Fitowar 19 Oktoba, 2026

Jefa tsabar kuɗi yana kama da bazuwar ne kawai. Idan ka san kowane dalla-dalla na jefawar, za ka iya hasashen sakamakon. Shin bazuwar kwantum za ta iya ɓoye irin waɗannan sirrikan dalla-dalla?

A 1943, wani masanin lissafi ɗan Birtaniya ya ce a'a. Hujjarsa ta yi amfani da photon ɗaya da polarizer. Ta tayar da martani goma sha ɗaya, goma daga cikinsu a mujallar Nature, a 1944 da 1945.

Martani na goma sha biyu ba a taɓa buga shi ba. Wani mai bincike a Geneva ya same shi a Vienna, a cikin ma'ajiyar takardun Erwin Schrödinger. Cikin 'yan layuka, Schrödinger ya gina daidai samfurin da aka ce ba zai yiwu ba.

Duk da haka shi ba mai imani da ƙaddara-tabbatacciya (determinism) ba ne; ya yi jayayya don bazuwa a cikin yanayi tun shekarun 1920. Samfurinsa hujja ce ta ka'ida: hujjar rashin yiwuwa tana tabbata ne kawai ga matsalar kamar yadda aka gabatar da ita.

Tushe: Schrödinger's unpublished hidden-variable model in the 1944–45 debate over the "Collapse of Determinism", https://arxiv.org/abs/2610.20752

Megabar ɗaya ya kai kusan sau miliyan na matsin iskar da ke kewaye da kai. Yana mayar da abubuwan da ba sa gudanar da wutar lantarki su zama ƙarfe. Samar da shi ta raƙuman girgiza yawanci yana buƙatar manyan lesar a dakunan gwaje-gwaje na ƙasa.

Wata ƙungiyar MIT ta sami gajeriyar hanya. Lesar ɗaya tana samar da zoben haske huɗu, mafi girma a harba shi da farko, mafi ƙanƙanta a ƙarshe. Kowane zobe yana bugun siririn ruwa mai baƙin tawada daidai lokacin da raƙumin girgiza mai nufar tsakiya yake wucewa, yana ƙara masa tura.

Zobe huɗu na millijoule 4.5 kowanne sun isa. Kwaikwayon da aka daidaita da hanyoyin raƙumin da aka auna sun sanya matsin a tsakiya sama da gigapascal 400. Zobe huɗu a jere sun fi zobe huɗu daban-daban nesa ba kusa ba.

Abu ɗaya da ya kamata a lura: an lissafa wannan matsin, ba a auna shi ba tukuna. Hasken X-ray ya kamata ya tabbatar da shi. Idan aka tabbatar, kimiyyar lissafin abubuwa a yanayi mai tsanani za ta iya dacewa a kan teburin dakin gwaji na yau da kullum.

Tushe: Megabar Shocks on a Benchtop, https://arxiv.org/abs/2610.20749

A 1900, Hilbert ya nemi wata hanya da za ta yanke hukunci ko kowace lissafin polynomial tana da mafita a lambobi cikakku (whole numbers). A 1970, amsar ta zo: babu irin wannan hanyar da za ta iya wanzuwa.

Ga lambobin rabo, amsar ta zo kwanan nan kawai. A farkon Oktoba 2026, wani preprint daga OpenAI ya tabbatar cewa amsar ita ma a'a ce. Babu algorithm da zai yanke hukunci ga kowace lissafi.

Zhi-Wei Sun, a Jami'ar Nanjing, ya gina a kan wannan hujja. Ya nuna cewa rashin yiwuwar yana tabbata har ma ga lissafin da ke da abubuwa 86 kawai da ba a sani ba. Kayan aikinsa su ne ɗigo a kan lanƙwasar elliptic da kuma iyalan lambobin firamare (prime numbers) da aka gina da kyau.

Ga lambobi cikakku, Sun ya riga ya saukar da adadin zuwa 10. A kan lambobin rabo, 86 shi ne abin da wannan gini yake bayarwa. Sun yana kiransa iyaka ta sama, ba iyakar da ba za a iya doke ta ba.

Tushe: 86 unknowns for Hilbert's tenth problem over the rationals, https://arxiv.org/abs/2610.20691

Gallium zai iya kama neutrino. Lokaci-lokaci, ɗaya yana mayar da zarrar gallium zuwa germanium. Ƙirga waɗannan zarrori yana auna adadin neutrino da suka ratsa.

