Taswirar Gartner kamar babban nuni ne ga waɗanda ke cikin masana'antar fasaha. Ta hanyar duba shi, za ku iya gano tun da wuri waɗanne kalmomi ne aka fi ɗauka a wannan kakar da abin da za ku ji a duk taron da ke tafe.

Mun fayyace abin da ke bayan kyawawan kalmomi a cikin wannan jadawali don ku iya magana da yaren kuma.

Don farawa da, kawai 'yan kalmomi game da wane irin jadawali ne wannan. Kowace shekara a watan Agusta, hukumar tuntuɓar Gartner tana fitar da rahoto - Gartner Hype Curve. A cikin Rashanci, wannan shine "ƙarfin haɓaka," ko, mafi sauƙi, ƙararrawa. Shekaru 30 da suka gabata, mawakan rap na ƙungiyar Maƙiyin Jama'a sun rera waƙa: “Kada ku yarda da zaɓen.” Ku yi imani da shi ko a'a, tambaya ce ta sirri, amma yana da daraja aƙalla sanin waɗannan kalmomi idan kuna aiki a fagen fasaha kuma kuna son sanin yanayin duniya.

Wannan jadawali ne na tsammanin jama'a daga wata fasaha ta musamman. A cewar Gartner, da kyau, fasaha tana tafiya ta matakai 5: ƙaddamar da fasaha, kololuwar tsammanin buri, kwarin rashin jin daɗi, gangaren wayewar kai, tudun mun tsira. Amma kuma yana faruwa cewa ya nutse a cikin "kwarin cizon yatsa" - zaku iya tunawa da misalan da kanku cikin sauƙi, ɗauki bitcoins iri ɗaya: da farko buga kololuwa a matsayin "kuɗin nan gaba", da sauri sun zame ƙasa lokacin da gazawar fasaha. ya zama a bayyane, da farko ƙuntatawa akan adadin ma'amaloli da kuma yawan adadin wutar lantarki da ake buƙata don samar da bitcoins (wanda ya riga ya haifar da matsalolin muhalli). Kuma ba shakka, kada mu manta cewa ginshiƙi na Gartner hasashe ne kawai: a nan, alal misali, zaku iya karanta cikakken bayani. labarin, inda aka tsara hasashen da ba a cika ba.

Don haka, bari mu wuce sabon ginshiƙi na Gartner. An raba fasahohin zuwa manyan rukunin jigo guda 5:

1. Advanced AI da Analytics
2. Ƙididdigar Postclassical da Comms
3. Hankali da Motsi
4. Augmented Adam
5. Tsarin Muhalli na Dijital

1. Advanced AI da Analytics

A cikin shekaru 10 da suka gabata mun ga mafi kyawun sa'a na zurfafa ilmantarwa. Waɗannan cibiyoyin sadarwa suna da inganci da gaske don yawan ayyukansu. A cikin 2018, Yann LeCun, Geoffrey Hinton da Yoshua Bengio sun sami lambar yabo ta Turing don binciken da suka yi - lambar yabo mafi girma, kwatankwacin lambar yabo ta Nobel a kimiyyar kwamfuta. Don haka, manyan abubuwan da ke faruwa a wannan yanki, waɗanda aka nuna akan ginshiƙi:

1.1. Canja wurin Koyo

Ba kwa horar da hanyar sadarwa ta jijiyoyi daga karce, amma ɗauki wanda aka riga aka horar kuma sanya shi wata manufa ta daban. Wani lokaci wannan yana buƙatar sake horar da sashin cibiyar sadarwa, amma ba duka cibiyar sadarwa ba, wanda ya fi sauri. Misali, ɗaukar hanyar sadarwar jijiyoyi ResNet50, wanda aka horar akan bayanan bayanan ImageNet1000, zaku sami algorithm wanda zai iya rarraba abubuwa daban-daban a cikin hoto a matakin zurfin zurfi (azuzuwan 1000 dangane da fasalulluka waɗanda 50 yadudduka na jijiya suka haifar. network). Amma ba dole ba ne ka horar da wannan hanyar sadarwa duka, wanda zai ɗauki watanni.

В kan layi kwas Samsung "Neural networks da kwamfuta hangen nesa", misali, a karshe Kaggle aiki tare da rarraba faranti a cikin tsabta da ƙazanta, ana nuna hanya cewa a cikin minti 5 yana ba ku hanyar sadarwa mai zurfi mai zurfi wanda zai iya bambanta faranti mai datti daga masu tsabta, wanda aka gina bisa ga gine-ginen da aka kwatanta a sama. Cibiyar sadarwa ta asali ba ta san abin da faranti ba kwata-kwata, kawai ta koyi bambanta tsuntsaye da karnuka (duba ImageNet).

source: kan layi kwas Samsung "Neural networks da kwamfuta hangen nesa"

Don Canja wurin Koyo kuna buƙatar sanin waɗanne hanyoyi ne ke aiki da waɗanne shirye-shiryen gine-ginen da aka ƙera. Gabaɗaya, wannan yana haɓaka bullowar aikace-aikace masu amfani na koyo na inji.

1.2. Generative Adversarial Networks (GAN)

Wannan na waɗancan lokuta ne lokacin da yake da wahala a gare mu mu tsara manufar koyo. Mafi kusancin aikin shine rayuwa ta ainihi, mafi fahimtar shi a gare mu ("kawo teburin gefen gado"), amma mafi wahala shine tsara shi azaman aikin fasaha. GAN ƙoƙari ne kawai na ceto mu daga wannan matsala.

