InterSystems IRIS - universal real-time AI/ML chikuva

Munyori: Sergey Lukyanchikov, nyanzvi yekuongorora paInterSystems

Chaiyo-nguva AI/ML komputa mafoni

Ngatitangei nemienzaniso kubva muchiitiko cheData Sayenzi kudzidzira paInterSystems:

• Iyo yakatakura mutengi portal yakabatana kune online kurudziro system. Pachave nekugadziridzwa kwekukwidziridzwa kune network yekutengesa (semuenzaniso, pachinzvimbo che "flat" mutsara wekusimudzira, "segment-tactics" matrix ikozvino ichashandiswa). Chii chinoitika kune injini dzekukurudzira? Chii chinoitika kune kuendesa uye kuvandudzwa kwedata kune injini yekukurudzira (iyo vhoriyamu yekupinza data yakawedzera ne25000 nguva)? Chii chinoitika pakuvandudzwa kwekurudziro (chido chekudzikisa chikumbaridzo chekusefa chemitemo yekurudziro nechuru nekuda kwekuwedzera kwechuru muhuwandu hwavo uye "renji")?
• Pane hurongwa hwekutarisa mukana wehurema huri kukura mumidziyo yemidziyo. Iyo automated process control system yakabatana kune yekutarisa sisitimu, ichiendesa zviuru zvetekinoroji maitiro paramita sekondi yega yega. Chii chinoitika kune yekutarisa sisitimu yakamboshanda pa "manual samples" (inokwanisa here kupa yechipiri-ne-sekondi-yechipiri mukana wekutarisa)? Chii chichaitika kana chivharo chitsva chemazana akati wandei makoramu chikaonekwa mune yekuisa data nekuverenga kubva kuma sensors achangobva kuwedzerwa kune process control system (zvichave zvichidikanwa uye kwenguva yakareba sei kumisa iyo yekutarisa system kuti ubatanidze data kubva kune nyowani sensors mukuongorora. )?
• Seti yeAI / ML nzira (kurudziro, kutarisa, kufanotaura) yakagadzirwa iyo inoshandisa mhedzisiro yebasa reumwe neumwe. Mangani emaawa emunhu anodiwa mwedzi wega wega kugadzirisa kushanda kweiyi yakaoma kune shanduko mune yekupinza data? Chii chinonzi "kuderera" kana chichitsigirwa nehutungamiri hwekuita sarudzo yakaoma (nguva yakawanda yekuitika kwemashoko matsva anotsigira mairi maererano nehuwandu hwekuitika kwe data itsva yekupinza)?

Tichipfupikisa iyi nemimwe mienzaniso yakawanda, tauya pakugadzirisa matambudziko anomuka kana tichienda kukushandiswa kwemuchina wekudzidza uye nzira dzehungwaru dzekugadzira munguva chaiyo:

• Isu takagutsikana here nekumhanya kwekugadzira uye kugadzirisa (kune shanduko yemamiriro) yeAI/ML budiriro mukambani yedu?
• Yakawanda sei iyo AI/ML mhinduro dzatinoshandisa dzinotsigira chaiyo-nguva bhizinesi manejimendi?
• Iyo AI/ML mhinduro dzatinoshandisa dzinokwanisa kuzvimiririra (pasina vanogadzira) kuchinjika kune shanduko mudata uye bhizinesi manejimendi maitiro?

Chinyorwa chedu chitarisiko chakakwana chekugona kweInterSystems IRIS papuratifomu maererano nerutsigiro rwepasirese rwekuendeswa kweAI/ML michina, kusangana (kubatanidzwa) kweAI/ML mhinduro, uye kudzidziswa (kuyedza) kweAI/ML mhinduro pane yakasimba. data hova. Tichatarisa kutsvagisa kwemusika, nyaya dzezvidzidzo zveAI/ML mhinduro, uye pfungwa dzezvatinodaidza kuti chaiyo-nguva AI/ML chikuva mune ino chinyorwa.

Zvatinoziva kubva kuongororo: real-time applications

Mhinduro ongororoyakaitwa pakati pevanoda kusvika mazana masere enyanzvi dzeIT muna 800 naLightbend, vanozvitaurira:

Mufananidzo 1 Kutungamira vatengi venguva chaiyo data

Ngatitorei zvidimbu zvakakosha zvemushumo pane zvakabuda muongororo iyi mushanduro yedu:

"... Mafambiro mukuzivikanwa kwezvishandiso zvekubatanidza hova dze data uye, panguva imwe chete, kutsigira komputa mumidziyo inopa mhinduro yekubatana kuchikumbiro chemusika chekuda kuteedzera, kunzwisisika, kusimba kwechikumbiro chemhinduro dzinoshanda. Yekufambisa data inotamisa ruzivo nekukurumidza kupfuura yechinyakare packet data. Kuwedzerwa kune izvi ndiko kugona kukurumidza kushandisa nzira dzemakomputa, senge, semuenzaniso, AI/ML-yakavakirwa mazano, kugadzira mabhenefiti emakwikwi kuburikidza nekuwedzera kugutsikana kwevatengi. Mujaho weagility unobatawo mabasa ese muDevOps paradigm - ichiita kuvandudza kwekushandisa uye kutumira zvakanyanya. … Mazana masere neina nyanzvi dzeIT dzakapa ruzivo rwekushandiswa kwekuyerera kwedata mumasangano avo. Vakapindura vainyanya kuwanikwa munyika dzekuMadokero (41% muEurope uye 37% kuNorth America) uye vakada kugovaniswa zvakaenzana pakati pemakambani madiki, epakati uye makuru. ...

... Artificial intelligence haisi hype. Makumi mashanu nesere muzana yeavo vari kutoshandisa data stream processing mukugadzirwa kweAI/ML application inosimbisa kuti kushandisa kwavo AI/ML kuchaona kukura kukuru mugore rinotevera (zvichienzaniswa nemamwe maapplication).

• Zvinoenderana neruzhinji rwevakapindura, kushandiswa kwedhata hova muAI / ML mamiriro achaona kukura kukuru mugore rinotevera.
• Zvishandiso muAI/ML zvichakura kwete nekuda kwemhando nyowani dzemamiriro, asiwo nekuda kwechinyakare mamiriro umo data-nguva-chaiyo iri kuwedzera kushandiswa.
• Pamusoro peAI/ML, nhanho yekufarira pakati pevashandisi veIoT data pombi inoshamisa - 48% yeavo vakatobatanidza data reIoT vanoti mamiriro ekuita pane iyi data achaona kuwedzera kukuru munguva pfupi iri kutevera. ..."