Lokacin da masana kimiyyar lissafi suka nufi gallium da tushe-tushen radiation masu ƙarfi, sun sami zarrorin germanium kusan kashi 20% ƙasa da yadda ake tsammani. Wannan gibin, wanda aka tabbatar a sama da sigma 4, shi ne "baƙon al'amarin gallium" (gallium anomaly).

Wasu masana kimiyyar lissafi biyu sun koma ga wani tushe: Rana. Wasu na'urori, kamar Borexino da SNO, suna auna fitar neutrino ɗinta ba tare da dogaro da gallium ba. Sun yi hasashen kamawa 67.5 a ma'aunin da aka saba. Gwaje-gwajen gallium da kansu sun auna 66.1 ± 3.1.

Abin da aka auna da abin da aka yi hasashe sun yi daidai: rabon shi ne 0.98. Don haka ƙimar kamawa a kan gallium ba ta yi kuskure ba, haka ma ƙirgawar. Baƙon al'amarin gaskiya ne, kuma sanadinsa yana wani wuri daban.

Tushe: Gallium Anomaly and Solar Neutrinos, https://arxiv.org/abs/2610.22219

Hoton X-ray na ƙirji yakan ɗauki hatimi a kusurwarsa: lambar asibiti, lakabi. AI zai iya koyon karanta wannan hatimi maimakon huhu. Sannan ya gaza, a ɓoye, a wani asibiti.

Masu bincike biyu a Jami'ar Makerere sun duba cikin wani AI na hotuna da aka saba amfani da shi. Daga cikin 'attention heads' 144, ɗaya ya fice. A kan kashi 62% na hotunan X-ray 200 na COVID-19, yana kallon rubutun da ke kusurwar sama a hagu.

Goge rubutun da siffar nama (tissue) ta gaskiya, sai hankalin wannan head ya koma zuwa huhu. Tsari ɗaya yana bayyana a cikin wasu samfurorin AI biyu da kuma a kan wani rukunin bayanai na biyu.

Juyin: kashe head ɗin, har da wasu biyu da ake zargi, kuma ganewar cututtuka ba sa canzawa yadda za a iya auna. Gwaje-gwajen da aka saba ba za su iya ganin gajeriyar hanyar ba. Duba cikin samfurin ne kawai ke bayyana ta.

Tushe: Spurious Attention Heads in Vision Transformers: Mechanistic Identification of Annotation-Driven Bias in Medical Chest X-Ray Analysis, https://arxiv.org/abs/2610.21638

Tafiye-tafiye biyu masu tsawo ɗaya za su iya ƙunsar hawa daban-daban sosai. Kuma kowane hawa yana cin makamashi, akai-akai, duk lokacin da aka bi shi da mota. Shin siffar birni tana la'akari da hakan?

Masu bincike a Spain sun ɗauki ainihin tsarin titunan birane 56. Sun riƙe tafiye-tafiye iri ɗaya, daga saƙonnin da ke ɗauke da wuri, sannan suka motsa kowane tsarin titi a kan ƙasarsa: an zame shi, an juya shi, an miƙe shi.

A cikin birane 26 daga cikin 30 da za a iya gwadawa gaba ɗaya, ainihin tsarin yana buƙatar mafi ƙarancin makamashi cikin dukkan sigogin da aka gwada. A cikin sauran huɗun, sigar da aka zame tana ajiye ƙasa da kashi 3%.

Bogotá za ta ajiye kashi 2.5% idan aka motsa ta mita 750 zuwa yamma. Amma wannan ƙasar ta haɗa da wuraren da ambaliya ke yawan shafa a gefen Kogin Bogotá. Wataƙila ainihin birnin ya riga ya zauna a kan ƙasa mafi arha da ake da ita.

Tushe: Mobility energetic balance frames spatial configuration of cities, https://arxiv.org/abs/2610.21558

Don a kwashe mutane kafin guguwa, ana buƙatar sanin inda za ta faɗa, kwanaki da yawa kafin lokaci. Samfurorin yanayi na AI sun riga sun yi kafaɗa da kafaɗa da mafi kyawun hasashen gargajiya a matsakaici. Amma fa a kan guguwa masu tsanani da ba kasafai ake ganin su ba?

Wata ƙungiya daga INRIA, Jami'ar Sorbonne da AXA ta sake kwaikwayon guguwa 178. Ta haɗa samfurorin AI uku waɗanda ke fitar da cikakkun rukunin hasashe da IFS-ENS, rukunin hasashe na ECMWF mai tushe a kimiyyar lissafi.

