Ana sabunta kullun1800 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.
Wata sabuwar takarda ta arXiv tana ba da shawarar ka'idar sake dubawa na wakilai uku wanda mai bita ya kimanta lambar wakili da kuma binciken masu sukar da ke bitar kafin a yi gyara. Marubutan sun ba da rahoton ingantattun sakamakon ƙididdiga sama da ginshiƙan da aka gwada, yayin da kuma suna gano yarjejeniya ta ƙarya azaman yanayin gazawa.
Fitilar arXiv tana kwatanta sifili-harbi da ingantaccen daidaitawa mai inganci don gano maganganun ƙiyayya a cikin Roman Urdu. Marubutan sun ba da rahoton cewa daidaitawar LoRA ta ɗaga aikin F1 daga 0.56 zuwa sama da 0.93 akan ƙungiyar da ke ɗauke da bayanai sama da 72,000.
Takardar arXiv ta gabatar da FraudBench, ma'auni don gwada ko masu amfani da kayan aiki na banki na iya gano zamba da ke bayyana a cikin tattaunawa. A cikin tantancewar gwaji guda na farko, wakilai huɗu sun sami kashi 49 zuwa 65% akan tsaron harin.
A study of more than 2,000 language-model training runs says scarce target data can be repeated 15–20 times in mixtures, with the best rate varying by scale and compute.
Apple researchers describe LINK, a pretraining intervention that replaces selected English words with word-level translations from a target language. The paper reports improvements across eight languages and five model sizes, including up to a twofold speedup in reaching equivalent downstream performance.
A position paper reports tacit collusion by DeepSeek-R1 agents in a simulated Bertrand pricing market, even after human prompts against collusion. It argues that observed-behavior certification should precede deployment of reasoning agents in economic markets; the evidence and safeguards remain preliminary.
A systematic review surveys how large language models are being studied for mental-health analysis, risk assessment, therapy support and multimodal monitoring, while stressing unresolved ethical and regulatory challenges.
An ICML 2026 position paper analyzing 500 Hugging Face model cards argues that open-weight foundation models need coordinated model cards, acceptable-use policies, and licenses to address safety and governance gaps.
A position paper proposes the Transition Complexity Profile, a standardized way to describe how unpredictable and long-range the dynamics of game environments are for game-world modeling and reinforcement learning.
A new arXiv benchmark places 15 language-model agents in a 20-year football-management simulation, testing whether they can make consistent decisions when short-term choices affect long-term outcomes.
A new arXiv benchmark reports that changing only the retrieval method raised a fixed model’s exact accuracy on financial reconciliation cases from 2.05% to 72.44%. The study also finds that a correct root-cause label often does not mean the system returned sufficient evidence for an auditable diagnosis.
Wani sabon binciken arXiv yana gabatar da gwaje-gwajen doka don ƙirar rubutu-zuwa-hoto a cikin lissafin kasafin kuɗi daga 10^19 zuwa 10^22 FLOPs. Marubutanta sun ba da rahoton cewa ƙirar hoto na buƙatar ƙarin bayanai a kowane siga fiye da ƙirar harshe don horarwa da inganci.
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