Kubva pane iyi ongororo inonakidza, zviri pachena kuti maonero ekudzidza kwemuchina uye mamiriro ehungwaru ekunyepedzera sevatungamiriri mukushandiswa kwemadhata atove "munzira." Asi kucherechedzwa kwakakosha zvakaenzana maonero eiyo chaiyo-nguva AI/ML kuburikidza nelenzi yeDevOps: pano isu tinogona kutotanga kutaura nezve shanduko yetsika ichiri kutonga ye "inoraswa AI/ML ine data rinowanikwa zvizere."

Chaiyo-nguva AI/ML papuratifomu pfungwa

Imwe yakajairika nzvimbo yekunyorera yechaiyo-nguva AI/ML inzira yekudzora mukugadzira. Tichishandisa muenzaniso wake uye tichifunga nezvemifungo yakapfuura, isu tichagadzira iyo pfungwa yeiyo chaiyo-nguva AI/ML chikuva.
Iko kushandiswa kwehungwaru hwekugadzira uye kudzidza muchina mukutonga kwemaitiro kune akati wandei maficha:

• Dhata pamusoro peiyo mamiriro eiyo tekinoroji maitiro anogamuchirwa zvakanyanya: neakakwira frequency uye neyakasiyana siyana paramita (kusvika makumi ezviuru emakumi ezviuru zveparameter tsika inofambiswa pasekondi kubva kune process control system)
• Dhata pamusoro pekuzivikanwa kwekuremara, tisingataure nezve data pakukura kwavo, pane zvinopesana, kushomeka uye kusajairika, inoratidzirwa nekusakwana typification yehurema uye nzvimbo yavo munguva (inowanzo miririrwa nepepa rekodhi)
• Kubva pamaonero anoshanda, chete "hwindo rekukosha" reiyo sosi data rinowanikwa pakudzidziswa nekushandisa mamodheru, zvichiratidza masimba eiyo tekinoroji maitiro pamusoro penguva inonzwisisika inotsvedza inopera neyekupedzisira kuverenga kukosha kweiyo maitiro paramita.

Aya maficha anotimanikidza, pamusoro pekugamuchira uye kwekutanga kugadzirisa munguva chaiyo yeyakadzika "broadband yekuisa chiratidzo" kubva kune tekinoroji maitiro, kuita (pamwe chete) mashandisirwo, kudzidziswa uye kutonga kwemhando yemhedzisiro yeAI / ML modhi - zvakare munguva chaiyo. Iyo "frame" iyo mamodheru edu "inoona" muhwindo rinotsvedza rekukosha rinogara richichinja - uye nayo, mhando yemhedzisiro yebasa reAI / ML modhi yakadzidziswa pane imwe ye "mafuremu" munguva yakapfuura inochinjawo. . Kana mhando yemhedzisiro yebasa reAI/ML modhi yakadzikira (semuenzaniso: kukosha kweiyo "alarm-norm" yekukanganisa yekuronga kwapfuura miganhu yatakatsanangura), kumwe kudzidziswa kwemamodheru kunofanirwa kutangwa otomatiki. imwe yazvino "furemu" - uye sarudzo yenguva yekutanga kudzidziswa kwemamwe mamodheru inofanirwa kufunga nezve iyo nguva yekudzidziswa pachayo, uye simba rekuderera kwemhando yebasa reiyo vhezheni yemhando (sezvo. shanduro dzemazuva ano dzemhando dzinoramba dzichishandiswa apo mienzaniso inodzidziswa, uye kusvikira shanduro dzavo "dzichangobva kudzidziswa" dzaumbwa).

InterSystems IRIS ine yakakosha papuratifomu kugona kugonesa AI/ML mhinduro dzechaiyo-nguva maitiro ekutonga. Aya masimba anogona kukamurwa kuita mapoka makuru matatu:

• Kuenderera mberi kuendeswa (Kuenderera mberi Kuendeswa / Kuendesa, CD) yezvitsva kana zvakagadziridzwa zviripo AI/ML masisitimu kuita mhinduro inobereka inoshanda munguva chaiyo paInterSystems IRIS papuratifomu.
• Kuenderera mberi Kubatanidzwa (CI) mune imwe chete inogadzira mhinduro yeinouya tekinoroji maitiro data hova, mitsetse yedata yekushandisa/kudzidziswa/kudzora kwemhando yeAI/ML nzira uye kuchinjana kwedata/kodhi/kudzora zviito nenzvimbo dzemasvomhu dzekuenzanisira, dzakarongwa munguva chaiyo. papuratifomu InterSystems IRIS
• Kuenderera mberi (kuzvidzora) kudzidzisa (Kunoenderera Kudzidzisa, CT) yeAI / ML nzira, dzinoitwa munzvimbo dzemasvomhu dzekuenzanisa uchishandisa data, kodhi uye zviito zvekutonga ("zvisarudzo zvakaitwa") zvinoparidzirwa neInterSystems IRIS platform.

Kurongeka kwehunyanzvi hwepuratifomu zvine chekuita nekudzidza kwemuchina uye hungwaru hwekugadzira mumapoka aya hazvina tsaona. Ngatitaurei nezve methodological chinyorwa Google, iyo inopa hwaro hwepfungwa hwechikamu ichi, mushanduro yedu:

"... Pfungwa yeDevOps, yakakurumbira mazuva ano, inovhara kuvandudzwa nekushanda kwehurongwa hwemashoko makuru. Zvakanakira kuita iyi pfungwa kudzikiswa kwenguva yemafambisirwo ebudiriro, kukurumidza kuendesa zviitiko, uye kushanduka mukuronga kuburitsa. Kuti uwane mabhenefiti aya, DevOps inosanganisira kuita kanenge maviri maitiro:

• Kuenderera mberi Kubatanidzwa (CI)
• Kuenderera mberi nekutumira (CD)

Aya maitiro anoshandawo kune AI/ML mapuratifomu kuve nechokwadi chekuvimbika uye kuita gungano rekugadzira AI/ML mhinduro.