Akwai cibiyoyin sadarwa guda biyu da ke aiki a nan: ɗaya janareta (Generative), ɗayan kuma mai nuna bambanci (Adversarial). Ɗayan hanyar sadarwa tana koyon yin aiki mai amfani (raba hotuna, gane sautuna, zana zane-zane). Kuma wata cibiyar sadarwa ta koyi koyar da wannan hanyar sadarwa: tana da misalai na gaske, kuma ta koyi samun wata dabarar da ba a san ta ba a baya don kwatanta samfuran sashin samar da hanyar sadarwa tare da abubuwa na zahiri (tsarin horo) dangane da ainihin mahimman halaye masu zurfi. : yawan idanu, kusanci da salon Miyazaki, daidaitaccen lafazin Turanci.

Misalin sakamakon hanyar sadarwa don ƙirƙirar haruffan anime. Source

Amma, ba shakka, yana da wuya a gina gine-gine a can. Bai isa kawai jefa neurons ba, suna buƙatar shirya. Kuma dole ne ku yi karatun makonni. Abokan aiki na a Cibiyar Intelligence Intelligence ta Samsung suna aiki akan batun GAN; wannan shine ɗayan mahimman tambayoyin binciken su. Misali, kamar wannan ci gaba: yin amfani da hanyoyin sadarwa na zamani don haɗa hotuna na gaskiya na mutanen da ke da matsayi daban-daban - alal misali, don ƙirƙirar ɗaki mai dacewa, ko haɗa fuska, wanda zai iya rage adadin bayanan da ake buƙatar adanawa ko aikawa don tabbatar da ingancin bidiyo mai inganci. sadarwa, watsa shirye-shirye ko kariyar bayanan sirri.

Source

1.3. AI mai bayyanawa

Ga wasu ayyuka da ba kasafai ba, ci gaba a cikin gine-gine masu zurfi sun kawo kwatsam kwatsam ikon hanyoyin sadarwa masu zurfi kusa da iyawar mutum. Yanzu an fara yaƙi don ƙara yawan irin waɗannan ayyuka. Misali, injin tsabtace mutum-mutumi na iya bambanta cat da kare cikin sauƙi a cikin taron gaba-gaba. Amma a mafi yawan yanayi na rayuwa, ba zai iya samun cat da ke barci a tsakanin lilin ko kayan daki (duk da haka, kamar mu, a mafi yawan lokuta ...).

Menene dalilin nasarar nasarar hanyoyin sadarwa mai zurfi? Suna haɓaka wakilcin matsalar ba bisa bayanan "bayyanannu ga ido tsirara" (pixels na hoto, canje-canje a cikin ƙarar sauti ...), amma akan sifofin da aka samu bayan ƙaddamar da wannan bayanin ta hanyar ɗaruruwan yadudduka na cibiyar sadarwa na jijiyoyi. Abin takaici, waɗannan alaƙar na iya zama marasa ma'ana, rashin daidaituwa, ko ɗaukar alamun rashin ƙarfi a cikin saitin bayanan asali. Misali, akwai karamin wasan kwamfuta game da abin da rashin tunani na AI wajen daukar ma'aikata zai iya haifar da shi Tsira Mafi Kyau.

Na'urar tagging na hoton an sanya wa mutumin da ke dafa abinci a matsayin mace, duk da cewa mutumin da ke cikin hoton a zahiri namiji ne (Source). Wannan lura a Cibiyar Nazarin Virginia.

Don bincika hadaddun dangantaka mai zurfi da zurfi waɗanda sau da yawa ba za mu iya tsara kanmu ba, ana buƙatar hanyoyin AI masu bayyanawa. Suna tsara fasalin hanyoyin sadarwar jijiyoyi masu zurfi don haka bayan horo, za mu iya nazarin wakilcin cikin gida da cibiyar sadarwar ta koya, maimakon dogaro da shawararta kawai.

1.4. Edge Analytics / AI

Duk abin da ke da kalmar Edge a zahiri yana nufin mai zuwa: canja wurin wani ɓangare na algorithms daga gajimare / uwar garken zuwa ƙarshen na'urar / matakin ƙofar. Irin wannan algorithm zai yi aiki da sauri kuma ba zai buƙaci haɗi zuwa uwar garken tsakiya don aikinsa ba. Idan kun saba da abstraction na "abokin ciniki na bakin ciki," to a nan muna yin wannan abokin ciniki ɗan kauri.
Wannan na iya zama mahimmanci ga Intanet na Abubuwa. Misali, idan injin ya yi zafi sosai kuma yana buƙatar sanyaya, yana da ma'ana don sigina wannan nan da nan, a matakin shuka, ba tare da jiran bayanan zuwa gajimare ba kuma daga can zuwa ma'aikacin motsi. Ko kuma wani misali: motoci masu tuka kansu na iya gano yanayin zirga-zirga da kansu, ba tare da tuntuɓar uwar garken tsakiya ba.

Source

Ko kuma wani misali na dalilin da ya sa hakan ke da mahimmanci ta fuskar tsaro: lokacin da kake buga rubutu a wayarka, yana tuna kalmomin da suka saba maka, ta yadda daga baya maballin wayar zai iya sa ka dace da su - wannan shi ake kira tsinkaya. shigar da rubutu. Aika duk abin da ka rubuta akan madannai naka zuwa cibiyar bayanai a wani wuri zai zama cin zarafin sirrinka kuma mara lafiya kawai. Saboda haka, horar da madannai yana faruwa ne kawai a cikin na'urarka kanta.

1.5. AI Platform azaman Sabis (AI PaaS)

PaaS - Platform-as-a-Service shine samfurin kasuwanci wanda muke samun damar yin amfani da haɗin gwiwar dandamali, gami da ajiyar bayanan tushen girgije da shirye-shiryen shirye-shiryen. Ta wannan hanyar, za mu iya 'yantar da kanmu daga ayyukan samar da ababen more rayuwa kuma mu mai da hankali sosai kan samar da wani abu mai amfani. Misalin dandamali na PaaS don ayyukan AI: IBM Cloud, Microsoft Azure, Koyon Injin Amazon, Google AI Platform.

