Wasannin bidiyo na ci gaba da ba da ƙalubale ga biliyoyin 'yan wasa a duniya. Wataƙila ba ku sani ba tukuna, amma algorithms na koyon injin sun fara haɓaka ƙalubalen su ma.
A halin yanzu akwai adadi mai yawa na bincike a fagen AI don ganin ko ana iya amfani da hanyoyin koyo na na'ura akan wasannin bidiyo. Babban ci gaba a wannan fanni ya nuna haka injin inji ana iya amfani da wakilai don yin koyi ko ma maye gurbin ɗan wasan ɗan adam.
Menene wannan ke nufi ga nan gaba na wasanin bidiyo?
Shin waɗannan ayyukan ne kawai don nishaɗi, ko akwai dalilai masu zurfi da ya sa yawancin masu bincike ke mai da hankali kan wasanni?
Wannan labarin zai ɗan bincika tarihin AI a cikin wasannin bidiyo. Bayan haka, za mu ba ku taƙaitaccen bayani game da wasu dabarun koyon injin da za mu iya amfani da su don koyon yadda ake doke wasanni. Zamu duba wasu aikace-aikace masu nasara na ragamar jijiyoyi don koyo da ƙware takamaiman wasannin bidiyo.
Takaitaccen Tarihin AI a cikin Wasanni
Kafin mu shiga dalilin da yasa gidajen yanar gizo suka zama mafi kyawun algorithm don magance wasannin bidiyo, bari mu ɗan bincika yadda masana kimiyyar kwamfuta suka yi amfani da wasannin bidiyo don haɓaka binciken su a AI.
Kuna iya jayayya cewa, daga farkonsa, wasanni na bidiyo sun kasance yanki mai zafi na bincike ga masu bincike masu sha'awar AI.
Duk da yake ba ainihin wasan bidiyo bane asalinsa, chess ya kasance babban fifiko a farkon zamanin AI. A cikin 1951, Dokta Dietrich Prinz ya rubuta shirin wasan dara ta amfani da kwamfuta mai lamba Ferranti Mark 1. Wannan ya kasance a baya a zamanin lokacin da waɗannan manyan kwamfutoci dole ne su karanta shirye-shiryen daga tef ɗin takarda.
Shirin da kansa ba cikakken dara ba AI. Saboda gazawar kwamfuta, Prinz zai iya ƙirƙirar shirin ne kawai wanda zai magance matsalolin darasi-da-biyu. A matsakaita, shirin ya ɗauki mintuna 15-20 don ƙididdige kowane motsi mai yuwuwa ga 'yan wasan Fari da Baƙar fata.
Aiki kan inganta dara dara da masu duba AI ya inganta a hankali cikin shekarun da suka gabata. Ci gaban ya kai kololuwa a cikin 1997 lokacin da IBM's Deep Blue ya doke babban malamin Ches na Rasha Garry Kasparov a wasanni biyu na wasanni shida. A zamanin yau, injunan dara da zaku iya samu akan wayar hannu zasu iya cin nasara akan Deep Blue.
Abokan adawar AI sun fara samun shahara a lokacin zinare na wasannin arcade na bidiyo. 1978's Space Invaders da 1980s Pac-Man wasu daga cikin majagaba na masana'antu wajen ƙirƙirar AI wanda zai iya ƙalubalantar ƙalubalen har ma da mafi yawan tsoffin 'yan wasa.
Pac-Man, musamman, sanannen wasa ne ga masu binciken AI don yin gwaji. Daban-daban gasa don Ms. Pac-Man an shirya don sanin ko wane kungiya za ta iya fito da mafi kyawun AI don doke wasan.
Game AI da heuristic algorithms sun ci gaba da haɓakawa yayin da ake buƙatar abokan adawar wayo. Misali, yaƙi AI ya tashi cikin shahara kamar yadda nau'ikan nau'ikan nau'ikan masu harbi na farko suka zama mafi al'ada.