AI/ML mapuratifomu anosiyana kubva kune mamwe masisitimu eruzivo mune zvinotevera zvinhu:

• Kugona kwechikwata: Paunenge uchigadzira mhinduro yeAI/ML, timu inowanzo sanganisira masayendisiti e data kana nyanzvi dze "dzidzo" mumunda wekutsvagisa data vanoongorora data, kugadzira uye kuyedza mhando. Idzi nhengo dzechikwata dzinogona kunge dzisiri nyanzvi dzinogadzira makodhi madhizaini.
• Kubudirira: AI / ML injini dzinoedza mune zvakasikwa. Kuti ugadzirise dambudziko nenzira yakanyatso shanda, zvinodikanwa kuti upfuure nemusanganiswa wakasiyana-siyana wekuisa mabhii, maalgorithms, nzira dzekuenzanisira uye modhi paramita. Kuoma kwekutsvaga kwakadaro kunowanikwa mukutsvaga "chakashanda / chisina kushanda", kuve nechokwadi chekuberekazve kwezvikamu, kugadzirwa kwezvinoitika zvekudzokorora zvakare.
• Kuedza: Kuedza mainjini eAI/ML kunoda huwandu hwakawanda hwebvunzo kupfuura zvimwe zvakawanda zvinobudirira. Pamusoro peyakajairwa yuniti uye yekubatanidza bvunzo, chokwadi chedata uye mhando yemhedzisiro yekushandisa modhi pakudzidziswa uye kutonga masampuli anoedzwa.
• Deployment: Deployment of AI/ML solutions is not limited to predictive services inoshandisa yakambodzidziswa modhi. AI/ML mhinduro dzakavakwa dzakatenderedza-akawanda-nhanho mapaipi ayo anoita otomatiki modhi kudzidzisa uye kushandisa. Kuisa mapaipi akadaro kunosanganisira kuita otomatiki zviito zvisiri-zvidiki zvagara zvichiitwa nemaoko nemasaenzi edata kuitira kuti vakwanise kudzidzisa uye kuyedza mhando.
• Kubudirira: AI / ML injini dzinogona kushaya chibereko kwete chete nekuda kwekusashanda hurongwa, asiwo nekuda kwekugara kuchichinja kweiyo data yekuisa. Mune mamwe mazwi, mashandiro eAI/ML maitiro anogona kudzikisira nekuda kwezvikonzero zvakawanda kupfuura kuita kwezvakajairwa zviitiko. Izvo zvinotungamira kune kudikanwa kwekutarisa (pamhepo) mashandiro eAI / ML injini dzedu, pamwe nekutumira zviziviso kana kuramba mhinduro kana zviratidzo zvekushanda zvisingaite zvinotarisirwa.

AI/ML mapuratifomu akafanana nemamwe masisitimu eruzivo mukuti ese ari maviri anoda kuenderera mberi kodhi kubatanidza neshanduro yekutonga, bvunzo dzeyuniti, yekubatanidza kuyedza, uye kuenderera mberi kwekusimudzira. Nekudaro, mune iyo AI/ML, pane akati wandei akakosha misiyano:

• CI (Kuramba Kubatanidzwa) haichagumiri pakuyedza uye kusimbisa kodhi yezvikamu zvakatumirwa - inosanganisirawo kuyedza uye kusimbisa data uye AI/ML modhi.
• CD (Kuenderera mberi Kuendesa / Kuendesa, kuenderera mberi kwekuendesa) hakugumiri pakunyora nekuburitsa mapakeji kana masevhisi, asi zvinoreva chikuva chekugadzira, kudzidziswa uye kushandiswa kweAI/ML mhinduro.
• CT (Kuramba Kudzidziswa, kuenderera mberi kudzidziswa) chinhu chitsva [approx. munyori wechinyorwa: chinhu chitsva chine chekuita neiyo yechinyakare pfungwa yeDevOps, umo CT iri, sekutonga, Kuenderera Kuedzwa], inowanikwa muAI/ML mapuratifomu, inotarisira kuzvitonga kwemaitiro ekudzidzisa nekushandisa AI. /ML mhando. ..."

Tinogona kutaura kuti kudzidza kwemichina uye hungwaru hwekugadzira hunoshanda pane chaiyo-nguva data inoda seti yakafara yezvishandiso uye hunyanzvi (kubva pakuvandudza kodhi kusvika pakurongeka kwemasvomhu ekuenzanisira nharaunda), kusanganisa kwepedyo pakati penzvimbo dzese dzinoshanda uye dzezvidzidzo, sangano rinoshanda revanhu uye. muchina zviwanikwa.

Chaiyo-nguva mamiriro: kuziva kukura kwekuremara mumapombi ekudya

Kuenderera mberi nekushandisa iyo nzira yekudzora nzvimbo semuenzaniso, funga nezvedambudziko chairo (takambotaura pakutanga): isu tinofanirwa kupa chaiyo-nguva yekutarisa kuvandudzwa kwehurema mumapombi zvichienderana nekuyerera kwemaitiro parameter maitiro uye mishumo kubva kuvashandi vekugadzirisa nezve hurema hwakaonekwa.

Mufananidzo 2 Dambudziko rekugadzirisa kutarisa kukura kwekukanganisa

Chinhu chemabasa mazhinji anounzwa nenzira iyi mukuita ndechekuti kugara uye kushanda zvakanaka kwekugamuchira data (APCS) kunofanirwa kutariswa zvichipesana neseri kwechiitiko cheepisodic uye chisingaite (uye kunyoreswa) kwehurema hwemhando dzakasiyana. Mune mamwe mazwi: data kubva kuhurongwa hwekutonga hunosvika kamwechete kwechipiri, yakarurama uye yakarurama, uye zvinyorwa pamusoro pekuremara zvinogadzirwa nepenzura yekemikari inoratidza zuva riri mubhuku rakawanda mumusangano (semuenzaniso: "12.01 - kunyura muchivharo. kubva kudivi rechitatu ").

Nokudaro, isu tinogona kuwedzera kuumbwa kwechinetso nekupikiswa kwakakosha kunotevera: isu tine "label" imwe chete yekuremara kweimwe mhando (kureva, muenzaniso wekukanganisa kweimwe mhando inomiririrwa nedata kubva padanho rekutonga. system pane yakatarwa zuva - uye isu hatina mimwe mienzaniso yekukanganisa kweiyi rudzi). Uku kudzikisira kunobva kwatitora kupfuura chiyero chekirasi yekudzidza muchina (inotariswa kudzidza), iyo panofanirwa kuve ne "ma tag" akawanda.

Mufananidzo 3 Kujekeswa kwebasa rekutarisa kukura kwekuremara

Tinogona here neimwe nzira "kuwanza" iyo chete "tagi" yatiinayo? Hongu tinogona. Ikozvino mamiriro epombi anoratidzirwa nehuwandu hwekufanana nekuremara kwakanyoreswa. Kunyangwe pasina kushandiswa kwehuwandu nzira, pamwero wekuona kwekuona, nekutarisa dynamics yedata data inosvika kubva kuhurongwa hwekutonga system, unogona kutodzidza zvakawanda:

Mufananidzo 4 Masimba epombi mamiriro ekumashure kwe "mucherechedzo" wehurema hwemhando yakapihwa.

