI ezinamandla kubonelela ngeqonga lesizukulwana esitsha sesoftware kunye nokwenza ngcono.
Ixesha lokuposa: Jan-12-2022