Koyon Inji a Wasan Bidiyo
Yayin da fasahohin koyon na'ura suka tashi da sauri cikin shahara, ayyukan bincike daban-daban sun yi ƙoƙarin amfani da waɗannan sabbin dabaru don yin wasannin bidiyo.
Wasanni kamar Dota 2, StarCraft, da Doom na iya aiki azaman matsaloli ga waɗannan mashin ilmin lissafi a warware. Algorithms na ilmantarwa mai zurfi, musamman, sun sami damar cimmawa har ma sun zarce aikin matakin ɗan adam.
The Muhallin Koyon Arcade ko ALE ya ba masu bincike hanyar sadarwa sama da ɗari Atari 2600 wasanni. Dandalin bude tushen ya baiwa masu bincike damar tantance aikin dabarun koyan injin akan wasannin bidiyo na Atari na gargajiya. Google ma ya buga nasu takarda ta amfani da wasanni bakwai daga ALE
A halin yanzu, ayyuka kamar VizDoom ya ba masu binciken AI damar horar da na'ura algorithms don kunna 3D masu harbi na farko.
Yadda Ake Aiki: Wasu Mahimman Ka'idoji
Networks
Yawancin hanyoyin magance wasannin bidiyo tare da koyon inji sun haɗa da nau'in algorithm da aka sani da hanyar sadarwa ta jijiya.
Kuna iya tunanin hanyar yanar gizo azaman shirin da ke ƙoƙarin yin kwaikwayi yadda kwakwalwa zata iya aiki. Hakazalika yadda kwakwalwarmu ta ƙunshi jijiyoyi masu watsa sigina, net ɗin jijiyoyi kuma yana ɗauke da ƙananan ƙwayoyin cuta.
Waɗannan jijiyoyi na wucin gadi kuma suna aika sigina zuwa juna, tare da kowace sigina ta zama ainihin lamba. Gidan yanar gizo yana ƙunshe da yadudduka da yawa tsakanin shigarwar da matakan fitarwa, wanda ake kira cibiyar sadarwa mai zurfi.
Sanin karantarwa
Wata dabarar koyon inji ta gama gari wacce ta dace da koyan wasannin bidiyo shine ra'ayin ƙarfafa koyo.
Wannan dabara ita ce tsarin horar da wakili ta hanyar amfani da lada ko hukunci. Da wannan hanya, wakili ya kamata ya iya samar da mafita ga matsala ta hanyar gwaji da kuskure.
Bari mu ce muna son AI don gano yadda ake buga wasan Snake. Manufar wasan mai sauƙi ce: sami maki da yawa gwargwadon yuwuwa ta hanyar cinye abubuwa da guje wa girma wutsiya.
Tare da ƙarfafa koyo, za mu iya ayyana aikin lada R. Aikin yana ƙara maki lokacin da maciji ya cinye abu kuma yana cire maki lokacin da Maciji ya sami matsala. Ganin yanayin da ake ciki na yanzu da jerin ayyuka masu yuwuwa, ƙirar ƙarfafawar mu za ta yi ƙoƙarin ƙididdige mafi kyawun 'manufofin' waɗanda ke haɓaka aikin lada.
Neuroevolution
Tsayawa cikin jigo tare da samun wahayi ta yanayi, masu bincike sun kuma sami nasara wajen amfani da ML zuwa wasannin bidiyo ta hanyar da aka sani da neuroevolution.
Maimakon amfani saukowa gradient don sabunta neurons a cikin hanyar sadarwa, zamu iya amfani da algorithms na juyin halitta don samun sakamako mafi kyau.
Algorithms na juyin halitta yawanci suna farawa ta hanyar samar da farkon yawan mutane bazuwar. Sannan muna kimanta waɗannan mutane ta amfani da wasu sharuɗɗa. An zaɓi mafi kyawun mutane a matsayin "iyaye" kuma an haɗa su tare don samar da sababbin mutane na mutane. Wadannan mutane za su maye gurbin mafi ƙarancin mutane a cikin jama'a.
Waɗannan algorithms kuma yawanci suna gabatar da wani nau'i na aiki na maye gurbi yayin tsallake-tsallake ko matakin “kiwo” don kula da bambancin jinsi.
