Metric-compatible coarse corrections transfer gradients across manifolds, reducing iterations and CPU time on Kohn-Sham, Gross-Pitaevskii, and segmentation problems.
Local scene graphs plus DINOv2-guided object codes lift KITTI PSNR to 30.31 while preserving object editing across long driving sequences.
Auditing 33,158 UK employment claims shows outcome-revealing phrases can deliver 56.07 Macro-F1 by themselves, close to full-text classifiers.
A BBGKY argument couples entropy and Fisher-information inequalities to extend Lacker’s chaos method to smooth mean-field diffusions.
A modal DG solver stores the Vlasov distribution in tensor-train form, reaching 10⁴-scale compression on weakly nonlinear benchmarks.
Calibrated proxy axes turn image-derived attributes into auditable subgroup posteriors across fundus, dermoscopy, and chest radiography datasets.
Instruction-level binary features train a classifier that recovers 91.75% of unsafe regions at a 6.16% false positive rate.
A three-port dq/DC admittance test separates self and coupling failures, guiding a DE-OSR controller that stabilizes weak-grid operation without extra tuning.
A Pareto-frontier analysis of GABD estimators identifies an extended $(\phi,\gamma)$ divergence that minimizes asymptotic variance at a chosen breakdown point.