1.6. Koyon Injin Adaɗi (Maɗaukakin ML)

Me zai faru idan muka bar ilimin wucin gadi ya daidaita... Kuna tambaya - wato, ta yaya?... Shin bai riga ya dace da aikin ba? Matsalar ita ce: muna zazzagewa da zayyana kowace irin wannan matsala kafin gina algorithm na hankali don magance ta. Za su amsa muku - ya bayyana cewa ana iya sauƙaƙe wannan sarkar.

Koyon na'ura na al'ada yana aiki akan ka'idar buɗe madauki: kuna shirya bayanai, fito da hanyar sadarwa ta jijiyoyi (ko wani abu), jirgin ƙasa, sannan duba alamun da yawa, kuma idan kuna son komai, zaku iya aika cibiyar sadarwar jijiyoyi zuwa wayoyi. don magance matsalolin masu amfani. Amma a aikace-aikace inda akwai bayanai da yawa kuma yanayinsa yana canzawa a hankali, ana buƙatar wasu hanyoyin. Irin waɗannan tsarin, waɗanda ke daidaitawa da koyar da kansu, an tsara su cikin rufaffiyar madaukai na koyo (rufe-madauki), kuma dole ne su yi aiki lafiya.

Aikace-aikace - wannan na iya zama stream analytics (Stream Analytics), a kan tushen da yawancin 'yan kasuwa ke yanke shawara, ko sarrafa samar da daidaitacce. A cikin sikelin aikace-aikacen yanzu kuma an ba da ƙarin fahimtar haɗarin mutane, dabarun da ke samar da mafita ga wannan matsalar duk an tattara su ƙarƙashin laima Adaptive AI.

Source

Duban wannan hoton, yana da wuya a kawar da jin daɗin da ba sa ciyar da gurasar futurologists - bari su koya wa robot numfashi ...

Ƙididdigar Postclassical da Comms

2.1. Sadarwar wayar hannu ta ƙarni na biyar (5G)

Wannan batu ne mai ban sha'awa wanda nan da nan za mu mayar da ku zuwa ga namu labarin. To, ga ɗan taƙaitaccen bayani. 5G, ta hanyar haɓaka yawan watsa bayanai, zai sa Intanet ta yi saurin sauri ba tare da haƙiƙa ba. Yana da wuya ga gajeren raƙuman ruwa don wucewa ta hanyar cikas, don haka zane na cibiyoyin sadarwa zai zama daban-daban: sau 500 ana buƙatar ƙarin tashoshin tushe.

Tare da sauri, za mu sami sababbin abubuwan mamaki: wasanni na lokaci-lokaci tare da gaskiyar haɓakawa, yin ayyuka masu rikitarwa (kamar tiyata) ta hanyar wayar tarho, hana hatsarori da mawuyacin yanayi a kan hanyoyi ta hanyar sadarwa tsakanin inji. A ƙarin bayanin kula: Intanet ta hannu a ƙarshe za ta daina faɗuwa yayin taron jama'a, kamar wasa a filin wasa.

Majiyar hoto - Reuters, Niantic

2.2. Ƙwaƙwalwar Ƙarni na gaba

Anan muna magana ne game da ƙarni na biyar na RAM - DDR5. Samsung ya ba da sanarwar cewa samfuran tushen DDR2019 za su kasance a ƙarshen 5. Ana tsammanin sabon ƙwaƙwalwar za ta zama sau biyu da sauri kuma sau biyu yayin riƙe da sandunan da ke cikin tsari na 32GB don kwamfutarmu. A nan gaba, wannan zai zama musamman dacewa ga wayoyin hannu (sabon ƙwaƙwalwar ajiya za ta kasance a cikin ƙaramin ƙarfin lantarki) da kuma kwamfyutocin kwamfyutoci (inda adadin DIMM ya iyakance). Kuma koyon inji shima yana buƙatar RAM mai yawa.

2.3. Tsarin Tauraron Dan Adam na Ƙasa-Ƙasa-Orbit

Tunanin maye gurbin nauyi, tsada, tauraron dan adam masu ƙarfi tare da ɗimbin ƙanana da arha ba shi da nisa daga sabo kuma ya bayyana a cikin 90s. Game da me Elon Musk zai rarraba Intanet ga kowa da kowa daga tauraron dan adam. Yanzu malalaci ne kawai ba su ji ba. Kamfanin da ya fi shahara a nan shi ne Iridium, wanda ya yi fatara a ƙarshen 90s, amma ya sami ceto a kuɗin Ma'aikatar Tsaro ta Amurka (kada a dame shi da iRidium, tsarin gida mai wayo na Rasha). Aikin Elon Musk (Starlink) ya yi nisa da daya tilo - Richard Branson (OneWeb - 1440 tauraron dan adam da aka tsara), Boeing (sauraron tauraron dan adam 3000), Samsung (sauraron tauraron dan adam 4600), da sauransu suna shiga cikin tseren tauraron dan adam.

Yadda abubuwa ke tsayawa a wannan yanki, yadda tattalin arzikin ya kasance a can - karanta a ciki bita. Kuma muna jiran gwajin farko na waɗannan tsarin ta masu amfani na farko, wanda ya kamata ya faru a shekara mai zuwa.

2.4. Nanoscale 3D Printing

3D bugu, kodayake bai shiga rayuwar kowane mutum ba (a cikin sigar da masana'antar filastik ta gida ta yi alkawari), duk da haka ya daɗe da barin fasahar fasaha don geeks. Kuna iya yin hukunci da gaskiyar cewa kowane ɗan makaranta ya san game da kasancewar aƙalla alkalan zane-zane na 3D, kuma da yawa suna mafarkin siyan akwati tare da masu gudu da extruder don ... "kamar haka" (ko kuma sun riga sun saya).

