Ana sabunta kullun1797 ingantattun labarai
Binciken tushen AI na ƙaddamar da samfur, sauye-sauyen manufofi, bincike na aminci, da motsin masana'antu, wanda ƙungiyar ilimi mai zaman kanta ta yi bayani a cikin Ingilishi a sarari.
Kowane labari yana da alaƙa da mafi ƙaƙƙarfan shaidar da ake samu: tushe na asali idan akwai, in ba haka ba a bayyane yake ba da rahoto.
Abin da ya faru, dalilin da yasa yake da mahimmanci, da abin da za a kalli - ba tare da jargon ba.
Lokacin da siginar ta yi bakin ciki, ba mu buga komai ba sai fakitin abinci.
Labarun AI da aka bincika tushen tushe, sabbin farko, don mutanen da ke buƙatar fahimtar AI ba tare da neman haɓaka ba.
Google DeepMind ta ce tana fadada binciken wasanninta ta hanyar haɗin gwiwa tare da Fenris Creations, ɗakin studio bayan duniyar EVE. Shirin zai gwada ƙwaƙwalwar ma'aikata, ci gaba da koyo, tsare-tsare na dogon lokaci da halayyar wakilai da yawa, farawa da yanayin EVE Online na layi.
Tabbatar da tushe: deepmind.googleAikin Gilashin da ke Kusa yana amfani da bayanan Bluetooth don tuta wasu na'urorin gilashin wayo akan Android da iOS. Takardun nasa ya ce tsarin na iya samar da abubuwan karya, rasa na'urori kuma bai kamata a yi amfani da shi don fuskantar mutane ba.
Google Barazana Intelligence Group ya ce abokan gaba suna amfani da AI don haɓaka bincike mai rauni, gujewa ganowa, sarrafa ayyukan malware da kai hari kan sarkar samar da software a kusa da tsarin AI.
cloud.google.comWani jigon arXiv ya ba da rahoton cewa ƙirar harshe na iya bayyana mahimman bayanai da aka sanya a cikin mahallin, koda bayan ƙin buƙatun cirewa kai tsaye. A cikin gwaje-gwajen da aka sarrafa, marubutan sun ce hare-haren daidaitawa sun dawo da gajerun sirri daga abubuwan samarwa na yau da kullun kuma sun fitar da cikakkun lambobin Tsaron Zamantakewa daga wakilin salon samarwa.
Wani sabon takarda na arXiv yana gabatar da Thinkingbox, akwatin yashi da ma'auni don gwada wakilan AI akan ayyukan kasuwanci masu yawa. Samfurin da aka gwada mafi ƙarfi ya sami 65.36% wucewa @ maki 1 amma kawai 25.25% wucewa^20 maki, yana nuna tazara tsakanin nasara lokaci-lokaci da aiwatar da abin dogaro.
Wani bugu na arXiv yana kimanta 3 LLMs a cikin hanyoyin matsawa guda 11 yana ba da rahoton cewa jimlar daidaito da ruɗani na iya ɓoye asarar ilimi mara daidaituwa, wuce gona da iri akan sabbin bayanan da suka ɓace, da kuma daidaita canje-canje a cikin son rai na rukuni.
Rubutun arXiv ya ba da rahoton cewa rubutun da bai dace da ɗawainiya akai-akai yana nuna son kai ga ƙirar harsunan multimodal a cikin hukunce-hukuncen gani, kuma yana bayyana tasirin a matsayin maƙasudin ma'auni a gefen yanke shawara.
Fitilar arXiv tana gano rashin daidaituwa tsakanin ma'aunin nauyi na cibiyar sadarwa na jijiyoyi a matsayin mai yuwuwar sanadin ɓata daidaitawa, sannan ya ba da shawarar yanke waɗannan dabi'u lokaci-lokaci a ci gaba da haɓaka koyo.
Wani sabon bugu na arXiv yana gabatar da bincike don yanke shawara lokacin da takamaiman yanki na ƙirar koma baya zai iya ƙetare ƙayyadaddun gauraya. A cikin gwaje-gwajen da marubutan suka siffanta, ƙaramin bincike-yanki mai lakabi ya annabta riba kuma ya taimaka ƙin turawa ɗaya wanda ya haifar da asarar fiye da sau 30.
arXiv preprint ya ba da rahoton shaida cewa Llama 3.1 8B a ciki yana bin bambance-bambancen farko a cikin jeri na ƙira na musamman. Binciken yana ba da tsarin da aka tsara don wannan ɗabi'a, amma iyakarsa da ƙarfinsa ba su da tabbas daga abin da aka kawo.
Rubutun da aka karɓa a MLHC 2026 yana kwatanta PAFIR, tsarin ƙarfafawa-koyo wanda ke zaɓar canzawa, takamaiman siginar haɗarin faɗuwa na mutum daga maimaita ma'aunin lafiya na multimodal. Takardar ta ba da rahoton mafi kyawun kamanni fiye da tushe, amma rikodin da aka kawo ba ya yin tasiri mai girma ko shaidar raguwar faɗuwa.
Wani samfurin arXiv ya bayyana DCGCNet, samfurin da ya haɗu da sake ginawa na ECG da rarrabuwa na atrial-fibrillation, yana ba da rahoton AUC sama da 0.98 a cikin saitunan saitin bayanan giciye guda bakwai da juriya ga nau'ikan amo da yawa.
Takaitaccen bayani guda ɗaya mai amfani kowane mako
Samu ingantattun labaran AI na mako, bayanan asali, kayan aiki masu amfani, zaɓen koyo, da sabbin ayyukan AI.
Hayar ƙwararren AI ko ƙaddamar da samfurin AI mai amfani? Sanya shi a gaban mutanen da suka zo nan don koyo da aiki.
Buga aikin AI Ƙaddamar da kayan aikin AI