Asi maonero ekuona (zvirinani ikozvino) haisi iyo yakanyatsokodzera jenareta ye "ma tags" mumamiriro edu ari kukurumidza kuchinja. Tichaongorora kufanana kwepombi yepombi yemazuva ano kune zvakasvibiswa zvakasvibiswa tichishandisa nhamba yekuongorora.

Mufananidzo 5 Kushandisa bvunzo dzenhamba kune data rinouya richipesana nekumashure kwechirema "label"

Muedzo wenhamba unotarisa mukana wekuti marekodhi ane kukosha kweiyo tekinoroji maitiro paramita mu "flow-packet" inogamuchirwa kubva ku process control system yakafanana nemarekodhi e "tag" yekukanganisa kweimwe mhando. Iko kukosha kwehuwandu (huwandu hwekufanana kwenhamba) yakaverengerwa semhedzisiro yekushandisa bvunzo yenhamba inoshandurwa kuita kukosha kwe0 kana 1, ichiva "chinyorwa" chekudzidza kwemichina mune rekodhi yega yega mupaketi iri kuongororwa kufanana. Ndokunge, mushure mekugadzirisa pasuru ichangobva kugashirwa yepombi mamiriro erekodhi ine bvunzo yehuwandu, isu tine mukana we (a) wekuwedzera iyi pasuru kune yekudzidziswa seti yekudzidzisa iyo AI / ML modhi uye (b) kuita kutonga kwemhando ye yazvino vhezheni yemodhi kana uchiishandisa kune iyi package.

Mufananidzo 6 Kushandisa modhi yekudzidza yemuchina kune data rinouya richipesana nekumashure kwechirema "label"

Mune imwe yedu yapfuura webinars Isu tinoratidza uye tinotsanangura kuti iyo InterSystems IRIS chikuva inobvumidza iwe kuita chero AI/ML mashandiro nenzira yekuenderera mberi nekuita bhizinesi maitiro anotarisisa kuvimbika kwemhedzisiro yekuenzanisira uye kugadzirisa modhi paramita. Kana tichiita prototype yemamiriro edu nemapombi, tinoshandisa ese InterSystems IRIS mashandiro akaunzwa panguva yewebhu - kuita muanalyzer maitiro sechikamu chemhinduro yedu kwete yekirasi inotariswa kudzidza, asi kusimbaradza kudzidza, iyo inongogadzirisa sarudzo yemhando dzekudzidzisa. . Sample yekudzidziswa ine marekodhi ekuti "yekuona kubvumirana" inomuka mushure mekushandisa zvese zviri zviviri bvunzo yenhamba uye yazvino vhezheni yemuenzaniso - kureva, zvese bvunzo yenhamba (mushure mekushandura indekisi yekufanana kuita 0 kana 1) uye modhi yakaburitsa mhedzisiro. pane zvinyorwa zvakadaro 1. Panguva yekudzidziswa kutsva kwemuenzaniso, panguva yekusimbiswa kwayo (iyo ichangobva kudzidziswa muenzaniso inoshandiswa kune yayo yekudzidzira muenzaniso, nekutanga kushandiswa kwenhamba yekuongorora kune iyo), zvinyorwa "zvisina kuchengetedza" mhedzisiro 1 mushure mekugadzirisa. nekuongororwa kwenhamba (nekuda kwekuvapo nguva dzose mukudzidziswa sampuli yezvinyorwa kubva kune yepakutanga "label" yehurema), inobviswa kubva pakudzidziswa seti, uye shanduro itsva yemuenzaniso inodzidza kubva ku "label" ye kukanganisa pamwe ne "zvakachengetwa" zvinyorwa kubva murukova.

Mufananidzo 7 Robotization yeAI / ML kuverenga muInterSystems IRIS

Kana paine kudikanwa kwerudzi rwe "pfungwa yechipiri" pamhando yekuonekwa yakawanikwa panguva yekuverenga kwenzvimbo muInterSystems IRIS, chirongwa cheanopa zano chinogadzirwa kuti chiite kudzidziswa uye kushandisa mamodheru padhata rekutonga uchishandisa makore masevhisi (semuenzaniso, Microsoft. Azure, Amazon Web Services, Google Cloud Platform, nezvimwewo):

Mufananidzo 8 Chechipiri Pfungwa kubva kuMicrosoft Azure yakarongwa neInterSystems IRIS

Iyo prototype yemamiriro edu ezvinhu muInterSystems IRIS yakagadzirirwa seajenti-yakavakirwa sisitimu yekuongorora maitiro ayo anopindirana nechinhu chemidziyo (pombi), masvomhu ekuenzanisira nharaunda (Python, R naJulia), uye ive nechokwadi chekuzvidzidza kwevose vanobatanidzwa AI/ ML maitiro - pane chaiyo-nguva data hova .

Mufananidzo 9 Kushanda kukuru kweiyo chaiyo-nguva AI / ML mhinduro muInterSystems IRIS

Mhedzisiro inoshanda yeprototype yedu:

• Chirema chemuenzaniso chinozivikanwa nemuenzaniso (Ndira 12):

• Chirema chinokura chakazivikanwa nemuenzaniso icho chisina kuverengerwa mumuenzaniso (Gunyana 11, chirema pachacho chakaonekwa neboka rekugadzirisa mazuva maviri chete gare gare, munaGunyana 13):

Simulation pane chaiyo data ine akati wandei episodhi yehurema hwakafanana yakaratidza kuti mhinduro yedu, yakaitwa paInterSystems IRIS papuratifomu, inotibvumira kuona kuvandudzwa kwehurema hwemhando iyi mazuva akati wandei zvisati zvaonekwa neboka rekugadzirisa.

InterSystems IRIS - yepasirese chaiyo-nguva AI/ML komputa chikuva

Iyo InterSystems IRIS chikuva inorerutsa kusimudzira, kutumira uye kushanda kweiyo-chaiyo-nguva data mhinduro. InterSystems IRIS inokwanisa kuita panguva imwe chete kuita transaction uye analytical data processing; tsigiro yakawiriraniswa data maonero zvichienderana nemhando dzakawanda (kusanganisira hukama, hierarchical, chinhu uye gwaro); ita sepuratifomu yekubatanidza huwandu hwakawanda hwe data masosi uye maapplication ega; ipa yepamusoro-chaiyo-nguva analytics pane yakarongeka uye isina kurongeka data. InterSystems IRIS inopawo maitiro ekushandisa ekunze ekuongorora maturusi uye inobvumira inoshanduka musanganiswa wekutambira mugore uye pamaseva emuno.