Stereolithography ( firintocin 3D na laser) yana ba da damar bugu tare da hoto ɗaya: ana bincika sabbin polymers waɗanda kawai ke buƙatar photon biyu don ƙarfafawa. Wannan zai ba da damar, a cikin yanayin da ba na dakin gwaje-gwaje ba, don ƙirƙirar sabbin matattara gabaɗaya, tudu, maɓuɓɓugan ruwa, capillaries, ruwan tabarau da ... zaɓuɓɓukanku a cikin sharhi! Kuma a nan ba shi da nisa daga photopolymerization - kawai wannan fasaha yana ba mu damar "buga" na'urori masu sarrafawa da na'urorin kwamfuta. Bugu da kari, wannan ba ita ce shekarar farko da aka yi ba fasaha don buga graphene 500 nm sassa uku masu girma dabam, amma ba tare da m ci gaba.

Source

3. Hankali da Motsi

3.1. Matsayin Tuki Mai Ikon Kai 4 & 5

Don kada a ruɗe a cikin ƙamus, yana da daraja fahimtar menene matakan yancin kai (wanda aka ɗauka daga cikakkun bayanai). labarai, wanda muke mayar da duk masu sha'awar:

Mataki na 1: Sarrafa jirgin ruwa: yana taimaka wa direba a cikin ƙananan yanayi (misali, riƙe motar a wani gudun da aka ba da bayan direba ya ɗauke ƙafarsa daga feda)
Mataki na 2: Iyakantaccen tuƙi da taimakon birki. Dole ne direba ya kasance a shirye don ɗaukar iko kusan nan take. Hannunsa na kan sitiyarin, idanunsa na kan hanya. Wannan wani abu ne wanda Tesla da General Motors suka rigaya suna da su.
Mataki na 3: Direba ba dole ne ya ci gaba da kallon hanya ba. Amma dole ne ya kasance a faɗake kuma ya kasance a shirye don ɗaukar iko. Wannan wani abu ne da motocin da ake da su na kasuwanci ba su samu ba tukuna. Duk waɗanda suke a halin yanzu suna matakin 1-2.
Mataki na 4: Gaskiya autopilot, amma tare da ƙuntatawa: kawai tafiye-tafiye a cikin wani yanki da aka sani wanda aka tsara taswirar a hankali kuma an san shi da tsarin, kuma a ƙarƙashin wasu yanayi: misali, in babu dusar ƙanƙara. Waymo da General Motors suna da irin waɗannan samfuran, kuma suna shirin ƙaddamar da su a cikin birane da yawa da gwada su a cikin yanayi na gaske. Yandex yana da yankunan gwaji don taksi marasa matuka a Skolkovo da Innopolis: tafiya yana faruwa a ƙarƙashin kulawar injiniyan da ke zaune a wurin zama na fasinja; A karshen shekara, kamfanin na shirin fadada rundunarsa zuwa motocin marasa matuka 100.
Mataki na 5: Cikakken tuƙi ta atomatik, cikakken maye gurbin direba mai rai. Irin waɗannan tsarin ba su wanzu, kuma da wuya su bayyana a cikin shekaru masu zuwa.

Yaya gaskiya yake ganin duk wannan a nan gaba? Anan zan so in tura mai karatu zuwa labarin "Me yasa ba zai yiwu a ƙaddamar da robotaxi nan da 2020 ba, kamar yadda Tesla ya yi alkawari". Wannan wani bangare ne saboda rashin haɗin 5G: akwai saurin 4G bai isa ba. Wani bangare saboda tsadar motoci masu cin gashin kansu: har yanzu ba su sami riba ba, ƙirar kasuwancin ba ta da tabbas. A cikin kalma, "komai yana da rikitarwa" a nan, kuma ba daidai ba ne cewa Gartner ya rubuta cewa hasashen aiwatar da taro na matakan 4 da 5 ba a baya fiye da shekaru 10 ba.

3.2. 3D Sensing kyamarori

Shekaru takwas da suka gabata, Microsoft's Kinect mai kula da wasan caca ya yi raƙuman ruwa ta hanyar ba da mafita mai sauƙi kuma mara tsada ga hangen nesa na 3D. Tun daga wannan lokacin, ilimin motsa jiki da wasanni na raye-raye tare da Kinect sun ɗanɗana ɗan gajeren hawansu da raguwa, amma an fara amfani da kyamarori na 3D a cikin mutummutumi na masana'antu, motocin da ba su da hannu, da wayoyin hannu don ganewar fuska. Fasaha ta zama mai rahusa, ƙarami kuma ta fi dacewa.

Wayar Samsung S10 tana da kyamarar Time-of-Flight wanda ke auna nisa zuwa abu don sauƙaƙe mai da hankali. Source

Idan kuna sha'awar wannan batu, to, za mu tura ku zuwa cikakken cikakken bita na kyamarori masu zurfi: part 1, part 2.

3.3. Jirage masu saukar ungulu don isar da ƙananan kaya (Jirgin Canjin Kaya mai Haske)

A wannan shekara, Amazon ya yi taguwar ruwa lokacin da ya nuna wani sabon jirgin sama mara matuki a wasan kwaikwayon wanda zai iya ɗaukar ƙananan lodi har zuwa kilogiram 2. Ga birni mai cunkoson ababen hawa, wannan yana kama da mafita mai kyau. Bari mu ga yadda wadannan jirage marasa matuka ke yi nan gaba kadan. Wataƙila yana da kyau a yi shakka a hankali a nan: akwai matsaloli da yawa, farawa tare da yiwuwar sauƙin sata na jirgi mara matuki, kuma yana ƙarewa tare da hani na doka akan UAVs. Amazon Prime Air ya kasance tsawon shekaru shida amma har yanzu yana cikin lokacin gwaji.

