Neutrino AI model maps hidden signals and ranks events by error
Preprint: The one-million-event PolarBERT analysis identified detector signals and tested an uncertainty head for ranking likely angular error, with calibration limits.
1341–1360
Preprint: The one-million-event PolarBERT analysis identified detector signals and tested an uncertainty head for ranking likely angular error, with calibration limits.
Preprint: A formal construction makes ghost confinement conditional on bound states that the study does not show can form.
Preprint: A formal point-free construction links Esakia frames with Heyting algebras and restricts Townsend’s localic Priestley duality without excluded middle or non-constructive choice.
A 2026 arXiv preprint reports a theorem for reduced plane-curve tuples and a narrower result in arbitrary dimensions.
Preprint reports gains in Ebola nowcasting, Nigerian food-security downscaling and US health and disaster-risk benchmarks.
Preprint: PlanSightRAG performed strongly on visual retrieval and controlled compliance tests, while real-plan rule extraction remained unreliable.
An arXiv preprint dated 26 August 2026 builds strictly positive definite examples that reach the guaranteed order, then fail at the next derivative.
An arXiv preprint reports 62.5% average accuracy for a visual-dependence approach trained sequentially on answer-unlabeled image-question pairs across six multimodal tasks.
Preprint: A 7,512-sample project combines exercise video, physiological signals and expert error labels, but the supplied text omits detailed performance figures.
Preprint analysis finds near-solar elemental abundances and an elevated carbon isotope ratio in the star.
Preprint: A methods paper shows how calibrated spectra can bound known torque signals and extra stationary stochastic noise.
Preprint modeling associates a spatial Hall-viscosity change with a stronger vortex in one chamber and a weaker or suppressed vortex in the other.
Preprint: A four-sightline analysis links some absorbing structures to the Local Bubble’s wall and places another more than 100 parsecs beyond it, with important uncertainties.
This preprint reports strong benchmark scores and 50% lower training computation, but the evidence comes largely from AI judges and small human studies.
The theoretical work examines whether spin and valley can unlock inside rhombohedral graphene walls and predicts a possible superconducting-junction voltage.
Preprint: A proof-based study extends the level-2 result to every bounded convex polygon with at least four sides.
Preprint: Zero-WAM averaged 46.95% success on seven unseen simulation tasks and posted higher rates than LingBot-VA in three real-world task families.
Preprint: Separate tests report stronger GUI ranking than BM25 and high self-rated usability for RaWi.
Preprint: Outputs using selected English claims had more supported additions and fewer hallucinations on average, but results varied by language and generator.
A mathematical two-trader model links interior equilibrium to price matching and finds first-mover effects, while reserve scale changes numerical welfare comparisons.