Tchati cha Gartner chili ngati chiwonetsero cha mafashoni apamwamba kwa omwe ali mumakampani aukadaulo. Poyang'ana pa izi, mutha kudziwa pasadakhale kuti ndi mawu ati omwe ali otchuka kwambiri nyengo ino komanso zomwe mungamve pamisonkhano yonse yomwe ikubwera.
Tazindikira zomwe zili kumbuyo kwa mawu okongola omwe ali pagrafuyi kuti nanunso mutha kulankhula chinenerocho.
Poyamba, mawu ochepa chabe onena za mtundu wa graph iyi. Chaka chilichonse mu Ogasiti, bungwe lothandizira la Gartner limatulutsa lipoti - Gartner Hype Curve. Mu Russian, iyi ndi "hype curve," kapena, mophweka, hype. Zaka 30 zapitazo, oimba oimba a m’gulu la Public Enemy anaimba kuti: “Musakhulupirire mawu osangalatsawo.” Khulupirirani kapena ayi, ndi funso laumwini, koma ndikofunikira kudziwa mawu osakira ngati mumagwira ntchito muukadaulo ndipo mukufuna kudziwa zomwe zikuchitika padziko lonse lapansi.
Ichi ndi chithunzi cha zomwe anthu amayembekezera kuchokera kuukadaulo wina. Malinga ndi Gartner, moyenera, ukadaulo umadutsa magawo 5: kukhazikitsidwa kwaukadaulo, kukwera kwa ziyembekezo zochulukira, chigwa cha zokhumudwitsa, malo otsetsereka a chidziwitso, malo opangira zokolola. Koma zimachitikanso kuti imamira mu "chigwa cha zokhumudwitsa" - mungathe kukumbukira zitsanzo nokha mosavuta, tengani ma bitcoins omwewo: poyambira kugunda pachimake monga "ndalama zamtsogolo", adatsika mofulumira pamene zofooka zaukadaulo. zinaonekeratu, choyamba zoletsa pa chiwerengero cha wotuluka ndi kuchuluka kwa magetsi zofunika kupanga bitcoins (zomwe zikuphatikizapo mavuto zachilengedwe). Ndipo, ndithudi, sitiyenera kuiwala kuti tchati cha Gartner ndi kulosera chabe: apa, mwachitsanzo, mukhoza kuwerenga mwatsatanetsatane. , kumene maulosi ochititsa chidwi kwambiri amene sanakwaniritsidwe amasanjidwa.
Chifukwa chake, tiyeni tidutse tchati chatsopano cha Gartner. Matekinoloje amagawidwa m'magulu 5 akulu akulu:
- Advanced AI ndi Analytics
- Postclassical Compute ndi Comms
- Sensing ndi Mobility
- Augmented Munthu
- Digital Ecosystems
1. Advanced AI ndi Analytics
Pazaka 10 zapitazi taona ola labwino kwambiri la kuphunzira mozama. Maukonde awa ndi othandizadi pantchito zawo zosiyanasiyana. Mu 2018, Yann LeCun, Geoffrey Hinton ndi Yoshua Bengio adalandira Mphotho ya Turing pazomwe adapeza - mphotho yapamwamba kwambiri, yofanana ndi Mphotho ya Nobel mu sayansi yamakompyuta. Choncho, zochitika zazikulu m'derali, zomwe zikuwonetsedwa pa tchati:
1.1. Kusamutsa Maphunziro
Simuphunzitsa neural network kuyambira pachiyambi, koma tengani yemwe waphunzitsidwa kale ndikumupatsa cholinga china. Nthawi zina izi zimafuna kubwerezanso gawo la maukonde, koma osati maukonde onse, omwe ali mwachangu kwambiri. Mwachitsanzo, kutenga neural network yopangidwa kale ResNet50, yophunzitsidwa pa dataset ya ImageNet1000, mupeza algorithm yomwe imatha kuyika zinthu zambiri pazithunzi pamlingo wakuya kwambiri (makalasi 1000 kutengera mawonekedwe opangidwa ndi zigawo 50 za neural). network). Koma simukuyenera kuphunzitsa maukonde onsewo, zomwe zingatenge miyezi.
В Samsung "Neural network ndi masomphenya apakompyuta", mwachitsanzo, pomaliza ndi gulu la mbale zoyera ndi zakuda, njira ikuwonetsedwa kuti mumphindi 5 imakupatsani mwayi wokhala ndi neural network yakuya yomwe imatha kusiyanitsa mbale zakuda ndi zoyera, zomangidwa molingana ndi kamangidwe kamene tafotokoza pamwambapa. Maukonde oyambirira sankadziwa kuti mbale zinali chiyani, adangophunzira kusiyanitsa mbalame ndi agalu (onani ImageNet).
Source: Samsung "Neural network ndi masomphenya apakompyuta"
Kuti muphunzire za Transfer muyenera kudziwa kuti ndi njira ziti zomwe zimagwira ntchito komanso zomwe zidapangidwa kale zomwe zilipo. Ponseponse, izi zimafulumizitsa kwambiri kutuluka kwa ntchito zothandiza pakuphunzirira makina.
1.2. Generative Adversarial Networks (GAN)
Izi ndizochitika pamene zimakhala zovuta kuti tipange cholinga chophunzirira. Kuyandikira kwa ntchitoyo ndi moyo weniweni, kumamveka bwino kwa ife ("bweretsani tebulo la pambali pa bedi"), koma zimakhala zovuta kwambiri kuzipanga ngati ntchito yaukadaulo. GAN ndi kuyesa chabe kutipulumutsa ku vutoli.
