arXiv subjects

Kyle Chen

Publications and source records attributed to Kyle Chen.

6 recordsLinked to original sources

arXiv2026

Explainable AI-based Decision Support for Nocturnal Hypoglycemia Prevention in Type 1 Diabetes

Purpose: Nocturnal hypoglycemia (NH) remains a challenge for individuals with type 1 diabetes (T1D), particularly those who are physically active or on multiple daily injections (MDI). We leveraged an explainable evidential neural network model that forecasts minimum overnight glucose to identify NH risk factors and generate recommendations for NH prevention. Methods: We analyzed the impact of glucose and physical activity (PA) factors on predicted NH probability using SHapley Additive exPlanations (SHAP). Data were from 20 adults with T1D on MDI (10 females; mean age 39 years; HbA1c 7\%) who participated in a cross-over study (NCT05967260). Results: A total of 502 nights were analyzed. Bedtime glucose was the strongest predictor of NH. Other relevant factors included recent exposure to high or low glucose, glucose variability before bedtime, and timing of PA. Some associations appeared physiologically counterintuitive, possibly reflecting behavioral adjustments. Based on the identified risk factors and their SHAP values, we determined key decision points and developed recommendations that may help prevent NH, such as consuming a bedtime snack or discussing potential adjustments to insulin therapy with a healthcare provider. Conclusion: Identifying predictors of NH offers insights for managing NH risk in clinical decision support.

q-bio.QM↗

arXiv2025

Realistic Oscillon Interactions

Oscillons are long-lived nonlinear pseudo-solitonic configurations of scalar fields and many plausible inflationary scenarios predict an oscillon-dominated phase in the early universe. Many possible aspects of this phase remain unexplored, particularly oscillon-oscillon interactions and interactions between oscillons and their environment. However the primary long range forces between oscillons are gravitational and thus slow-acting relative to the intrinsic timescales of the oscillons themselves. Given that simulations with local gravity are computationally expensive we explore these effects by extracting individual specimens from simulations and then engineering interactions. We find that oscillons experience friction when moving in an inhomogeneous background and, because oscillons in non-relativistic collisions bounce or merge as a function of their relative phases, the outcomes of interactions between ``wild'' oscillons depend on their specific trajectories.

astro-ph.CO↗

arXiv2025

Mixed Planewave and Localized Orbital Basis for Sparse-Stochastic Hybrid TDDFT

We present a mixed basis-set approach to obtain optical absorption spectra within a generalized Kohn-Sham time-dependent density functional theory framework. All occupied valence molecular orbitals (MOs) are expanded in a plane-wave (PW) basis, while unoccupied MOs are derived primarily from localized atomic basis functions. The method accelerates spectral convergence when compared to fully PW-based simulations, with a $2-3$ fold reduction in the number of unoccupied MOs entering the Casida equation. The mixed-basis is placed on a common real-space grid, enabling our previously developed deterministic/sparse-stochastic evaluation of the exact exchange operator (J. Chem. Theory Comput. 2023, 19, 9239-9247). This chemically intuitive and computationally efficient approach is validated across various molecular systems, including $\pi$-conjugated polymethine dyes, aromatic hydrocarbons, and a chlorophyll monomer.

physics.chem-ph↗

arXiv2024

Extracting Geometry and Topology of Orange Pericarps for the Design of Bioinspired Energy Absorbing Materials

As a result of evolution, many biological materials have developed irregular structures that lead to outstanding mechanical properties, like high stiffness-to-weight ratios and good energy absorption. To reproduce these properties in synthetic materials, biomimicry typically replicates the irregular natural structure, often leading to fabrication challenges. Here, we present a bioinspired material design method that instead reduces the irregular natural structure to a finite set of building blocks, also known as tiles, and rules to connect them, and then uses these elements as instructions to generate synthetic materials with mechanical properties similar to the biological materials. We demonstrate the method using the pericarp of the orange, a member of the citrus family known for its protective, energy-absorbing capabilities. We generate polymer samples and characterize them under quasi-static and dynamic compression and observe spatially-varying stiffness and good energy absorption, as seen in the biological materials. By quantifying which tiles and connectivity rules locally deform in response to loading, we determine how to spatially control the stiffness and energy absorption.

cond-mat.mtrl-sci↗

arXiv2024

Dynamics of Symmetry-Protected Topological Matter on a Quantum Computer

Control of topological edge modes is desirable for encoding quantum information resiliently against external noise. Their implementation on quantum hardware, however, remains a long-standing problem due to current limitations of circuit depth and noise, which grows with the number of time steps. By utilizing recently developed constant-depth quantum circuits in which the circuit depth is independent of time, we demonstrate successful long-time dynamics simulation of bulk and surface modes in topological insulators on noisy intermediate-scale quantum (NISQ) processors, which exhibits robust signatures of localized topological modes. We further identify a class of one-dimensional topological Hamiltonians that can be readily simulated with NISQ hardware. Our results provide a pathway towards stable long-time implementation of topological quantum spin systems on present day quantum processors.

quant-ph↗

arXiv2020

Single-atom catalysts boost nitrogen electroreduction reaction

Ammonia (NH3) is mainly produced through the traditional Haber-Bosch process under harsh conditions with huge energy consumption and massive carbon dioxide (CO2) emission. The nitrogen electroreduction reaction (NERR), as an energy-efficient and environment-friendly process of converting nitrogen (N2) to NH3 under ambient conditions, has been regarded as a promising alternative to the Haber-Bosch process and has received enormous interest in recent years. Although some exciting progress has been made, considerable scientific and technical challenges still exist in improving the NH3 yield rate and Faradic efficiency, understanding the mechanism of the reaction and promoting the wide commercialization of NERR. Single-atom catalysts (SACs) have emerged as promising catalysts because of its atomically dispersed activity sites and maximized atom efficiency, unsaturated coordination environment, and its unique electronic structure, which could significantly improve the rate of reaction and yield rate of NH3. In this review we briefly introduce the unique structural and electronic features of SACs, which contributes to comprehensively understand the reaction mechanism owing to their structural simplicity and diversity, and in turn expedite the rational design of fantastic catalysts at the atomic scale. Then, we summarize the most recent experimental and computational efforts on developing novel SACs with excellent NERR performance, including precious metal-, nonprecious metal- and nonmetal-based SACs. Finally, we present challenges and perspectives of SACs on NERR, as well as some potential means for advanced NERR catalyst.

cond-mat.mes-hall↗