Sabon drone na Amazon, wanda aka nuna wannan bazara. Akwai wani abu Star Wars game da shi. Source

Baya ga Amazon, akwai wasu 'yan wasa a wannan kasuwa (akwai cikakken bayani обзор), amma ba samfurin da aka gama ba: duk abin da ke cikin mataki na gwaji da yakin tallace-tallace. Na dabam, yana da daraja lura quite ban sha'awa sosai na musamman likita ayyuka a Afirka: isar da jinin da aka bayar a Ghana (kasuwa 14, kamfanin Zipline) da Rwanda (kamfanin Matternet).

3.4. Motoci Masu Gudun Hijira

Yana da wuya a ce wani abu tabbatacce a nan. A cewar Gartner, wannan ba zai bayyana ba a baya fiye da shekaru 10. Gabaɗaya, akwai matsaloli iri ɗaya a nan kamar a cikin motoci masu tuƙi, kawai suna samun sabon girma - a tsaye. Porsche, Boeing, da Uber sun sanar da burinsu na gina motar haya mai tashi sama.

3.5. Augmented Reality Cloud (AR Cloud)

Kwafin dijital na dindindin na ainihin duniya, yana ba ku damar ƙirƙirar sabon Layer na gaskiya gama gari ga duk masu amfani. A cikin ƙarin sharuɗɗan fasaha, muna magana ne game da buɗe dandamalin girgije wanda masu haɓakawa zasu iya haɗa aikace-aikacen su na AR. Samfurin samun kuɗi a bayyane yake; wani nau'in analog ne na Steam. Tunanin ya zama mai kaifi sosai har wasu sun yi imanin cewa AR ba tare da gajimare ba ne kawai mara amfani.

Yadda wannan zai iya kama a nan gaba ana nuna shi a cikin ɗan gajeren bidiyo. Yana kama da wani taron Black Mirror:

Hakanan zaka iya karantawa a labarin sake dubawa.

4. Adam Augmented

4.1. Emotion AI

Yadda za a auna, kwaikwaya da amsa motsin zuciyar ɗan adam? Wasu daga cikin abokan ciniki a nan kamfanoni ne waɗanda ke yin mataimakan murya kamar Amazon Alexa. Suna iya yin amfani da gidaje da gaske idan sun koyi fahimtar yanayin: fahimtar dalilin rashin gamsuwar mai amfani, kuma suyi ƙoƙarin gyara halin da ake ciki. Gabaɗaya, akwai ƙarin bayani a cikin mahallin fiye da saƙon kansa. Kuma mahallin shine yanayin fuska, yawan magana, da halin rashin magana.

Sauran aikace-aikace masu amfani: nazarin motsin zuciyarmu yayin ganawar aiki (dangane da tambayoyin bidiyo), tantance halayen tallace-tallace ko wasu abubuwan bidiyo (murmushi, dariya), taimako a cikin koyo (misali, don yin aiki mai zaman kanta a cikin fasahar magana da jama'a).

Yana da wuya a yi magana mafi kyau akan wannan batu fiye da marubucin gajeren fim na minti 6 Satar Ur Ji. Bidiyon mai hankali da salo yana nuna yadda zaku iya auna motsin zuciyarmu don dalilai na talla, kuma daga halayen ɗan lokaci na fuskar ku, gano ko kuna son pizza, karnuka, Kanye West, har ma menene matakin samun ku da kuma ƙimar IQ. Ta ziyartar gidan yanar gizon fim ɗin ta amfani da hanyar haɗin da ke sama, za ku zama ɗan takara a cikin bidiyo mai mu'amala ta amfani da ginanniyar kyamarar kwamfutar tafi-da-gidanka. An riga an nuna fim ɗin a bukukuwan fina-finai da yawa.

Source

Akwai ma irin wannan binciken mai ban sha'awa: yadda ake gane zagi a cikin rubutu. Mun dauki tweets tare da hashtag #sarcasm kuma mun yi tsarin horo na tweets 25 tare da ba'a da tweets na yau da kullun 000 game da duk abin da ke ƙarƙashin rana. Mun yi amfani da ɗakin karatu na TensorFlow, mun horar da tsarin, kuma ga sakamakon:

Source

Saboda haka, yanzu, idan ba ku da tabbas game da abokin aikinku ko aboki - ya ce wani abu a gare ku da gaske ko ba'a, za ku iya rigaya amfani. horar da jijiyoyi cibiyar sadarwa!

4.2. Haɓaka Hankali

Aiwatar da aikin hankali ta atomatik ta amfani da hanyoyin koyon injin. Zai zama kamar ba sabon abu ba? Amma kalmar da kanta tana da mahimmanci a nan, musamman ma da yake ta zo daidai a takaice tare da Artificial Intelligence. Wannan ya dawo da mu ga muhawara game da "ƙarfi" da "rauni" AI.
Strong AI shine basirar wucin gadi ɗaya daga fina-finan almara na kimiyya wanda ya yi daidai da tunanin ɗan adam kuma yana sane da kansa a matsayin mutum. Wannan bai wanzu ba tukuna kuma babu tabbas ko zai wanzu kwata-kwata.

Mai rauni AI ba mutum ne mai zaman kansa ba, amma mataimaki na ɗan adam. Ba ya da'awar cewa yana da tunani irin na ɗan adam, amma kawai ya san yadda za a warware matsalolin bayanai, misali, ƙayyade abin da aka nuna a hoto ko fassara rubutu.

Source

A wannan ma'anar, da ke da hankali na Maris "mai rauni Ai" a cikin tsari mai tsabta, kuma da alama ba ta gabatar da rikice-rikice ba "ko da yake da kowa da kowa da kowa" wanda kowa yayi mafarki tuna da yawa tattaunawa game da "motocin tawaye"). Yin amfani da kalmar Augmented Intelligence, nan da nan muka zama jarumawa na wani fim: daga almara kimiyya (kamar Asimov's "I, Robot") mun sami kanmu a cikin cyberpunk ("augmentations" a cikin wannan nau'in duk nau'in implants ne wanda ke fadada ikon ɗan adam).