Pali maukonde awiri omwe akugwira ntchito pano: imodzi ndi jenereta (Generative), inayo ndi yatsankho (Adversarial). Maukonde amodzi amaphunzira kugwira ntchito zothandiza (kugawa zithunzi, kuzindikira mawu, kujambula zojambulajambula). Ndipo netiweki ina imaphunzira kuphunzitsa maukonde: ili ndi zitsanzo zenizeni, ndipo imaphunzira kupeza chilinganizo chovuta chomwe sichinadziwikepo chofanizira zinthu za gawo lopangira maukonde ndi zinthu zenizeni zapadziko lapansi (zophunzitsira) zochokera pamikhalidwe yozama kwambiri. : chiwerengero cha maso, pafupi ndi kalembedwe ka Miyazaki, matchulidwe olondola a Chingerezi.
Chitsanzo cha zotsatira za netiweki yopangira zilembo za anime.
Koma, ndithudi, ndizovuta kumanga zomangamanga kumeneko. Sikokwanira kungoponya ma neurons, amafunika kukonzekera. Ndipo muyenera kuphunzira kwa milungu ingapo. Anzanga ku Samsung Artificial Intelligence Center akugwira ntchito pa mutu wa GAN; ili ndi limodzi mwamafunso awo ofunikira. Mwachitsanzo, monga chonchi : kugwiritsa ntchito maukonde opanga kupanga zithunzi zenizeni za anthu omwe ali ndi mawonekedwe osinthika - mwachitsanzo, kupanga chipinda choyenera, kapena kupanga nkhope, zomwe zingachepetse kuchuluka kwa zidziwitso zomwe ziyenera kusungidwa kapena kufalitsidwa kuti zitsimikizire kanema wapamwamba kwambiri. kuyankhulana, kuwulutsa kapena kuteteza deta yanu.
1.3. Kufotokozera kwa AI
Pazochita zina zomwe zimachitika kawirikawiri, kupita patsogolo kwa zomangamanga mozama kwabweretsa mwadzidzidzi kuthekera kwa ma neural network pafupi ndi kuthekera kwamunthu. Tsopano nkhondo ili kuonjezera kuchuluka kwa ntchito zotere. Mwachitsanzo, makina otsuka ma loboti amatha kusiyanitsa mosavuta mphaka ndi galu pamisonkhano yapamsonkhano. Koma m'mikhalidwe yambiri ya moyo, sangathe kupeza mphaka akugona pakati pa nsalu kapena mipando (komabe, monga ife, nthawi zambiri ...).
Chifukwa chiyani chakuchita bwino kwa ma neural network akuya? Amapanga chithunzithunzi cha vutolo osati pazidziwitso "zowoneka ndi maso" (ma pixel azithunzi, kusintha kwa voliyumu ya mawu ...), koma pazinthu zomwe zidapezedwa pambuyo pokonza chidziwitsochi ndi magawo mazana angapo a neural network. Tsoka ilo, maubwenziwa atha kukhala opanda tanthauzo, osagwirizana, kapena amakhala ndi zolakwika mu seti yoyambirira. Mwachitsanzo, pali masewera ang'onoang'ono apakompyuta okhudza zomwe kugwiritsa ntchito mopanda nzeru kwa AI polemba anthu kungayambitse .
Makina olembera zithunzi amatcha munthu amene akuphika kuti ndi mkazi, ngakhale kuti amene ali pachithunzipa ndi mwamuna (). izo ku Virginia Institute.
Kuti tiwunike maubale ovuta komanso ozama omwe nthawi zambiri sitingathe kupanga tokha, Njira zofotokozera za AI ndizofunikira. Amapanga mawonekedwe a neural network yakuya kotero kuti pambuyo pophunzitsidwa, titha kusanthula zoyimira zamkati zomwe maukonde aphunzira, m'malo mongodalira chisankho chake.
1.4. Edge Analytics / AI
Chilichonse chomwe chili ndi mawu akuti Edge kwenikweni chimatanthauza izi: kusamutsa gawo la ma aligorivimu kuchokera pamtambo / seva kupita ku chipangizo chomaliza / pachipata. Algorithm yotereyi idzagwira ntchito mwachangu ndipo sidzafunikira kulumikizana ndi seva yapakati kuti igwire ntchito. Ngati mumadziwa za kuchotsedwa kwa "kasitomala woonda," ndiye apa tikupangitsa kasitomala uyu kukhala wokhuthala pang'ono.
Izi zitha kukhala zofunikira pa intaneti ya Zinthu. Mwachitsanzo, ngati makina atenthedwa ndipo akufunika kuziziritsa, ndizomveka kuwonetsa izi nthawi yomweyo, pamtunda wa zomera, popanda kuyembekezera kuti deta ipite kumtambo ndi kuchoka kumeneko kupita kwa woyang'anira wosuntha. Kapena chitsanzo china: magalimoto odziyendetsa okha amatha kudziwa momwe magalimoto alili pawokha, osalumikizana ndi seva yapakati.
Kapena chitsanzo china cha chifukwa chake izi ndizofunikira pachitetezo chachitetezo: mukalemba zolemba pafoni yanu, zimakumbukira mawu omwe amafanana ndi inu, kotero kuti pambuyo pake kiyibodi ya foni imatha kukupangitsani kukhala nawo - izi zimatchedwa kuneneratu. kulemba mawu. Kutumiza chilichonse chomwe mumalemba pa kiyibodi yanu kumalo osungira data kwinakwake kungakhale kuphwanya zinsinsi zanu komanso kukhala osatetezeka. Chifukwa chake, maphunziro a kiyibodi amapezeka mkati mwa chipangizo chanu chokha.
1.5. AI Platform monga Service (AI PaaS)
PaaS - Platform-as-a-Service ndi chitsanzo cha bizinesi chomwe timapeza mwayi wopita ku nsanja yophatikizika, kuphatikizapo kusungirako deta yochokera kumtambo ndi njira zokonzekera. Mwanjira iyi, titha kudzipulumutsa tokha ku ntchito zomanga ndikuyang'ana kwambiri kupanga china chake chothandiza. Chitsanzo cha nsanja za PaaS za ntchito za AI: IBM Cloud, Microsoft Azure, Amazon Machine Learning, Google AI Platform.