Zvikumbiro zvakavakirwa paInterSystems IRIS papuratifomu zvinoiswa munzvimbo dzakasiyana siyana maindasitiri, zvichibatsira makambani kuona zvakakosha zvehupfumi kubva muhungwaru uye mashandiro emaonero, kuwedzera kuita sarudzo ine ruzivo uye kuvhara mikaha pakati pechiitiko, kuongorora uye kuita.

Mufananidzo 10 InterSystems IRIS dhizaini mune yechokwadi-nguva AI/ML

Kufanana nedhayagiramu yapfuura, dhayagiramu iri pasi apa inobatanidza itsva "coordinate system" (CD / CI / CT) nemufananidzo wekuyerera kwemashoko pakati pezvinhu zvinoshanda zvepuratifomu. Iko kuona kunotanga ne macromechanism CD uye inoenderera neiyo macromechanism CI uye CT.

Mufananidzo 11 Dhiagiramu yeruzivo inoyerera pakati peAI / ML zvinhu zveInterSystems IRIS chikuva.

Izvo zvakakosha zveCD mechanism muInterSystems IRIS: vashandisi vepuratifomu (vagadziri veAI/ML mhinduro) vanogadzirisa zviripo uye/kana kugadzira mitsva yeAI/ML mitsva vachishandisa yakasarudzika kodhi dhizaini yeAI/ML maitiro: Jupyter (zita rizere: Jupyter Notebook; Nekupfupika, magwaro akagadzirwa mune ino edhita anodaidzwawo dzimwe nguva). MuJupyter, mugadziri ane mukana wekunyora, kugadzirisa uye kuona mashandiro (kusanganisira kushandisa mifananidzo) yeiyo chaiyo AI/ML budiriro isati yaiswa ("yakaiswa") muInterSystems IRIS. Zviri pachena kuti kuvandudzwa kutsva kwakagadzirwa nenzira iyi kuchawana chete kugadzirisa kwekutanga (sezvo, kunyanya, Jupyter isingashande neyechokwadi-nguva data streams) - izvi zviri muhurongwa hwezvinhu, nokuti mhedzisiro huru yebudiriro muJupyter. chisimbiso chekushanda kwakakosha kweimwe nzira yeAI / ML ("inoratidza mhedzisiro inotarisirwa pane data data"). Saizvozvo, imwe nzira yakatoiswa pachikuva (ona anotevera macro-mechanisms) isati yagadziriswa muJupyter ingangoda "kudzoreredza" kune "pre-platform" fomu (kuverenga data kubva kumafaira, kushanda nedata kuburikidza nexDBC pachinzvimbo chetafura, kudyidzana kwakananga nepasirese - multidimensional data arrays InterSystems IRIS - nezvimwewo).

Chinhu chakakosha chekushandiswa kweCD muInterSystems IRIS: kubatanidzwa kwebidirectional kwave kuchiitwa pakati pepuratifomu neJupyter, zvichibvumira zviri muPython, R naJulia kuti zviendeswe kuchikuva (uye, zvichizoitwa pachikuva) (ese ari matatu ari programming. mitauro mumitauro inoenderana inotungamira yakavhurika-sosi). sosi yemasvomhu modhi yenzvimbo). Nekudaro, vanogadzira zvemukati zveAI/ML vane mukana wekuita "kuenderera mberi kwekuendesa" kwezvirimo papuratifomu, vachishanda mune yavo yavanoziva Jupyter mupepeti, ane maraibhurari anozivikanwa anowanikwa muPython, R, Julia, uye kuita zvekutanga debugging (kana zvichidikanwa) kunze kwechikuva .

Ngatienderei kuCI macro mechanism muInterSystems IRIS. Iyo dhizaini inoratidza macro process ye "chaiyo-nguva robotizer" (yakaomarara yedata zvimiro, bhizinesi maitiro uye zvimedu zvekodhi zvakarongwa navo mumitauro yemasvomhu uye ObjectScript - mutauro wekuvandudza weInterSystems IRIS). Basa reiyi macro process ndeyekuchengetedza mitsara yedata inodiwa pakushanda kweAI/ML masisitimu (zvichienderana nekuyerera kwedata kunopfuudzwa papuratifomu munguva chaiyo), kuita sarudzo nezve kutevedzana kwekushandisa uye "assortment" yeAI/ ML nzira (idzo zvakare "mathematical algorithms", "modheru", nezvimwewo - inogona kudaidzwa zvakasiyana zvichienderana neiyo chaiyo yekuitwa uye terminological zvido), chengetedza zvimiro zvedata kusvika zvino kuti uongorore mhedzisiro yebasa reAI/ ML michina (cubes, matafura, multidimensional data arrays, etc.) etc. - zvemishumo, dashboards, nezvimwewo).

Chinhu chakakosha chekuitwa kweCI muInterSystems IRIS: kubatanidzwa kwebidirectional kwaitwa pakati pepuratifomu nenzvimbo dzemasvomhu dzekuenzanisira, zvichikubvumidza kuti uite zvirimo zvakagarwa papuratifomu muPython, R naJulia munzvimbo dzavo dzakasiyana uye kugamuchira mhedzisiro yekuurayiwa. Kubatanidzwa uku kunoitwa zvese mu "terminal mode" (kureva, AI/ML yemukati inogadzirwa seObjectScript kodhi inofona kune zvakatipoteredza) uye mu "business process mode" (kureva, AI/ML yemukati inogadzirwa senge bhizinesi maitiro uchishandisa graphical editor, kana dzimwe nguva kushandisa Jupyter, kana kushandisa IDE - IRIS Studio, Eclipse, Visual Studio Code). Kuvepo kwemaitiro ebhizinesi ekugadzirisa muJupyter kunoratidzwa kuburikidza nekubatana pakati peIRIS padanho reCI uye Jupyter padanho reCD. Imwe ongororo yakadzama yekubatanidzwa nenzvimbo dzemasvomhu yekuenzanisira inopiwa pazasi. Panguva ino, semaonero edu, pane zvikonzero zvese zvekuona kuti puratifomu ine zvese zvinodiwa zvekushandisa "kuenderera mberi kubatanidzwa" kweAI/ML budiriro (inobva ku "kuenderera mberi kuendeswa") mune chaiyo-nguva AI/ML mhinduro.