A lokacin guguwar 2025, samfurin AI mai suna FGN ya sanya yiwuwa mafi yawa a ainihin wurin guguwar fiye da IFS-ENS ga kusan kashi 90% na guguwar. Ya kuma fi sanya wuraren iskoki mafi ƙarfi da matsin iska mafi ƙanƙanta.

Raunin: guguwar AI sun yi rauni fiye da kima. Iskokinsu mafi ƙarfi ba su kai ba, yayin da IFS-ENS ya fi kusantar ma'aunin. Amma alaƙar da ke tsakanin iska da matsin iska a cikin kowace guguwa ta fito daidai.

Tushe: Can AI Forecast the Extremes? Evaluating Probabilistic Weather Models on Tropical Cyclones, https://arxiv.org/abs/2610.22077

Ƙurar sararin samaniya tana rage haske ta wata hanya ta musamman. A cikin Milky Way ɗinmu, tana haƙa rami a cikin hasken ultraviolet kusa da ångström 2,175. Ƙananan ƙwayoyin carbon ne ake zargi yawanci.

Wata ƙungiyar ƙasa da ƙasa da Jami'ar Padua ke jagoranta ta yi nazarin wani galaxy da aka yi wa laƙabi da Baƙar Lu'u-lu'u (Black Pearl). Na'urar hangen nesa ta Webb tana ganinsa kamar yadda yake kusan shekaru miliyan 500 bayan Big Bang.

Hasken sa yana nuna wannan ramin ƙura, an gano shi a sigma 17.6. Shi ne mafi nisa da aka bayar da rahoto zuwa yanzu. Amma yana a ångström 2,132, ya kauce zuwa shuɗi.

Sanannun ramukan daga galaxy da suka zo ɗan daga baya sun kauce zuwa ɗaya ɓangaren. Don haka ƙurar farko ba duka iri ɗaya ba ce. Ƙananan ƙwayoyin graphite ko ƙwayoyin carbon da ake kira PAH duka za su iya dacewa.

Tushe: An Attenuation Bump at z=9.3: Probing Dust Diversity at Cosmic Dawn, https://arxiv.org/abs/2610.20722

Bari shirye-shiryen koyo masu sauƙi biyu su sanya farashi gaba da juna. Suna ƙarewa da cajin da ya wuce farashin gasa sosai. Wani babban bincike na 2020 ya ga barazana a bayansa: ka rage farashi a kaina zan mayar da martani.

Farashin da ake karewa ta ramuwa yana kama da cartel. Wannan fassarar ce ta ba sakamakon nauyi a wurin masu kula da gasa. Masu bincike biyu a São Paulo sun sa shi a gwaje-gwaje biyar.

Ka ɗaga farashinka maimakon rage shi, abokin hamayya zai "rama" haka nan. Ka haɗa wakili (agent) da baƙo, sai haɗin kai ya rushe da kashi 80%. Kuma kaucewa mafi wayo zai biya: ƙarin riba ta kashi 17.5% a matsakaici.

Don haka farashi yana kasancewa sama ba don ana hukunta yaudara ba, amma don ba a taɓa gwada ta ba. Juya waɗanda ke wasa da juna yana karya wannan al'ada.

Tushe: Algorithmic Collusion Without Deterrence: Auditing the Punishment Mechanism in Q-Learning Pricing Agents, https://arxiv.org/abs/2610.21867

Zana babbar da'ira a kan takarda mai murabba'ai ka ƙirga ɗigon raga da ke ciki. Adadin yana kusa da fili ɗin, ba daidai ba taɓa. Har yaya gibin zai iya girma?

Gauss ya nuna cewa gibin yana girma aƙalla kamar X zuwa ƙarfin 1/2. A nan X shi ne murabba'in radius. Masana lissafi suna tsammanin 1/4, kuma ba za a iya yin fiye da haka ba.

Fiye da ƙarni ɗaya, mai nunawa (exponent) yana saukowa a hankali. Masana lissafi huɗu a Beijing yanzu sun kai 209/670, kusan 0.3119. Tarihin da ya gabata ya kasance kusan 0.3145.

Ƙimarsu tana ƙasa da 5/16 kaɗan. Huxley ya taɓa ba da shawarar 5/16 a matsayin iyaka mai yiwuwa ta hanyar da yake ingantawa. Sabbin gyare-gyare biyu na fasaha sun wuce ta. Takardar preprint ce, kuma hanyar zuwa 1/4 har yanzu doguwa ce.

Tushe: A New Bound for the Gauss Circle Problem, https://arxiv.org/abs/2610.21730