1.6. Kuphunzira kwa Makina Osinthira (Adaptive ML)
Nanga bwanji ngati tilola luntha lochita kupanga kuti lisinthe… Mumafunsa - ndiye kuti, bwanji? Vuto ndi ili: timapanga movutikira vuto lililonse lotere tisanapange algorithm yanzeru kuti tithetse. Adzakuyankhani - zikuwoneka kuti unyolo uwu ukhoza kuphweka.
Kuphunzira makina ochiritsira kumagwira ntchito pa mfundo yotseguka: mumakonzekera deta, bwerani ndi neural network (kapena chirichonse), phunzitsani, kenako yang'anani zizindikiro zingapo, ndipo ngati mukufuna chirichonse, mukhoza kutumiza neural network ku mafoni a m'manja. - kuthetsa mavuto ogwiritsa ntchito. Koma m'mapulogalamu omwe pali deta yambiri ndipo chikhalidwe chake chimasintha pang'onopang'ono, njira zina ndizofunikira. Machitidwe otere, omwe amasintha ndikudziphunzitsa okha, amapangidwa kukhala zotsekeka, zodzipangira tokha (zotsekeka), ndipo ziyenera kugwira ntchito bwino.
Mapulogalamu - izi zitha kukhala kusanthula kwamayendedwe (Stream Analytics), pamaziko omwe mabizinesi ambiri amapanga zisankho, kapena kasamalidwe kosintha kapangidwe. Pamlingo wazomwe zimagwiritsidwa ntchito pano ndikupatsidwa zoopsa zomwe anthu amazimvetsetsa bwino, njira zomwe zimapanga njira yothetsera vutoli zonse zimasonkhanitsidwa pansi pa mawu akuti Adaptive AI.
Kuyang'ana chithunzichi, n'zovuta kuchotsa kumverera komwe sikudyetsa mkate wa futurologists - aloleni aphunzitse robot kupuma ...
Postclassical Compute ndi Comms
2.1. Kulumikizana kwa mafoni a m'badwo wachisanu (5G)
Uwu ndi mutu wosangalatsa kotero kuti timakutumizirani nthawi yomweyo ku zathu . Chabwino, nachi chidule chachidule. 5G, pakuwonjezera kuchuluka kwa kufalitsa kwa data, ipangitsa kuti intaneti ifulumire mopanda nzeru. Zimakhala zovuta kuti mafunde afupiafupi adutse zopinga, kotero mapangidwe a maukonde adzakhala osiyana kotheratu: 500 nthawi XNUMX masiteshoni owonjezera pakufunika.
Pamodzi ndi liwiro, tidzapeza zochitika zatsopano: masewera a nthawi yeniyeni ndi zenizeni zowonjezera, kuchita ntchito zovuta (monga opaleshoni) kudzera pa telepresence, kuteteza ngozi ndi zovuta m'misewu kudzera mukulankhulana pakati pa makina. Pazambiri: intaneti yam'manja imasiya kutsika pamisonkhano yayikulu, monga machesi pabwalo lamasewera.
Chithunzi chojambula - Reuters, Niantic
2.2. Memory ya m'badwo Wotsatira
Apa tikukamba za m'badwo wachisanu wa RAM - DDR5. Samsung idalengeza kuti zopangira DDR2019 zizipezeka kumapeto kwa 5. Zikuyembekezeka kuti kukumbukira kwatsopano kudzakhala kuwirikiza kawiri komanso kuwirikiza kawiri ndikusunga mawonekedwe omwewo, ndiye kuti, titha kupeza zomata zokhala ndi mphamvu mpaka 32GB pakompyuta yathu. M'tsogolomu, izi zidzakhala zofunikira makamaka kwa mafoni a m'manja (chikumbukiro chatsopano chidzakhala chochepa mphamvu) ndi ma laputopu (kumene chiwerengero cha DIMM slots ndi chochepa). Ndipo kuphunzira pamakina kumafunanso kuchuluka kwa RAM.
2.3. Malo Otsika-Earth-Orbit Satellite Systems
Lingaliro lakusintha ma satelayiti olemera, okwera mtengo, amphamvu okhala ndi tinthu tating'ono komanso otsika mtengo ndilatsopano ndipo adawonekera m'ma 90s. Za chiyani Tsopano aulesi okhawo sanamve. Kampani yotchuka kwambiri pano ndi Iridium, yomwe idasokonekera kumapeto kwa zaka za m'ma 90, koma idapulumutsidwa chifukwa cha US Department of Defense (osasokonezedwa ndi iRidium, dongosolo lakunyumba laku Russia). Pulojekiti ya Elon Musk (Starlink) ili kutali ndi yokhayo - Richard Branson (OneWeb - 1440 satellites akufuna), Boeing (3000 satellites), Samsung (4600 satellites), ndi ena akuchita nawo mpikisano wa satana.
Momwe zinthu zikuyimira mdera lino, momwe chuma chikuwonekera kumeneko - werengani mkati . Ndipo tikuyembekezera mayesero oyambirira a machitidwewa ndi ogwiritsa ntchito oyambirira, omwe ayenera kuchitika chaka chamawa.
2.4. Kusindikiza kwa Nanoscale 3D
Kusindikiza kwa 3D, ngakhale sikunalowe m'moyo wa munthu aliyense (mu mawonekedwe omwe adalonjezedwa ndi fakitale yapulasitiki yapanyumba), komabe kwasiya kale ukadaulo wa geeks. Mukhoza kuweruza kuti mwana aliyense wasukulu amadziwa za kukhalapo kwa zolembera zosachepera za 3D, ndipo ambiri amalota kugula bokosi ndi othamanga ndi extruder kwa ... "monga choncho" (kapena adagula kale).