Yadda ya ce Erik Brynjolfsson da Andrew McAfee: “A cikin shekaru 10 masu zuwa, abin da zai faru ke nan. AI ba zai maye gurbin manajoji ba, amma waɗancan manajojin da ke amfani da AI za su maye gurbin waɗanda ba su yi hakan ba tukuna. ”

misalai:

• Magunguna: Jami'ar Stanford ta haɓaka da algorithm, wanda ke jure wa aikin gane cututtukan cututtuka a kan rayukan X-ray a matsakaici kamar yadda yawancin likitoci suka yi nasara.
• Ilimi: taimako ga ɗalibai da malamai, nazarin martanin ɗalibi ga kayan, gina yanayin koyo na mutum ɗaya.
• Nazarin kasuwanci: ƙaddamar da bayanai, bisa ga ƙididdiga, yana ɗaukar 80% na lokacin mai bincike, kuma kawai 20% na gwajin kanta.

4.3. Biochips

Wannan shine jigon da aka fi so na duk fina-finai da littattafai na cyberpunk. Gabaɗaya, microchipping dabbobi ba sabon abu bane. Amma yanzu an fara dasa waɗannan kwakwalwan kwamfuta a cikin mutane.

A cikin wannan yanayin, ana iya danganta ƙarar da wani lamari mai ban sha'awa a cikin Kasuwar Square Uku na Amurka. A can, ma'aikaci ya fara ba da kyauta don dasa kwakwalwan kwamfuta a ƙarƙashin fata don musanya kuɗi. Guntu tana ba ku damar buɗe kofofin, shiga cikin kwamfutoci, siyan kayan ciye-ciye daga injin siyarwa - wato, irin wannan katin ma'aikaci na duniya. Bugu da ƙari, irin wannan guntu yana aiki daidai azaman katin shaida; ba shi da tsarin GPS, don haka ba zai yuwu a iya bin diddigin duk wanda ke amfani da shi ba. Kuma idan mutum yana son cire guntu daga hannunsa, yana ɗaukar mintuna 5 tare da taimakon likita.

Ana dasa kwakwalwan kwamfuta yawanci tsakanin babban yatsan yatsa da yatsa. Source

Kara karantawa labarin game da halin da ake ciki tare da chipping a duniya.

4.4. Wurin Aiki Mai Immersive

"Immersive" wata sabuwar kalma ce wacce kawai ba ta da kuɓuta. Yana ko'ina. Gidan wasan kwaikwayo mai zurfi, nuni, silima. Me kuke nufi? Immersion shine ƙirƙirar tasiri mai zurfi, lokacin da iyaka tsakanin marubucin da mai kallo, da kama-da-wane da ainihin duniya ta ɓace. A wurin aiki, mai yiwuwa, wannan yana nufin ɓata layin tsakanin mai yin da mai farawa da ƙarfafa ma'aikata don ɗaukar matsayi mai ƙarfi ta hanyar sake fasalin yanayin su.

Tun da yanzu muna da Agile, sassauci, da haɗin gwiwa a ko'ina, wuraren aiki ya kamata su kasance masu sauƙin daidaitawa kamar yadda zai yiwu kuma ya kamata ya karfafa aikin rukuni. Tattalin Arziki yana ƙayyadad da sharuɗɗansa: akwai ƙarin ma'aikata na wucin gadi, farashin hayar ofis yana ƙaruwa, kuma a cikin kasuwa mai fa'ida, kamfanonin IT suna ƙoƙarin haɓaka gamsuwar ma'aikata daga aiki ta hanyar ƙirƙirar wuraren shakatawa da sauran fa'idodi. Kuma duk wannan yana nunawa a cikin zane na wuraren aiki.

Daga rahoto sani

4.5. Keɓantawa

Kowa ya san menene keɓancewa a cikin talla. Wannan shi ne lokacin da a yau kuke tattaunawa da abokin aikinku cewa iskar da ke cikin dakin ta dan bushe, kuma ya kamata ku sayi na'urar humidifier don ofis, kuma washegari za ku ga wani talla a dandalin sadarwar ku - "sayi mai humidifier" (a hakikanin lamarin da ya faru da ni).

Source

Keɓancewa, kamar yadda Gartner ya ayyana, amsa ce ga haɓakar damuwa na masu amfani game da amfani da bayanan keɓaɓɓen su don dalilai na talla. Manufar ita ce samar da hanyar da za a nuna mana tallace-tallace wanda ya dace da yanayin da muka sami kanmu, ba mu kanmu ba. Misali, wurinmu, nau'in na'urar, lokacin rana, yanayin yanayi - wannan wani abu ne da ba ya keta bayanan sirrinmu, kuma ba ma jin daɗin jin daɗin “sa ido”.

Karanta game da bambanci tsakanin waɗannan ra'ayoyin biyu bayanin kula Andrew Frank blogs akan gidan yanar gizon Gartner. Akwai irin wannan bambance-bambancen da hankali da irin waɗannan kalmomi waɗanda ku, ba ku san bambancin ba, kuna hadarin yin jayayya na dogon lokaci tare da mai magana da ku, ba tare da zargin cewa, a gaba ɗaya, duka biyu daidai ne (kuma wannan ma wani lamari ne na gaske wanda ya faru da marubuci).

4.6. Biotech – Al'ada ko Artificial Tissue

Wannan shi ne, da farko, ra'ayin girma naman wucin gadi. A lokaci guda, ƙungiyoyi da yawa a duniya suna shagaltuwa da haɓaka dakin gwaje-gwaje "Nama 2.0" - ana tsammanin zai zama mai rahusa fiye da yadda aka saba, kuma abinci mai sauri sannan manyan kantunan zasu canza zuwa gare ta. Masu saka hannun jari a wannan fasaha sun hada da Bill Gates, Sergey Brin, Richard Branson da sauransu.