Uye iyo huru macro mechanism: CT. Pasina iyo, hapazove neAI/ML chikuva (kunyangwe "nguva chaiyo" ichaitwa kuburikidza neCD/CI). Izvo zvakakosha zveCT ibasa repuratifomu ne "artifacts" yekudzidza kwemuchina uye hungwaru hwekugadzira zvakananga mumasesheni ekushanda emamiriro ekuenzanisa masvomhu: modhi, matafura ekugovera, matrix vectors, zvidimbu zveneural network, nezvimwe. Iri "basa", kazhinji, rinosanganisira kugadzira izvo zvakataurwa munzvimbo dzakasimukira (mune mamiriro emhando, semuenzaniso, "kusikwa" kunosanganisira kugadzika kweiyo modhi uye kunotevera kusarudzwa kwehunhu hwema paramita ayo - iyo inonzi "kudzidziswa" yemuenzaniso), kushandiswa kwavo (kune mamodheru: kuverenga nerubatsiro rwavo rwe "modhi" kukosha kwezvakatarwa zvinosiyana - kufanotaura, nhengo yeboka, mukana wechiitiko, nezvimwewo) uye kuvandudzwa kwekare. yakagadzirwa uye yakashandiswa zvigadzirwa (somuenzaniso, kutsanangudza seti yemodhi yekuisa shanduko zvichienderana nemhedzisiro yekushandisa - kuitira kuvandudza kufanotaura kwechokwadi, sechisarudzo). Chinhu chakakosha pakunzwisisa basa reCT ndeye "kubvisa" kubva kune chokwadi cheCD neCI: CT ichaita zvese zvigadzirwa, ichitarisa pane computational uye masvomhu chaiwo eiyo AI / ML mhinduro mukati mekugona kunopihwa neakananga nharaunda. Basa re "kupa zvekushandisa" uye "kuendesa zvinobuda" richava basa reCD neCI.

Chinhu chakakosha pakuitwa kweCT chakanyanya muInterSystems IRIS: kushandisa kubatanidzwa nenzvimbo dzemasvomhu dzekuenzanisira dzatotaurwa pamusoro, puratifomu inokwanisa kuburitsa izvo chaizvo zvigadzirwa kubva muzvikamu zvebasa zvinomhanya pasi pekutonga kwayo munzvimbo dzemasvomhu uye (kunyanya kukosha) kutendeuka. ivo mupuratifomu data zvinhu. Semuenzaniso, tafura yekugovera iyo ichangobva kugadzirwa mumusangano wePython unoshanda inogona kuva (pasina kumisa musangano wePython) inotamirwa kupuratifomu nenzira ye, semuenzaniso, yepasirese (a multidimensional InterSystems IRIS data array) - uye yakashandiswa. yekuverenga mune imwe AI/ML- mashandiro (yakashandiswa mumutauro weimwe nharaunda - semuenzaniso, muR) - kana tafura chaiyo. Mumwe muenzaniso: mukuwirirana ne "yakajairika maitiro" ekushanda kwemuenzaniso (mumusangano wePython wekushanda), "auto-ML" inoitwa pane yayo yekuisa data: otomatiki kusarudzwa kwezvakanyanya kupinza zvinosiyana uye parameter kukosha. Uye pamwe chete ne "nguva dzose" kudzidziswa, modhi inobudirira munguva chaiyo inogamuchirawo "chikumbiro chekugadzirisa" chirevo chayo - umo seti yezvipo zvinoshanduka zvinoshanduka, iyo parameter inoshanduka (haisisiri semhedzisiro yekudzidziswa. muPython, asi semugumisiro wekudzidziswa ne "imwe nzira" vhezheni yayo pachayo, senge H2O stack), ichibvumira iyo yakazara AI/ML mhinduro kutarisana nekuzvimirira nekuchinja kusingatarisirwe muchimiro cheiyo data yekupinza uye zviitiko zviri kuenzanisirwa. .

Ngatizivei zvakadzama nepuratifomu AI/ML mashandiro eInterSystems IRIS, tichishandisa muenzaniso weiyo chaiyo-yehupenyu prototype.

Mumufananidzo uri pasi apa, kune rumwe rutivi rworuboshwe rwe slide pane chikamu chekushanda kwebhizimisi rinoshandisa kushandiswa kwezvinyorwa muPython uye R. Muchikamu chepakati pane zvinyorwa zvekuona zvekuitwa kwezvimwe zvezvinyorwa izvi, maererano. muPython uye R. Pakarepo shure kwavo kune mienzaniso yezvinyorwa pane imwe uye imwe mutauro, inotamirwa kuurayiwa kunzvimbo dzakakodzera. Kumagumo kurudyi kune zviratidziro zvinoenderana nemhedzisiro yescript execution. Maonesheni ari kumusoro akaitwa paIRIS Analytics (data yakatorwa kubva kuPython ichipinda muInterSystems IRIS data platform uye yakaratidzwa pane dashboard uchishandisa chikuva), pazasi yakaitwa yakananga muR work session uye yakabuda kubva ipapo kusvika kugraphic files. . Chinhu chakakosha: chidimbu chakaunzwa mune prototype chine basa rekudzidzisa modhi (chikamu chemidziyo yenyika) pane data rakagamuchirwa munguva chaiyo kubva kune yemidziyo simulator maitiro, pakuraira kubva kumhando yemhando yekutarisa maitiro akaonekwa panguva yekushandiswa kweiyo modhi. Kuitwa kweAI/ML mhinduro muchimiro cheseti yekudyidzana maitiro ("magents") ichakurukurwa zvakare.

Mufananidzo 12 Kudyidzana nePython, R naJulia muInterSystems IRIS

Matanho epuratifomu (iwo zvakare "maitiro ebhizinesi", "analytical process", "mapombi", nezvimwewo - zvichienderana nemamiriro ezvinhu), chekutanga, anogadziridzwa mune graphical bhizinesi process mupepeti papuratifomu pachayo, uye mune yakadaro nzira iyo ese ari maviri ayo dhizaini dhizaini uye inoenderana AI/ML michina (chirongwa kodhi) inogadzirwa panguva imwe chete. Kana tichiti "muchina weAI/ML wawanikwa," tinotanga tichireva hybridity (mukati mekuita kamwe chete): zvirimo mumitauro yenzvimbo dzemasvomhu dzekuenzanisa zviri padyo nezviri muSQL (kusanganisira mawedzero kubva IntegratedML), muInterSystems ObjectScript, nemimwe mitauro inotsigirwa. Uyezve, chirongwa chepuratifomu chinopa mikana yakafara kwazvo ye "kupa" muchimiro chezvimedu zvakasimudzwa (sezvinogona kuonekwa mumuenzaniso mudhayagiramu pazasi), izvo zvinokutendera kuti unyatso kuronga kunyangwe zvakaomesesa zvemukati pasina "kudonha" yemhando yegraphical (mu "asiri-graphical") mafomati). »mitoo/makirasi/maitirwo, nezvimwewo). Ndokunge, kana zvichidikanwa (uye zvinofanooneswa mumapurojekiti mazhinji), zvachose iyo AI/ML mhinduro inogona kuitwa mune graphical self-documenting fomati. Ndapota cherechedza kuti muchikamu chepakati chemufananidzo uri pasi apa, iyo inomiririra yakakwirira "yero ye nesting", zviri pachena kuti kunze kwebasa chairo rekudzidzisa muenzaniso (uchishandisa Python uye R), kuongororwa kwezvinonzi. ROC curve yemuenzaniso wakadzidziswa inowedzerwa, ichibvumira zvinoonekwa (uye computationally futi) kuongorora kunaka kwekudzidziswa - uye kuongororwa uku kunoitwa mumutauro weJulia (akaurayiwa, zvinoenderana, munzvimbo yeJulia yemasvomhu).

