Preprint reports near-linear bounds for Khatri-Rao sketches
A theorem says the required sketching dimension can scale almost linearly with subspace size for any fixed tensor order under specific assumptions.
Search
2,451 results
441–460
New MOF models improve predictions of gas adsorption
Preprint: A large training set and literature benchmark accompany reported gains on selected gas-adsorption tests.
Verifier-guided method posts higher scores in long AI tasks
Preprint: A method called VICT recorded higher language-model agent scores on two benchmarks, but its proof system sometimes declined to assign credit.
Preprint links softmax attention rank to interaction geometry
A theorem gives a sharp worst-case size law, while native-coordinate features added no detectable held-out information in tested model heads.
Curved Lattice Model Finds Sharp Change in Energy Response
Preprint: A Haldane-model calculation finds a third-order energy-response discontinuity close to a relativistic benchmark, while its overall size remains model dependent.
AI Video Model Reports Tighter Camera Control in Tests
A preprint reports that MANIFOLD4D used a one-time 3D render to guide video re-shoots, with visual quality close to Vista4D in benchmark tests.
Preprint finds a controlled route through spin-boson models
The mathematical analysis controls a non-closing flow under gap and small-coupling assumptions, while broader extensions remain unproved.
New framework reduces time-dependent systems with symmetries
Preprint: A geometric method keeps the link between two phase spaces while allowing symmetries that shift time.
String Theory Study Links Scattering Equations to K-Identities
Preprint: An analytical study proposes a generating function that organizes an infinite hierarchy of identities in hard string scattering.
Enteral Nutrition Associated With Better Biomarker Trends in ICU Stroke
A seven-day cohort of 73 critically ill adults found different I-FABP and MMP-9 patterns, but clinicians chose the nutrition route rather than assigning it at random.
Missed specialist consultation linked to higher admission on ED return
A peer-reviewed retrospective study at a Turkish university hospital found higher admission among adults returning within 72 hours when specialist input had been missed at the first visit.
Most sampled patients had positive bile cultures, review finds
A review of 16 observational reports on biliary infection and cholangitis found a 77.3% sampling yield, but wide study differences limit the result.
Jazan school study finds stunting in 12.3% of adolescents
A cross-sectional study of 324 adolescents found stunting was more common than thinness and linked increasing age with higher odds, but it could not establish cause.
Protocol sets out test of 20-minute cool-water burn first aid
The plan will track whether emergency and EMS teams deliver cool running water within three hours, alongside grafting and other patient outcomes.
Butterfly genomes point to a patchwork of early reproductive isolation
A comparison across six Heliconius erato populations found the clearest link between chromatin accessibility and genetic divergence at intermediate separation, with important limits.
BL Lacertae flare hints at an ultra-compact gamma-ray source
Preprint: Two 2024-25 flares showed hour-scale changes, while a possible sub-minute signal remained tentative.
Preprint maps two boundary ensembles onto one holographic model
A type IIB calculation assigns the sign of an added topological term to fixed flux or fixed potential, then matches both perturbative boundary partition functions.
Surface Acoustic Waves Track Charge-Density-Wave Transitions
Preprint: Thin films of NbSe3 and 2H-TaSe2 showed temperature-dependent acoustoelectric currents, with polarity changes that varied with wave direction.
Preprint reports a smaller coarse space for multiscale solvers
A two-level method uses high-permeability features for selected small time steps, while full NLMC or an augmented space is proposed for general steps.
AI Training Filter Reports Higher Math and Code Scores
This preprint reports higher benchmark averages from a trajectory filter than from standard on-policy distillation in tested math and code comparisons, with nearly the same training time.