Source

Dalilan da yasa kowa ke sha'awar naman wucin gadi:

1. Dumamar duniya: iskar methane daga gonaki. Wannan shine kashi 18% na yawan iskar gas da ke shafar yanayin duniya.
2. Girman yawan jama'a. Bukatar nama yana girma, kuma ba zai yiwu a ciyar da kowa da kowa tare da nama na halitta ba - yana da tsada kawai.
3. Rashin sarari. An riga an yanke kashi 70% na gandun daji na Amazon don kiwo.
4. La'akari da da'a. Akwai wadanda wannan yana da mahimmanci a gare su. Tuni dai kungiyar kare hakkin dabbobi ta PETA ta bayar da kyautar dala miliyan daya ga masanin kimiyyar da ke kawo naman kajin da aka sanya a kasuwa.

Maye gurbin nama na gaske tare da waken soya wani yanki ne na bayani, saboda mutane na iya fahimtar bambancin dandano da laushi, kuma da wuya su daina naman nama don neman waken soya. Don haka kuna buƙatar nama na gaske, wanda aka shuka. Yanzu, rashin alheri, naman wucin gadi yana da tsada sosai: daga $ 12 a kowace kilogram. Wannan shi ne saboda tsarin fasaha mai rikitarwa na shuka irin wannan nama. Karanta game da shi duka labarin.

Idan muka yi magana game da wasu lokuta na girma nama - riga a cikin magani - to batun tare da gabobin wucin gadi yana da ban sha'awa: alal misali, "patch" don tsokar zuciya. buga firintar 3D na musamman. Sanin labaru kamar zuciyar linzamin kwamfuta da aka girka ta wucin gadi, amma gabaɗaya komai yana cikin iyakokin gwajin asibiti. Don haka da wuya mu ga Frankenstein a cikin shekaru masu zuwa.

Anan Gartner yayi taka tsantsan a cikin kiyasin nasa, da alama yana tunawa da hasashensa na 2015 da ya gaza cewa a shekarar 2019, kashi 10% na al'ummar kasashen da suka ci gaba za su sami na'urar buga lafiya ta 3D. Saboda haka, yana nufin lokacin da za a kai ga tudu mai yawan aiki shine aƙalla shekaru 10.

5. Digital Ecosystems

5.1. Yanar Gizo Mai Rarraba

Wannan ra'ayi yana da alaƙa da sunan wanda ya ƙirƙira gidan yanar gizo, wanda ya lashe lambar yabo ta Turing Sir Tim Burners-Lee. A gare shi, tambayoyi na ɗabi'a a kimiyyar kwamfuta koyaushe suna da mahimmanci kuma ainihin haɗin Intanet yana da mahimmanci: aza harsashi na hypertext, ya gamsu cewa hanyar sadarwa yakamata ta yi aiki kamar yanar gizo, ba kamar matsayi ba. Wannan lamarin ya kasance a farkon matakin ci gaban cibiyar sadarwa. Duk da haka, yayin da Intanet ke girma, tsarinsa ya zama tsakiya saboda dalilai daban-daban. Ya bayyana cewa ana iya toshe damar shiga hanyar sadarwar ga ƙasa baki ɗaya cikin sauƙi tare da taimakon ƴan masu samarwa. Kuma bayanan masu amfani sun zama tushen wutar lantarki da samun kudin shiga ga kamfanonin Intanet.

"An riga an lalata Intanet," in ji Burners-Lee. “Matsalar ita ce injin bincike guda ɗaya, babban hanyar sadarwar zamantakewa ɗaya, dandamali ɗaya na microblogging ya mamaye. Ba mu da matsalolin fasaha, amma muna da na zamantakewa. "

A cikin nasa budaddiyar wasika Domin cika shekaru 30 da kafa yanar gizo ta duniya, mahaliccin gidan yanar gizon ya zayyana manyan matsalolin Intanet guda uku:

1. Lalacewar da aka yi niyya kamar hacking na gwamnati, aikata laifuka, da cin zarafi akan layi
2. Tsarin tsarin tsarin, wanda, don lalata mai amfani, ya haifar da ƙasa don irin waɗannan hanyoyin kamar: abubuwan ƙarfafa kuɗi don dannawa da yada kwayar cutar ta bayanan karya.
3. Sakamakon da ba a yi niyya ba na ƙirar tsarin da ke haifar da rikici da rage ingancin tattaunawar kan layi

Kuma Tim Berners-Lee ya riga ya ba da amsa kan waɗanne ka'idoji "Intanet na Mutumin Lafiya" zai iya dogara da shi, ba tare da lambar matsala ta 2 ba: "Ga masu amfani da yawa, samun kudin shiga tallar ya kasance kawai samfurin yin hulɗa da Intanet. Ko da mutane suna jin tsoro game da abin da ke faruwa da bayanan su, suna shirye su yi yarjejeniya tare da na'urar talla don samun damar karɓar abun ciki kyauta. Ka yi tunanin duniyar da biyan kaya da ayyuka ke da sauƙi kuma mai daɗi ga ɓangarorin biyu. ” Daga cikin zaɓuɓɓukan yadda za a iya tsara wannan: mawaƙa za su iya sayar da rikodin su ba tare da masu shiga tsakani a cikin nau'in iTunes ba, kuma shafukan labarai na iya amfani da tsarin biyan kuɗi na micropayment don karanta labarin ɗaya, maimakon samun kuɗi daga talla.

A matsayin samfur na gwaji don wannan sabuwar Intanet, Tim Berners-Lee ya ƙaddamar da aikin SOLID, ainihin abin da ke cikin shi shine adana bayanan ku a cikin "pod" - ma'ajiyar bayanai, kuma yana iya samar da wannan bayanan ga aikace-aikacen ɓangare na uku. Amma bisa ka'ida, ku da kanku ne masu sarrafa bayanan ku. Duk waɗannan suna da alaƙa ta kut-da-kut da ra'ayin hanyoyin sadarwa na abokan gaba, wato, kwamfutarka ba kawai buƙatar sabis ba, har ma tana ba da su, don kada ku dogara ga uwar garken guda ɗaya a matsayin tashar guda ɗaya.

