Chronological feed of everything captured from Patrick McKenzie.
Using STM in a vector magnetic field, researchers discovered a previously unreported staggered charge-density-wave (CDW) in UTe₂ and demonstrated that a modest 1.7 T field aligned along the a-axis (quasi-1D uranium chain direction) completely quenches it, while fields along other crystallographic axes leave it intact. Counterintuitively, the same field orientation simultaneously suppresses the 5f Kondo resonance and modifies the hybridization gap. Modeling attributes this correlated behavior to a field-driven switch of the dominant hybridization channel from Te 5p–U 5f to U 6d–U 5f coupling, implying an intimate mechanistic link between the CDW order and the Kondo effect. This establishes directional magnetic field tuning as an effective experimental knob for disentangling and controlling the intertwined orders that underlie UTe₂'s unconventional superconductivity.
This paper formally introduces semilattices of Mal'cev blocks (SMB algebras) — algebraic structures where each element of a semilattice is "blown up" into a Mal'cev algebra — and proves that all such algebras induce tractable templates of the Constraint Satisfaction Problem (CSP). Beyond tractability, the authors demonstrate structural convergence between the two independent proofs of the CSP Dichotomy Theorem (Bulatov's and Zhuk's), showing they behave more similarly than previously recognized when restricted to SMB algebras. This is the second paper in a series aimed at deepening understanding of the Dichotomy Theorem's proof landscape.
This paper presents a spatial transfer-learning framework that fuses sparse ground-station measurements with chemical transport model (CTM) outputs to generate daily, fine-grid NO₂ predictions across Italy for 2023, with uncertainty quantification. Built on the LatticeKrig geostatistical framework, the method uses an image-to-image neural network to amortize estimation of millions of spatially varying covariance parameters from gridded CTM data in seconds, then transfers that nonstationary, anisotropic correlation structure to point-level station data. A likelihood-based refinement step corrects for the spatial smoothing inherent in CTM outputs, recovering fine-scale variability. The result outperforms stationary formulations and is designed to serve downstream applications in epidemiology and environmental policy.