Stereolithography (osindikiza a laser 3D) amalola kusindikiza ndi mafotoni apawokha: ma polima atsopano akufufuzidwa omwe amangofunika mafotoni awiri kuti akhale olimba. Izi zidzalola, muzinthu zopanda ma laboratory, kupanga zosefera zatsopano, zokwera, akasupe, ma capillaries, ma lens ndi ... zomwe mungasankhe mu ndemanga! Ndipo apa si patali ndi photopolymerization - luso lokhalo limatithandiza "kusindikiza" mapurosesa ndi mabwalo apakompyuta. Kuphatikiza apo, ino sichaka choyamba chomwe chakhalapo , koma popanda chitukuko champhamvu.
3. Kuzindikira ndi Kuyenda
3.1. Autonomous Driving Level 4 & 5
Kuti musasokonezedwe m'mawu, ndikofunikira kumvetsetsa kuti ndi magawo ati akudziyimira pawokha omwe amasiyanitsidwa (kuchokera mwatsatanetsatane. , komwe timalozera onse omwe ali ndi chidwi):
Level 1: Cruise control: imathandiza dalaivala pakanthawi kochepa (mwachitsanzo, kugwira galimoto pa liwiro lomwe dalaivala achotsa phazi lake)
Gawo 2: Thandizo lowongolera pang'ono ndi mabuleki. Dalaivala ayenera kukhala wokonzeka kuwongolera nthawi yomweyo. Manja ake ali pa chiwongolero, maso ake ali panjira. Izi ndi zomwe Tesla ndi General Motors ali nazo kale.
Gawo 3: Dalaivala sakuyeneranso kuyang'ana msewu nthawi zonse. Koma ayenera kukhala tcheru ndi kukhala wokonzeka kulamulira. Ichi ndi chinthu chomwe magalimoto omwe amapezeka pamalonda alibe. Onse omwe alipo pano ali pa level 1-2.
Mzere wa 4: Woyendetsa ndege weniweni, koma ndi zoletsa: maulendo okha m'dera lodziwika bwino lomwe limapangidwa mosamala komanso lodziwika bwino ndi dongosolo, komanso pansi pazikhalidwe zina: mwachitsanzo, popanda chipale chofewa. Waymo ndi General Motors ali ndi ma prototypes oterowo, ndipo akukonzekera kuwayambitsa m'mizinda ingapo ndikuwayesa m'malo enieni. Yandex ili ndi madera oyesera ma taxi opanda anthu ku Skolkovo ndi Innopolis: ulendowu umachitika moyang'aniridwa ndi injiniya wokhala pampando wa okwera; pofika kumapeto kwa chaka, kampaniyo ikukonzekera kukulitsa zombo zake ku magalimoto 100 opanda anthu.
Level 5: Kuyendetsa mokhazikika, kusinthira kwathunthu kwa dalaivala wamoyo. Machitidwe amenewa kulibe, ndipo n'zokayikitsa kuonekera mu zaka zikubwerazi.
Kodi ndi zoona bwanji kuona zonsezi m'tsogolomu? Apa ndikufuna ndikulozeranso owerenga nkhaniyi . Izi ndi zina chifukwa cha kusowa kwa kulumikizana kwa 5G: kuthamanga komwe kulipo kwa 4G sikokwanira. Mwa zina chifukwa cha kukwera mtengo kwa magalimoto odziyimira pawokha: iwo sali opindulitsa, mtundu wa bizinesi sudziwika bwino. Mwachidule, "zonse ndizovuta" pano, ndipo sizodabwitsa kuti Gartner akulemba kuti kulosera kwa kukhazikitsidwa kwakukulu kwa Levels 4 ndi 5 sikunayambepo kuposa zaka 10.
3.2. Makamera a 3D Sensing
Zaka zisanu ndi zitatu zapitazo, wowongolera masewera a Kinect a Microsoft adapanga mafunde popereka njira yofikirika komanso yotsika mtengo ya masomphenya a 3D. Kuyambira pamenepo, masewera olimbitsa thupi ndi masewera ovina omwe ali ndi Kinect adakwera pang'ono ndikuchepa, koma makamera a 3D adayamba kugwiritsidwa ntchito m'maloboti amakampani, magalimoto osayendetsedwa ndi anthu, komanso mafoni am'manja kuti adziwe nkhope. Zipangizo zamakono zakhala zotsika mtengo, zowonjezereka komanso zowonjezereka.
Foni ya Samsung S10 ili ndi kamera ya Time-of-Flight yomwe imayesa mtunda wa chinthu kuti kuyang'ana kwake kukhale kosavuta.
Ngati muli ndi chidwi ndi mutuwu, tikulozeraninso kuwunika kwatsatanetsatane kwamakamera akuya: , .
3.3. Ma Drones operekera katundu waung'ono (Ma Drones a Light Cargo Delivery)
Chaka chino, Amazon idapanga mafunde pomwe idawonetsa drone yatsopano yowuluka pachiwonetsero yomwe imatha kunyamula katundu wocheperako mpaka 2 kg. Kwa mzinda womwe uli ndi kuchuluka kwa magalimoto pamsewu, izi zikuwoneka ngati yankho labwino. Tiyeni tiwone momwe ma droneswa amachitira posachedwa. Mwina ndizoyenera kukayikira mosamala apa: pali mavuto ambiri, kuyambira ndi kuthekera kwa kuba kosavuta kwa drone, ndikutha ndi zoletsa zamalamulo pa ma UAV. Amazon Prime Air yakhalapo kwa zaka zisanu ndi chimodzi koma ikadali muyeso.
Drone yatsopano ya Amazon, yowonetsedwa masika. Pali chinachake cha Star Wars za iye.
Kuphatikiza pa Amazon, pali osewera ena pamsika (pali mwatsatanetsatane ), koma palibe chinthu chimodzi chomaliza: chirichonse chiri pa siteji ya kuyesa ndi kutsatsa malonda. Payokha, m'pofunika kudziwa chidwi kwambiri zapaderazi zachipatala mu Africa: kutumizidwa kwa magazi operekedwa ku Ghana (14 obweretsa, kampani ya Zipline) ndi Rwanda (kampani ya Matternet).