Mufananidzo 13 Inotaridzika nharaunda yekuumbwa kweAI/ML mhinduro muInterSystems IRIS

Sezvambotaurwa, kusimudzira kwekutanga uye (mune dzimwe nguva) kugadziridzwa kweAI/ML nzira dzakatoitwa pachikuva kuchaita/kunogona kuitwa kunze kwepuratifomu muJupyter mupepeti. Mumufananidzo uri pazasi tinoona muenzaniso wekugadzirisa iripo papuratifomu process (yakafanana neiyo mudhayagiramu iri pamusoro) - iyi ndiyo nzira iyo chidimbu chine basa rekudzidzisa modhi chinoratidzika senge muJupyter. Python zvemukati zviripo pakugadzirisa, kugadzirisa, uye magiraidhi anobuda zvakananga muJupyter. Shanduko (kana zvichidikanwa) inogona kuitwa nekuyananisa ipapo muchirongwa chepuratifomu, kusanganisira vhezheni yayo inogadzira. Zvitsva zvemukati zvinogona kuendeswa kupuratifomu nenzira yakafanana (itsva yepuratifomu process inogadzirwa otomatiki).

Mufananidzo 14 Kushandisa Jupyter Notebook kugadzirisa iyo AI/ML injini muInterSystems IRIS chikuva.

Kugadziriswa kwechikuva chepuratifomu kunogona kuitwa kwete chete mune graphical kana laptop fomati - asiwo mu "total" IDE (Integrated Development Environment) fomati. Aya maIDE iRIS Studio (yekuzvarwa IRIS studio), Visual Studio Code (InterSystems IRIS yekuwedzera yeVSCode) uye Eclipse (Atelier plugin). Mune zvimwe zviitiko, zvinokwanisika kuti timu yekusimudzira ishandise ese matatu maIDE panguva imwe chete. Dhiagiramu iri pazasi inoratidza muenzaniso wekugadzirisa maitiro akafanana muIRIS studio, muVisual Studio Code uye muEclipse. Chokwadi zvese zvirimo zviripo pakugadzirisa: Python/R/Julia/SQL, ObjectScript, uye bhizinesi maitiro.

Mufananidzo 15 Kuvandudzwa kweInterSystems IRIS bhizinesi maitiro mumaIDE akasiyana-siyana

Zvishandiso zvekutsanangura nekuita InterSystems IRIS bhizinesi maitiro muBusiness Process Mutauro (BPL) inofanirwa kutaurwa zvakanyanya. BPL inoita kuti zvikwanise kushandisa "zvakagadzirirwa-zvakagadzirwa zvekubatanidza zvikamu" (zviitwa) mumabhizinesi maitiro - izvo, kutaura zvazviri, zvinopa zvikonzero zvese zvekutaura kuti "kuenderera kubatanidzwa" kunoitwa muInterSystems IRIS. Yakagadzirirwa-yakagadzirwa bhizinesi maitiro ezvikamu (zviitwa uye zvinongedzo pakati pazvo) isimba rekumhanyisa kuunganidza mhinduro yeAI/ML. Uye kwete magungano chete: nekuda kwezviitwa uye kubatana pakati pawo pamusoro pekusiyana kweAI/ML kuvandudza uye masisitimu, "inozvimiririra manejimendi layer" inomuka, inokwanisa kuita sarudzo zvinoenderana nemamiriro ezvinhu, munguva chaiyo.

Mufananidzo 16 Yakagadzirira-yakagadzirwa bhizinesi maitiro ezvikamu zvekuenderera mberi kubatanidzwa (CI) paInterSystems IRIS chikuva.

Pfungwa yeagent system (inozivikanwawo se "multi-agent systems") ine nzvimbo yakasimba mu robotization, uye InterSystems IRIS platform organically inotsigira kuburikidza ne "product-process" kuvaka. Pamusoro peiyo isingaverengeki mikana ye "kurongedza" yega yega maitiro nekushanda kunodiwa kune iyo yakazara mhinduro, kupa iyo sisitimu yemapuratifomu maitiro nemidziyo ye "agency" inobvumidza iwe kugadzira mhinduro dzinoshanda dzekusagadzikana kwakanyanya simulated zviitiko (maitiro emagariro / biosystems, zvishoma zvinoonekwa tekinoroji maitiro, nezvimwewo).

Mufananidzo 16 Kushanda kweAI/ML mhinduro seagent-based business process system muInterSystems IRIS.

Isu tinoenderera mberi nekuongorora kwedu kweInterSystems IRIS nenyaya yekushandiswa kwepuratifomu kugadzirisa makirasi ese ematambudziko enguva-chaiyo (sumo yakadzama kune mamwe emhando yepamusoro yepuratifomu AI/ML paInterSystems IRIS inogona kuwanikwa mune imwe. yedu yapfuura webinars).

Kupisa pazvitsitsinho zvemufananidzo wekare, pasi apa pane dhiyabhorosi yakanyatsojeka yeagent system. Iyo dhizaini inoratidza yakafanana prototype, ese mana ejenti maitiro anooneka, hukama huri pakati pavo hwakadhirowewa schematically: GENERATOR - inogadzirisa kugadzirwa kwedata nema sensors emidziyo, BUFFER - inobata mitsara yedata, ANALYZER - inoita muchina kudzidza pachayo, TARIRA - inotarisisa iyo mhando yekudzidza kwemuchina uye inotumira chiratidzo nezve kudiwa kwekudzokorora modhi.

Mufananidzo 17 Kuumbwa kweAI/ML mhinduro muchimiro cheagent-based business process system muInterSystems IRIS.