Source

5.2. Ƙungiyoyi masu cin gashin kai

Kungiya ce da ake tafiyar da ita bisa ka’idojin da aka rubuta ta hanyar manhajar kwamfuta. Ayyukanta na kuɗi sun dogara ne akan blockchain. Manufar kafa irin wadannan kungiyoyi shi ne don kawar da jihar daga matsayin masu shiga tsakani da samar da yanayi na aminci ga abokan zamanta, wanda ba na kowa ba ne, na kowa ne tare. Wato, a ka'idar, wannan ya kamata, idan ra'ayin ya samo asali, a soke notaries da sauran cibiyoyin tabbatarwa da aka saba.

Misalin da ya fi shahara a irin wannan kungiya shi ne, The DAO mai mayar da hankali kan harkokin kasuwanci, wanda ya tara dala miliyan 2016 a shekarar 150, inda nan da nan aka sace dala 50 ta hanyar wani rami na doka. Wani mawuyaci mai wahala ya taso nan da nan: ko dai a mirgina a mayar da kuɗin, ko kuma a yarda cewa cire kuɗin doka ne, domin ba ta wata hanya ta keta ka'idodin dandamali. A sakamakon haka, don dawo da kuɗi ga masu zuba jari, masu yin halitta dole ne su lalata DAO, sake rubuta blockchain da kuma karya ka'idodinsa - rashin canzawa.

Comic game da Ethereum (hagu) da DAO (dama). Source

Wannan duka labarin ya ɓata sunan ainihin ra'ayin DAO. Wannan aikin da aka yi a kan tushen da Ethereum cryptocurrency, version Ether 2.0 ana sa ran shekara mai zuwa - watakila mawallafa (ciki har da sanannen Vitalik Buterin) za su yi la'akari da kurakurai da kuma nuna wani sabon abu. Wataƙila dalilin da ya sa Gartner ya sanya DAO akan layi.

5.3.Synthetics Data

Don horar da hanyoyin sadarwar jijiyoyi, ana buƙatar adadi mai yawa na bayanai. Lakabi bayanai da hannu babban aiki ne wanda mutum kaɗai zai iya yi. Saboda haka, yana yiwuwa a ƙirƙira bayanan wucin gadi. Misali, tarin tarin fuskokin mutane akan shafin https://generated.photos. An ƙirƙira su ta amfani da GAN - algorithms waɗanda aka riga aka ambata a sama.

Wadannan fuskokin ba na mutane ba ne. Source

Babban fa'idar irin waɗannan bayanan shine cewa babu matsalolin doka a cikin amfani da su: babu wanda zai ba da izini ga sarrafa bayanan sirri.

5.4. Digital Ops

Suffix "Ops" ya zama abin ban sha'awa mai ban sha'awa tun lokacin da DevOps ya sami tushe a cikin jawabinmu. Yanzu game da abin da DigitalOps yake - kawai ƙaddamarwa ne na DevOps, DesignOps, MarketingOps ... Shin kun gaji tukuna? A takaice, shi ne canja wurin tsarin DevOps daga yankin software zuwa duk sauran bangarorin kasuwanci - tallace-tallace, ƙira, da dai sauransu.

Source

Manufar DevOps ita ce kawar da shinge tsakanin Ci gaban kanta da Ayyuka (hanyoyin kasuwanci), ta hanyar ƙirƙirar ƙungiyoyi na gama gari, inda akwai masu shirye-shirye, masu gwadawa, ƙwararrun tsaro, da masu gudanarwa; aiwatar da wasu ayyuka: ci gaba da haɗin kai, abubuwan more rayuwa kamar lamba, raguwa da ƙarfafa sarƙoƙin amsawa. Manufar ita ce a hanzarta samfurin zuwa kasuwa. Idan kuna tunanin wannan yayi kama da Agile, kun yi gaskiya. Yanzu a hankali canza wannan hanyar daga fagen haɓaka software zuwa haɓaka gabaɗaya - kuma kun fahimci menene DigitalOps.

5.5. Hotunan Ilimi

Hanyar software don ƙirar yanki na ilimi, gami da yin amfani da algorithms na koyon inji. An gina jadawali na ilimi a saman bayanan da ake da su don haɗa dukkan bayanai: duka biyun da aka tsara (jerin abubuwan da suka faru ko mutane) da waɗanda ba a tsara su ba (rubutun labarin).

Misali mafi sauki shine katin da zaku iya gani a cikin sakamakon binciken Google. Idan kana neman mutum ko cibiyar, za ka ga kati a hannun dama:


Lura cewa "Abubuwan da ke zuwa" ba kwafin bayanai ba ne daga Taswirorin Google ba, amma haɗin jadawali tare da Yandex.Afisha: zaku iya ganin wannan cikin sauƙi idan kun danna abubuwan da suka faru. Wato shine haɗewar tushen bayanai da yawa tare.

Idan ka nemi jerin sunayen - alal misali, "Shahararrun daraktoci" - za a nuna maka carousel:

Kyauta ga waɗanda suka karanta har ƙarshe

Kuma yanzu da muka bayyana wa kanmu ma'anar kowane maki, za mu iya kallon wannan hoto, amma a cikin Rashanci:

Raba shi kyauta akan cibiyoyin sadarwar jama'a!

Tatyana Volkova - Mawallafin shirin horarwa don Intanet na Abubuwan IT hanya a Samsung Academy, ƙwararre a cikin shirye-shiryen alhakin zamantakewa a Cibiyar Bincike ta Samsung.


source: www.habr.com