3.4. Magalimoto Oyenda Pawokha
Ndizovuta kunena chilichonse chotsimikizika apa. Malinga ndi Gartner, izi sizidzawoneka kale kuposa zaka 10. Kawirikawiri, pali mavuto omwewo monga magalimoto oyendetsa okha, okhawo amapeza gawo latsopano - ofukula. Porsche, Boeing, ndi Uber alengeza zokhumba zawo zopanga taxi yowuluka.
3.5. Augmented Reality Cloud (AR Cloud)
Kope ya digito yokhazikika yadziko lenileni, kukulolani kuti mupange gawo latsopano la zenizeni zomwe zimafanana ndi ogwiritsa ntchito onse. M'mawu aukadaulo, tikukamba za kupanga nsanja yotseguka yamtambo momwe opanga amatha kuphatikiza mapulogalamu awo a AR. Njira yopangira ndalama ndiyomveka bwino; ndi mtundu wa analogi wa Steam. Lingaliroli lakhazikika kwambiri kotero kuti ena tsopano akukhulupirira kuti AR popanda mtambo ndiyopanda ntchito.
Momwe izi zingawonekere m'tsogolomu zikuwonetsedwa muvidiyo yayifupi. Zikuwoneka ngati gawo lina la Black Mirror:
Mukhozanso kuwerenga pa
4. Munthu Wowonjezera
4.1. Emotion AI
Kodi mungayeze bwanji, kutsanzira komanso kuchitapo kanthu ndi momwe anthu akumvera? Ena mwamakasitomala pano ndi makampani omwe amapanga othandizira mawu ngati Amazon Alexa. Amatha kuzolowerana ndi nyumba ngati aphunzira kuzindikira momwe akumvera: kumvetsetsa chifukwa cha kusakhutira kwa wogwiritsa ntchito, ndikuyesa kukonza vutoli. Nthawi zambiri, pali zambiri zambiri m'nkhaniyo kuposa uthenga womwewo. Ndipo nkhani ndi maonekedwe a nkhope, kalankhulidwe, ndi khalidwe lopanda mawu.
Ntchito zina zothandiza: kusanthula malingaliro pa nthawi yofunsa mafunso (kutengera zoyankhulana za kanema), kuwunika momwe amachitira ndi malonda kapena mavidiyo ena (kumwetulira, kuseka), kuthandizira pakuphunzira (mwachitsanzo, kuchita zodziyimira pawokha mu luso lolankhula pagulu).
Ndizovuta kuyankhula bwino pamutuwu kuposa wolemba filimu yayifupi ya mphindi 6 . Kanema wanzeru komanso wotsogola akuwonetsa momwe mungayezere malingaliro athu pazolinga zamalonda, komanso momwe nkhope yanu ikuwonera kwakanthawi, fufuzani ngati mumakonda pizza, agalu, Kanye West, komanso kuchuluka kwa ndalama zomwe mumapeza komanso IQ pafupifupi. Poyendera tsamba la filimuyo pogwiritsa ntchito ulalo womwe uli pamwambapa, mukhala otenga nawo gawo mu kanema wokambitsirana pogwiritsa ntchito kamera yokhazikika ya laputopu yanu. Filimuyi yawonetsedwa kale pamaphwando angapo amafilimu.
Palinso phunziro losangalatsa lotere: momwe mungazindikire mawu achipongwe m'mawu. Tinatenga ma tweets ndi hashtag #sarcasm ndipo tinapanga maphunziro a ma tweets 25 ndi mawu achipongwe komanso ma tweets 000 nthawi zonse pa chilichonse pansi pa dzuwa. Tidagwiritsa ntchito laibulale ya TensorFlow, kuphunzitsa dongosolo, ndipo zotsatirazi ndi izi:
Chifukwa chake, tsopano, ngati simukutsimikiza za mnzanu kapena bwenzi - adakuuzani mozama kapena monyodola, mutha kugwiritsa ntchito kale. !
4.2. Augmented Intelligence
Kugwiritsa ntchito mwanzeru ntchito pogwiritsa ntchito njira zophunzirira makina. Zingawoneke ngati zatsopano? Koma mawuwo ndiwofunikira pano, makamaka chifukwa amagwirizana ndi chidule cha Artificial Intelligence. Izi zimatibweretsanso ku mkangano wokhudza AI "yamphamvu" ndi "yofooka".
AI yamphamvu ndi luntha lochita kupanga lomwelo kuchokera m'mafilimu opeka asayansi omwe ali ofanana kwathunthu ndi malingaliro amunthu ndipo amadzizindikira yekha ngati payekha. Izi palibe pano ndipo sizikudziwika ngati zidzakhalapo konse.
Ofooka AI si munthu wodziimira yekha, koma wothandizira waumunthu. Sadzinenera kuti ali ndi malingaliro onga munthu, koma amangodziwa momwe angathetsere mavuto a chidziwitso, mwachitsanzo, kudziwa zomwe zikuwonetsedwa pa chithunzi kapena kumasulira malemba.
M'lingaliro limeneli, Augmented Intelligence ndi "AI yofooka" mu mawonekedwe ake oyera, ndipo mapangidwe ake amawoneka opambana, chifukwa samayambitsa chisokonezo ndi chiyeso chowona pano "AI yamphamvu" yomwe aliyense amalota (kapena mantha, ngati ife titati tipeze chisokonezo). kukumbukira zokambirana zambiri za "magalimoto opanduka"). Pogwiritsa ntchito mawu akuti Augmented Intelligence, nthawi yomweyo timakhala ngwazi za filimu ina: kuchokera ku nthano za sayansi (monga "I, Robot" ya Asimov) timadzipeza tokha mu cyberpunk ("augmentations" mumtundu uwu ndi mitundu yonse ya implants yomwe imakulitsa luso laumunthu).