Dhiagiramu iri pazasi inoratidza kuzvimiririra kushanda kweimwe marobhoti prototype (kuzivikanwa kwemanzwiro emavara ezvinyorwa) kwenguva yakati. Muchikamu chepamusoro ndiko kushanduka kwechiratidzo chemhando yekudzidzira kwemuenzaniso (unhu huri kukura), muchikamu chezasi pane simba rechiratidzo chemhando yekushandiswa kwemuenzaniso uye chokwadi chekudzokorora kudzidziswa (maronda matsvuku). Sezvauri kuona, mhinduro yakadzidza pachayo zvine mutsindo uye yakazvimiririra, uye inoshanda pane yakapihwa mhando yemhando (mhando yemhando yemhando haiwire pasi pe80%).

Mufananidzo 18 Kuenderera mberi (kuzvidzidzisa) (CT) paInterSystems IRIS papuratifomu

Isu takambotaura nezve "auto-ML" pakutanga, asi dhayagiramu pazasi inoratidza kushandiswa kweichi chiitiko zvakadzama uchishandisa muenzaniso weimwe prototype. Iyo graphical dhizaini yechidimbu chebhizinesi maitiro inoratidza chiitiko chinokonzeresa kuenzanisira mune H2O stack, inoratidza mhedzisiro yeiyi modhi (iyo yakajeka kutonga kweiyo inoguma modhi pane "yakagadzirwa nevanhu" modhi, zvinoenderana neiyo dhayagiramu yekuenzanisa ROC curves, pamwe chete nekuzivikanwa kwega kwe "anonyanya kupesvedzera akasiyana" anowanikwa mune yekutanga data set). Chinhu chakakosha pano kuchengetedza nguva uye zviwanikwa zvehunyanzvi zvinowanikwa kuburikidza ne "auto-ML": izvo zvatinoita papuratifomu yedu muhafu yeminiti (kutsvaga uye kudzidzisa iyo yakakwana modhi) inogona kutora nyanzvi kubva pavhiki kuenda kumwedzi.

Mufananidzo 19 Kubatanidzwa kwe "auto-ML" muAI / ML mhinduro pane InterSystems IRIS chikuva.

Dhiagiramu iri pasi apa inopotsa pfungwa yacho zvishoma, asi inzira yakanaka yekugumisa nyaya pamusoro pemakirasi ematambudziko chaiwo-nguva ari kugadziriswa: tinokuyeuchidza kuti nezvose zvinokwanisa zveInterSystems IRIS platform, maitiro ekudzidzira pasi pekutonga kwayo. kwete zvinosungirwa. Iyi puratifomu inogona kugamuchira kubva kunze iyo inodaidzwa kuti PMML kutsanangurwa kweiyo modhi, yakadzidziswa mune chishandiso chisiri pasi pekutonga kwepuratifomu - uye shandisa iyi modhi munguva chaiyo kubva panguva yainotengeswa kunze kwenyika. PMML tsanangudzo. Izvo zvakakosha kuti titarise kuti hazvisi zvese zveAI / ML zvigadzirwa zvinogona kuderedzwa kusvika kune PMML kutsanangurwa, kunyangwe kana mazhinji ezvakajairwa zvigadzirwa achibvumira izvi. Saka, iyo InterSystems IRIS chikuva "yakavhurika loop" uye haireve "platform uranda" kune vashandisi.

Mufananidzo 20 Kubatanidzwa kwe "auto-ML" muAI / ML mhinduro pane InterSystems IRIS chikuva.

Ngatinyorei mamwe mabhenefiti epuratifomu eInterSystems IRIS (yekujekesa, zvine chekuita nekutonga kwemaitiro), ayo akakosha zvikuru mu automation yehungwaru hwekugadzira uye chaiyo-nguva yekudzidza muchina:

• Yakagadziridzwa maturusi ekubatanidza ane chero masosi data uye vatengi (process control system/SCADA, midziyo, MRO, ERP, nezvimwewo)
• Yakavakwa-mukati akawanda-modhi DBMS yepamusoro-inoshanda transaction uye analytical process (Hybrid Transaction/Analytical Processing, HTAP) yechero vhoriyamu yetekinoroji maitiro data.
• Zvishandiso zvekuvandudza zvekuenderera mberi kwekuendesa kweAI/ML injini dzemhinduro dzenguva-chaiyo yakavakirwa paPython, R, Julia.
• Adaptive bhizinesi maitiro ekuenderera mberi kwekubatanidza uye (kuzvimiririra) kudzidza kweiyo chaiyo-nguva AI/ML mhinduro injini.
• Yakavakwa-mukati Bhizinesi Intelligence maturusi ekuona maitiro data uye mhedzisiro yeAI/ML mhinduro
• API Management yekuunza mhedzisiro yeAI/ML mhinduro yekugadzirisa kudzora masisitimu / SCADA, ruzivo uye analytical masisitimu, kutumira chenjedzo, nezvimwe.

AI/ML mhinduro paInterSystems IRIS papuratifomu inokwana nyore muiyo iripo IT zvivakwa. Iyo InterSystems IRIS papuratifomu inovimbisa kuvimbika kwakanyanya kweAI/ML mhinduro nekutsigira kushivirira-kushivirira uye kushivirira-kugadzika zvigadziriso uye kuchinjika kutumirwa munzvimbo dzakaita, pamaseva emuviri, mumakore akavanzika uye eruzhinji, uye Docker midziyo.

Saka, InterSystems IRIS inzvimbo yepasirese chaiyo-nguva AI/ML computing chikuva. Iyo yepasirese yepuratifomu yedu inosimbiswa mukuita nekushaikwa kwezvirambidzo zve de facto pakuoma kwekuverenga kwakashandiswa, kugona kweInterSystems IRIS kusanganisa (munguva chaiyo) kugadzirisa mamiriro kubva kwakasiyana maindasitiri, uye kuchinjika kwakasarudzika kwe. chero mabasa epuratifomu uye nzira kune izvo zvinodiwa nevashandisi.

Mufananidzo 21 InterSystems IRIS - yepasirese chaiyo-nguva AI/ML komputa chikuva

Kuti uwane kudyidzana kwakanyanya nevaverengi vedu vanofarira izvo zviri kuratidzwa pano, isu tinokurudzira kuti usazviverengera pakuverenga nekuenderera mberi nenhaurirano "live." Tichafara kupa rutsigiro nekugadzirwa kweiyo chaiyo-nguva AI / ML mamiriro ane chekuita neiyo kambani yako, ita pamwe chete prototyping paInterSystems IRIS papuratifomu, gadzira uye shandisa mukudzidzira nzira yekuunza hungwaru hwekugadzira uye kudzidza muchina. mune yako kugadzira uye manejimendi maitiro. Yedu AI/ML Nyanzvi Chikwata Bata Email - [email inodzivirirwa].

Source: www.habr.com