Samfuran Bincike akan Koyan Injin a Wasannin Bidiyo
BudeAI Five
BudeAI Five shiri ne na kwamfuta ta OpenAI wanda ke da nufin kunna DOTA 2, sanannen wasan fagen fama na wayar hannu (MOBA).
Shirin ya yi amfani da dabarun koyo na ƙarfafawa, wanda aka ƙaddara don koyo daga miliyoyin firam a sakan daya. Godiya ga tsarin horarwa da aka rarraba, OpenAI ya sami damar buga wasanni masu darajar shekaru 180 kowace rana.
Bayan lokacin horo, OpenAI Five ya sami damar cimma aikin matakin ƙwararru da nuna haɗin gwiwa tare da 'yan wasan ɗan adam. A cikin 2019, OpenAI biyar ya sami damar shan kashi 99.4% na 'yan wasa a wasannin jama'a.
Me yasa OpenAI ta yanke shawarar wannan wasan? A cewar masu binciken, DOTA 2 yana da injiniyoyi masu rikitarwa waɗanda ba sa iya samun zurfin da ke akwai ƙarfafa ilmantarwa algorithms.
Super Mario Bros.
Wani aikace-aikacen mai ban sha'awa na gidajen yanar gizo a cikin wasannin bidiyo shine amfani da neuroevolution don kunna dandamali kamar Super Mario Bros.
Misali, wannan shiga hackathon yana farawa da rashin sanin wasan kuma sannu a hankali yana gina tushen abin da ake buƙata don ci gaba ta matakin.
Gidan yanar gizo mai haɓaka kansa yana ɗaukar yanayin wasan a halin yanzu azaman grid na tayal. Da farko, gidan yanar gizo ba shi da fahimtar abin da kowane tayal ke nufi, kawai cewa fale-falen "iska" sun bambanta da "fale-falen ƙasa" da "fale-falen abokan gaba."
Aiwatar da aikin hackathon na juyin halittar neuroevolution yayi amfani da tsarin kwayoyin halitta NEAT don ƙirƙira tarun jijiyoyi daban-daban zaɓe.
Muhimmanci
Yanzu da kuka ga wasu misalan gidajen yanar gizo na wasan bidiyo, kuna iya yin mamakin menene ma'anar wannan duka.
Tunda wasannin bidiyo sun ƙunshi hadaddun hulɗa tsakanin wakilai da mahallin su, shine cikakkiyar filin gwaji don yin AI. Muhalli na zahiri amintattu ne kuma ana iya sarrafawa kuma suna ba da wadataccen bayanai mara iyaka.
Binciken da aka yi a wannan fanni ya baiwa masu bincike fahimtar yadda za a iya inganta gidajen yanar gizo don koyan yadda ake magance matsaloli a duniyar gaske.
Cibiyoyin Neural an yi wahayi zuwa ga yadda kwakwalwa ke aiki a cikin duniyar halitta. Ta hanyar nazarin yadda ƙwayoyin jijiya na wucin gadi ke yin hali yayin koyon yadda ake yin wasan bidiyo, za mu iya samun haske game da yadda kwakwalwa na mutum aiki.
Kammalawa
Kamanceceniya tsakanin hanyoyin sadarwa na jijiyoyi da kwakwalwa sun haifar da fahimta a bangarorin biyu. Ci gaba da bincike kan yadda gidajen yanar gizo za su iya magance matsalolin wata rana na iya haifar da ƙarin ci-gaba nau'ikan wucin gadi hankali.
Ka yi tunanin yin amfani da AI wanda aka keɓance da ƙayyadaddun bayanai naka wanda zai iya buga wasan bidiyo gaba ɗaya kafin ka saya don sanar da kai ko ya cancanci lokacinka. Shin kamfanonin wasan bidiyo za su yi amfani da gidajen yanar gizo don inganta ƙirar wasan, matakin tweak, da wahalar abokan hamayya?
Me kuke tunanin zai faru lokacin da gidajen yanar gizo suka zama manyan yan wasa?
Leave a Reply