Kodi Erik Brynjolfsson ndi Andrew McAffee: “Pazaka 10 zikubwerazi, izi ndi zomwe zidzachitike. AI sidzalowa m'malo mwa oyang'anira, koma oyang'anira omwe amagwiritsa ntchito AI alowa m'malo mwa omwe sanapangepo. "
zitsanzo:
- Mankhwala: Yunivesite ya Stanford idapangidwa , amene akulimbana ndi ntchito yozindikira matenda pachifuwa X-rays pafupifupi bwino monga madokotala ambiri
- Maphunziro: thandizo kwa ophunzira ndi aphunzitsi, kusanthula mayankho a ophunzira ku zipangizo, kumanga njira yophunzirira payekha.
- Ma analytics a bizinesi: kukonzanso kwa data, malinga ndi ziwerengero, kumatenga 80% ya nthawi ya ofufuza, ndipo 20% yokha ya kuyesa komweko.
4.3. Biochips
Uwu ndiye mutu womwe mumakonda kwambiri pamakanema ndi mabuku onse a cyberpunk. Nthawi zambiri, ziweto zazing'ono sichizoloŵezi chatsopano. Koma tsopano tchipisi izi zayamba kubzalidwa mwa anthu.
Pachifukwa ichi, hype nthawi zambiri imagwirizanitsidwa ndi nkhani yochititsa chidwi mu kampani yaku America ya Three Square Market. Kumeneko, bwanayo anayamba kupereka ndalama zoikamo tchipisi pansi pa khungu kuti amulipirire. Chip chimakulolani kuti mutsegule zitseko, kulowa mu makompyuta, kugula zokhwasula-khwasula kuchokera ku makina ogulitsa - ndiko kuti, khadi la antchito onse. Komanso, chip choterocho chimagwira ntchito ngati chizindikiritso cha khadi; ilibe gawo la GPS, kotero ndizosatheka kutsatira aliyense amene akugwiritsa ntchito. Ndipo ngati munthu akufuna kuchotsa chip pa mkono wake, zimatenga mphindi 5 mothandizidwa ndi dokotala.
Tchipisi nthawi zambiri zimayikidwa pakati pa chala chachikulu ndi chala cholozera.
Werengani zambiri za momwe zinthu zilili padziko lapansi.
4.4. Immersive Workspace
“Kumiza” ndi liwu linanso latsopano lomwe silingathe kuthawira. Zili paliponse. Immersive zisudzo, chiwonetsero, cinema. Mukutanthauza chiyani? Kumizidwa ndiko kupanga kozama, pamene malire pakati pa wolemba ndi wowonera, zenizeni ndi dziko lenileni zimatayika. Kuntchito, mwina, izi zikutanthauza kusokoneza mzere pakati pa wochita ndi woyambitsa ndi kulimbikitsa ogwira ntchito kuti achitepo kanthu pokonzanso chilengedwe chawo.
Popeza tsopano tili ndi Agile, kusinthasintha, ndi mgwirizano wapamtima kulikonse, malo ogwira ntchito ayenera kukhala osinthika mosavuta momwe angathere ndipo ayenera kulimbikitsa ntchito zamagulu. Chuma chimayang'ana mawu ake: pali antchito osakhalitsa, mtengo wobwereketsa ofesi ukukwera, ndipo mumsika wogwira ntchito wopikisana, makampani a IT akuyesera kuonjezera chikhutiro cha ogwira ntchito kuchokera kuntchito popanga malo osangalatsa ndi zopindulitsa zina. Ndipo zonsezi zikuwonekera pamapangidwe a malo ogwira ntchito.
Kuchokera kulira
4.5. Munthu
Aliyense amadziwa zomwe makonda pa malonda. Apa ndi pamene lero mukukambirana ndi mnzanu kuti mpweya m'chipindamo ndi wouma pang'ono, ndipo muyenera kugula chonyowa ku ofesi, ndipo tsiku lotsatira muwona zotsatsa pa malo ochezera a pa Intaneti - "gulani humidifier" (a zochitika zenizeni zomwe zidandichitikira).
Kupanga makonda, monga momwe Gartner amafotokozera, ndikuyankha ku nkhawa zomwe zikuchulukirachulukira kwa ogwiritsa ntchito pakugwiritsa ntchito deta yawo pazotsatsa. Cholinga chake ndi kupanga njira yomwe timawonetsera malonda omwe ali ogwirizana ndi zomwe timakumana nazo, osati kwa ife payekha. Mwachitsanzo, malo athu, mtundu wa chipangizo, nthawi ya tsiku, nyengo - ichi ndi chinthu chomwe sichimaphwanya deta yathu, ndipo sitimva kumverera kosautsa kwa "kuyang'aniridwa".
Werengani za kusiyana kwa mfundo ziwirizi Andrew Frank amalemba mabulogu patsamba la Gartner. Pali kusiyana kobisika komanso mawu ofanana kotero kuti inu, osadziwa kusiyana kwake, mumakhala pachiwopsezo chotsutsana kwa nthawi yayitali ndi wolankhulana naye, osakayikira kuti, onsewo ndi olondola (ndipo izi ndizochitika zenizeni zomwe zidachitika kwa wolemba).
4.6. Biotech - Minofu Yokhazikika kapena Yopanga
Ichi ndi, choyamba, lingaliro la kukula nyama yokumba. Panthawi imodzimodziyo, magulu angapo padziko lonse lapansi ali otanganidwa kupanga labotale "Nyama 2.0" - zikuyembekezeka kukhala zotsika mtengo kuposa masiku onse, ndipo zakudya zofulumira komanso masitolo akuluakulu zidzasintha. Otsatsa ndalama muukadaulo uwu akuphatikizapo Bill Gates, Sergey Brin, Richard Branson ndi ena.
