Search for: All records

Deep learning–based classifiers are widely used for perception in autonomous Cyber-Physical Systems (CPS’s). However, such classifiers rarely offer guarantees of perfect accuracy while being optimized for efficiency. To support safety-critical perception, ensembles of multiple different classifiers working in concert are typically used. Since CPS’s interact with the physical world continuously, it is not unreasonable to expect dependencies among successive inputs in a stream of sensor data. Prior work introduced a classification technique that leverages these inter-input dependencies to reduce the average time to successful classification using classifier ensembles. In this paper, we propose generalizations to this classification technique, both in the improved generation of classifier cascades and the modeling of temporal dependencies. We demonstrate, through theoretical analysis and numerical evaluation, that our approach achieves further reductions in average classification latency compared to the prior methods. 

The topological Hall effect (THE), a quantum phenomenon arising from the emergent magnetic field generated by a topological spin texture, is a key method for detecting non-coplanar spin structures like skyrmions in magnetic materials. Here, we investigate a bilayer structure of Pt and the conducting ferrimagnet NiCo2O4 (NCO) of perpendicular magnetic anisotropy and demonstrate a giant THE across a temperature range of 2–350 K. The absence of THE in a single-layer Pt and NCO, as well as in Pt/Cu/NCO, suggests its interfacial origin. The maximum THE occurring just before the NCO coercive field indicates its connection to magnetic nucleation centers, which are topologically equivalent to skyrmions. The large normalized THE, based on the emergent-field model, points to a high population density of small magnetic nucleation centers. This aligns with the seemingly unresolvable domain structures by the employed techniques during magnetization reversal, even though clear domain structures are detected after zero-field cooling. These results establish heavy metal/NCO as a promising system for exploring topological spin structures. 

In this study, two pairs of 0D chiral copper iodide clusters were synthesized. The structural rigidity is increased by halogen modulation to obtain a near unity PLQY. The applications in white LED and X-ray imaging are extremely promising. 

We theoretically investigate the spin structure of weakly bound diatomic van der Waals molecules formed by two identical bosonic alkali atoms. Our studies were performed using known Born-Oppenheimer potentials while developing a reduced interaction potential model. Such reduced potential models are currently a key for solving certain classes of few-body problems of atoms as they decrease the numerical burden on the computation. Although the reduced potentials are significantly shallower than actual Born-Oppenheimer potentials, they still capture the main properties of the near-threshold bound states, including their spin structure, and the scattering states over a broad range of magnetic fields. At zero magnetic field, we find that the variation in spin structure across different alkali species originates from the interplay between electronic spin exchange and hyperfine interactions. To characterize this competition we introduce a single parameter that is a function of the singlet and triplet scattering lengths, the atomic hyperfine splitting constant, and the molecular binding energy. We show that this parameter can be used to classify the spin structure of vdW molecules for each atomic species. Published by the American Physical Society2025