Kuchokera
Zifukwa zomwe aliyense amakonda kwambiri nyama yopangira:
- Kutentha kwapadziko lonse: mpweya wa methane wochokera m'mafamu. Ichi ndi 18% ya kuchuluka kwa mpweya wapadziko lonse womwe umakhudza nyengo.
- Kuchuluka kwa anthu. Kufunika kwa nyama kukukulirakulira, ndipo sikungatheke kudyetsa aliyense ndi nyama zachilengedwe - ndizokwera mtengo.
- Kusowa malo. 70% ya nkhalango za Amazon zadulidwa kale kuti zikhale msipu.
- Malingaliro amakhalidwe abwino. Pali ena omwe izi ndizofunikira. Bungwe lomenyera ufulu wa nyama PETA lapereka kale mphotho ya $ 1 miliyoni kwa wasayansi yemwe amabweretsa nyama ya nkhuku yopangira msika.
Kusintha nyama yeniyeni ndi soya ndi njira yothetsera pang'ono, chifukwa anthu amatha kuyamikira kusiyana kwa kukoma ndi kapangidwe kake, ndipo sangathe kusiya steak mokomera soya. Chifukwa chake mumafunikira nyama yeniyeni, yolimidwa mwachilengedwe. Tsopano, mwatsoka, nyama yochita kupanga ndi yokwera mtengo kwambiri: kuchokera ku $ 12 pa kilogalamu. Izi zimachitika chifukwa chaukadaulo wovuta kukulitsa nyama yotere. Werengani za izo zonse .
Ngati tilankhula za matenda ena akukula kwa minofu - kale muzamankhwala - ndiye kuti mutu wokhala ndi ziwalo zopangira ndi wosangalatsa: mwachitsanzo, "chigamba" cha minofu yamtima, chosindikizira chapadera cha 3D. Wodziwika monga mtima wa mbewa wokulirapo, koma zonse zikadali mkati mwa mayeso azachipatala. Chifukwa chake sitingathe kuwona Frankenstein m'zaka zikubwerazi.
Apa Gartner ali wosamala kwambiri pakuyerekeza kwake, akuwoneka kuti akukumbukira zomwe adaneneratu zomwe zidalephera mu 2015 kuti mu 2019, 10% ya anthu akumayiko otukuka adzakhala ndi choyika chachipatala chosindikizidwa cha 3D. Chifukwa chake, zikutanthauza kuti nthawi yofikira malo opangira zokolola ndi zaka 10.
5. Digital Ecosystems
5.1. Decentralized Web
Lingaliro ili likugwirizana kwambiri ndi dzina la woyambitsa intaneti, wopambana mphoto ya Turing Sir Tim Burners-Lee. Kwa iye, mafunso a zamakhalidwe mu sayansi yamakompyuta anali ofunikira nthawi zonse ndipo mfundo zonse zapaintaneti zinali zofunika: kuyika maziko a hypertext, adatsimikiza kuti intaneti iyenera kugwira ntchito ngati ukonde, osati ngati utsogoleri. Izi zinali choncho kumayambiriro kwa chitukuko cha intaneti. Komabe, pomwe intaneti idakula, mawonekedwe ake adakhazikika pazifukwa zosiyanasiyana. Zinapezeka kuti mwayi wopita ku netiweki wa dziko lonse ukhoza kutsekedwa mosavuta mothandizidwa ndi opereka ochepa chabe. Ndipo deta ya ogwiritsa ntchito yakhala gwero la mphamvu ndi ndalama kwa makampani a intaneti.
Burners-Lee anati: “Intaneti ndi yoletsedwa kale. "Vuto ndiloti injini imodzi yosakira, malo ochezera a pa Intaneti, malo amodzi a microblogging amalamulira. Tilibe mavuto aukadaulo, koma tili ndi mavuto azachuma. ”
Mwa iye Kwa zaka 30 za World Wide Web, mlengi wa Webusaiti adalongosola zovuta zazikulu zitatu za intaneti:
- Zovulaza zomwe anthu amakumana nazo monga kubera mothandizidwa ndi boma, umbanda, komanso kuzunza anthu pa intaneti
- Mapangidwe omwewo, omwe, mowononga wogwiritsa ntchito, amapangitsa njira zopangira njira monga: zolimbikitsira ndalama za clickbait ndi kufalikira kwa ma virus abodza.
- Zotsatira zosayembekezereka zamakina adongosolo zomwe zimadzetsa mikangano ndikuchepetsa makambirano a pa intaneti
Ndipo Tim Berners-Lee ali kale ndi yankho pa mfundo zomwe "Intaneti ya Munthu Wathanzi" ingakhazikitsidwe, yopanda vuto nambala 2: "Kwa ogwiritsa ntchito ambiri, ndalama zotsatsa zimakhalabe njira yokhayo yolumikizirana ndi intaneti. Ngakhale anthu akuwopa zomwe zimachitika ku deta yawo, ali okonzeka kupanga mgwirizano ndi makina otsatsa kuti apeze mwayi wolandira zomwe zili kwaulere. Tangoganizani dziko limene kulipirira katundu ndi ntchito n’kosavuta ndiponso kosangalatsa kwa onse awiri.” Zina mwazosankha za momwe izi zingakonzedwere: oimba amatha kugulitsa zojambulira zawo popanda oyimira pakati pa iTunes, ndipo masamba ankhani amatha kugwiritsa ntchito njira yolipirira ma micropayments powerenga nkhani imodzi, m'malo mopanga ndalama pakutsatsa.
Monga chitsanzo choyesera pa intaneti yatsopanoyi, a Tim Berners-Lee adayambitsa pulojekiti ya SOLID, makamaka yomwe imasunga deta yanu mu "pod" - malo osungiramo zidziwitso, ndipo angapereke detayi kwa anthu ena. Koma kwenikweni, inu nokha ndinu ambuye a deta yanu. Zonsezi zimagwirizana kwambiri ndi lingaliro la ma intaneti a anzawo, ndiko kuti, kompyuta yanu simangopempha mautumiki, komanso imapereka, kuti musadalire seva imodzi ngati njira yokhayo.
5.2. Decentralized Autonomous Organizations
Ndi bungwe limene limayendetsedwa ndi malamulo olembedwa ngati pulogalamu ya pakompyuta. Ntchito zake zachuma zimachokera ku blockchain. Cholinga chokhazikitsa mabungwe oterowo ndikuchotsa boma paudindo wa mkhalapakati ndikupanga malo odalirika ogwirizana, omwe si a aliyense payekhapayekha, koma omwe ali ndi aliyense palimodzi. Ndiko kuti, m'malingaliro, izi zikuyenera, ngati lingalirolo lizika mizu, kuthetseratu notary ndi mabungwe ena otsimikizira.
Chitsanzo chodziwika bwino cha bungwe loterolo chinali DAO yoyang'ana malonda, yomwe inakweza $ 2016 miliyoni mu 150, yomwe $ 50 inabedwa nthawi yomweyo kupyolera mu dzenje lalamulo mu malamulo. Vuto lovuta lidayamba nthawi yomweyo: mwina kubweza ndikubweza ndalamazo, kapena kuvomereza kuti kuchotsedwa kwa ndalama kunali kovomerezeka, chifukwa sikuphwanya malamulo a nsanja. Chotsatira chake, kuti abwezeretse ndalama kwa osunga ndalama, olenga adayenera kuwononga DAO, kulembanso blockchain ndikuphwanya mfundo yake yayikulu - kusasinthika.
Comic za Ethereum (kumanzere) ndi The DAO (kumanja).
Nkhani yonseyi yawononga mbiri ya lingaliro lomwe la DAO. Ntchito imeneyo inapangidwa pamaziko a Ethereum cryptocurrency, Baibulo Etere 2.0 akuyembekezeredwa chaka chamawa - mwina olemba (kuphatikizapo wotchuka Vitalik Buterin) adzaganizira zolakwa ndi kusonyeza chinachake chatsopano. Mwina ndichifukwa chake Gartner adayika DAO pamzere.
5.3.Zophatikizika Zambiri
Kuphunzitsa ma neural network, kuchuluka kwa data kumafunika. Kulemba deta pamanja ndi ntchito yaikulu yomwe ingakhoze kuchitidwa ndi munthu. Choncho, n'zotheka kupanga ma data ochita kupanga. Mwachitsanzo, zosonkhanitsira yemweyo wa nkhope za anthu pa malo . Amapangidwa pogwiritsa ntchito GAN - ma algorithms omwe atchulidwa kale pamwambapa.
Nkhope zimenezi si za anthu.
Ubwino waukulu wa deta yotere ndikuti palibe zovuta zamalamulo pakuzigwiritsa ntchito: palibe amene angavomereze kukonzedwa kwazinthu zaumwini.
5.4.Digital Ops
Chokwanira "Ops" chakhala chapamwamba kwambiri kuyambira pomwe DevOps idakhazikika pamalankhulidwe athu. Tsopano za zomwe DigitalOps ili - ndizongowonjezera za DevOps, DesignOps, MarketingOps ... Kodi mwatopa panobe? Mwachidule, ndikusuntha kwa njira ya DevOps kuchokera kumalo a mapulogalamu kupita kuzinthu zina zonse zamalonda - malonda, mapangidwe, ndi zina zotero.
Lingaliro la DevOps linali kuchotsa zotchinga pakati pa Chitukuko chokha ndi Ntchito (njira zamabizinesi), kudzera pakupanga magulu wamba, komwe kuli opanga mapulogalamu, oyesa, akatswiri achitetezo, ndi oyang'anira; kukhazikitsa njira zina: kuphatikiza kosalekeza, zomangamanga monga ma code, kuchepetsa ndi kulimbikitsa maunyolo oyankha. Cholinga chake chinali kufulumizitsa nthawi yogulitsira malonda. Ngati mumaganiza kuti izi ndizofanana ndi Agile, munalondola. Tsopano sinthani njira iyi m'malingaliro kuchokera kugawo lachitukuko cha mapulogalamu kupita ku chitukuko chonse - ndipo mumamvetsetsa kuti DigitalOps ndi chiyani.
5.5. Zithunzi Zachidziwitso
Njira yamapulogalamu yowonetsera gawo lachidziwitso, kuphatikiza kugwiritsa ntchito makina ophunzirira makina. Grafu ya chidziwitso imamangidwa pamwamba pa nkhokwe zomwe zilipo kuti zilumikize zonse: zonse zokonzedwa (mndandanda wa zochitika kapena anthu) komanso zosakonzedwa (zolemba zankhani).
Chitsanzo chosavuta kwambiri ndi khadi lomwe mutha kuwona pazotsatira zakusaka kwa Google. Ngati mukuyang'ana munthu kapena bungwe, mudzawona khadi kumanja:
Chonde dziwani kuti "Zochitika Zomwe Zikubwera" sizomwe zimachokera ku Google Maps, koma kugwirizanitsa ndondomeko ndi Yandex.Afisha: mukhoza kuona izi mosavuta ngati mutsegula pazochitikazo. Ndiko kuti, ndi kuphatikiza angapo deta magwero pamodzi.
Ngati mupempha mndandanda - mwachitsanzo, "otsogolera otchuka" - mudzawonetsedwa carousel:
Bonasi kwa omwe amawerenga mpaka kumapeto
Ndipo tsopano kuti tadzifotokozera tokha tanthauzo la mfundo iliyonse, tikhoza kuyang'ana chithunzi chomwecho, koma mu Russian:
Gawani momasuka pama social network!
Tatyana Volkova - Wolemba pulogalamu yophunzitsira ya Internet of Things IT track ku Samsung Academy, katswiri wamapulogalamu okhudzana ndi chikhalidwe cha anthu ku Samsung Research Center.
Source: www.habr.com