arXiv subjects

Wei-Cheng Lee

Publications and source records attributed to Wei-Cheng Lee.

At least 19 recordsLinked to original sources

arXiv2026

Stripe antiferromagnetism and chiral superconductivity in tWSe$_2$

The layer-dependent Hamiltonians of parallel-stacked MoTe$_2$ and WSe$_2$ homobilayer moiré materials are topologically non-trivial, both in real space and in momentum space, and have been shown to support integer and fractional quantum anomalous Hall states, as well as antiferromagnetic and superconducting states. Here, we address the interplay between the antiferromagnetic and superconducting states observed in tWSe$_2$ when the Fermi level is close to its $M$-point van Hove singularity and the displacement field is small. We combine DFT with path-integrals to construct a minimal moiré band model that accounts for lattice relaxation along the $c$-axis and perform Hartree-Fock calculations to identify competing charge and spin ordered states. For tWSe$_2$ at $θ=2.7^\circ$ and $θ=3.65^\circ$, we find that a layer antiferromagnet (AFM), a stripe spin-density-wave (SDW), and the ferromagnetic Chern insulator (FM) are the primary candidates for the ground state at zero displacement field, and argue that antiferromagnetic spin interactions on the next neighbor bond $J_2$ can induce a time-reversal symmetry breaking chiral superconducting state.

cond-mat.str-el↗

arXiv2026

A Single Stepsize Suffices for Unprojected Linear TD(0): Simultaneous Robust and Fast Rates via Polyak--Ruppert Averaging

We study linear TD(0) under Markovian sampling, where data are generated along a single trajectory. We provide high-probability guarantees for a plain unprojected TD(0) algorithm with Polyak-Ruppert (PR) averaging, using a single stepsize schedule $η_t \propto \frac{1}{τ_{\mathrm{mix}}\log(t)\sqrt{t}}$ that depends on the mixing time but requires no prior knowledge of the curvature parameter $ω$. Our first result shows that such a choice of the stepsize guarantees that the TD(0) iterates are automatically and uniformly bounded with high probability, without projections and without any stability argument based on $ω$. Building on this result, we establish a simultaneous high-probability convergence guarantee for the PR average: the same stepsize yields both a robust curvature-free $\widetilde{\mathcal{O}}\!\left(\frac{τ_{\mathrm{mix}}}{\sqrt{T}}\right)$ rate and a fast curvature-dependent $\widetilde{\mathcal{O}}\!\left(\frac{τ_{\mathrm{mix}}^2}{ωT}\right)$rate, with the bound taking the minimum of the two. The core technical ingredient is a Poisson-equation toolkit for geometrically mixing Markov chains, which decomposes Markov noise into a martingale term plus a controlled remainder and enables a new self-bounding inductive argument for pathwise stability.

cs.LG↗

arXiv2026

A Robust $\widetilde{\mathcal{O}}(1/\sqrt{T})$ Rate for Unprojected TD Learning with Linear Function Approximation

We investigate the finite-time convergence properties of Temporal Difference (TD) learning with linear function approximation, a cornerstone of reinforcement learning. We are interested in the so-called ``robust'' setting, where the convergence guarantee does not depend on the potential function's minimal curvature. While prior work has established convergence guarantees in this setting, these results typically rely on the artificial assumption that each iterate is projected onto a bounded set. Removing such a condition was left as an open problem by Bhandari et al. (COLT'18), hypothesizing the need for additional ``regularity conditions''. In this paper, we show that the simple unprojected TD(0) converges with a rate of $\widetilde{\mathcal{O}}\left(\frac{\|θ^*\|^2_2}{\sqrt{T}}\right)$ in expectation, even in the presence of Markovian noise. We do not require an additional regularity condition, but only a minor polylog correction to the learning rate. Our analysis reveals a novel self-bounding property of the TD updates and exploits it to guarantee bounded iterates.

cs.LG↗

arXiv2026

Imaging the Magnetically Driven Reconstruction of the Electronic States in the Antiferromagnetic Topological Insulator EuSn$_2$As$_2$

The realization of the axion insulator phase in magnetic topological insulators is often hindered by crystalline symmetries that protect gapless surface states, even when time-reversal symmetry is broken. Here, we use variable-temperature scanning tunneling microscopy (STM) and spectroscopy (STS), complemented with density functional theory (DFT), to investigate the local electronic structure of the antiferromagnetic (AFM) topological insulator EuSn$_2$As$_2$ across its Néel transition at $T_N = 24$ K. On the (001) surface, we observe a substantial density of intrinsic Sn vacancies that introduce nanoscale electronic inhomogeneity and p-type doping. Upon cooling below $T_N$, we resolve the emergence of two distinct magnetically driven gaps: a $\sim$100 meV gap near the Fermi level and a $\sim$50 meV gap at the ARPES-resolved Dirac point. We attribute the former gap to AFM Brillouin-zone folding and hybridization. The characteristics of the 50 meV gap point toward the lifting of mirror-symmetry protection by Sn vacancies and the consequent mass gapping of the Dirac point, although contributions from AFM-induced folding hybridization cannot be entirely ruled out. Our findings provide real-space evidence for strong coupling between localized moments and itinerant topological states, highlighting exfoliable EuSn$_2$As$_2$ as a potential candidate for realizing axion-insulator-based devices.

cond-mat.str-el↗

arXiv2025

A Best-of-Both-Worlds Proof for Tsallis-INF without Fenchel Conjugates

In this short note, we present a simple derivation of the best-of-both-world guarantee for the Tsallis-INF multi-armed bandit algorithm from J. Zimmert and Y. Seldin. Tsallis-INF: An optimal algorithm for stochastic and adversarial bandits. Journal of Machine Learning Research, 22(28):1-49, 2021. URL https://jmlr.csail.mit.edu/papers/volume22/19-753/19-753.pdf. In particular, the proof uses modern tools from online convex optimization and avoid the use of conjugate functions. Also, we do not optimize the constants in the bounds in favor of a slimmer proof.

cs.LG↗

arXiv2025

Topologically trivial semiconducting behavior and polaronic effects in antiferromagnetic EuZn$_2$As$_2$ and EuCd$_2$Sb$_2$

The Eu-based EuA$_2$X$_2$ (A = Zn, Cd, In, Sn; X = P, As, Sb) family of compounds has recently attracted significant attention as a promising platform for exploring magnetic topological materials, with several members either predicted or reported to exhibit nontrivial topological properties. We investigate the previously reported topological semimetals, EuZn$_2$As$_2$ and EuCd$_2$Sb$_2$, using scanning tunneling microscopy and spectroscopy, complemented by various first-principles computational approaches. Through examination of the cleaved surfaces, step-edges, and defect states, we determine the trivial semiconducting behavior in both material systems, with no evidence of topological surface or edge states. These experimental results are consistent with our theoretical analysis revealing the absence of topological band inversion in either system. Our experimental observations also reveal numerous intrinsic defects that trap charge carriers. These defects may facilitate the formation of magnetic polarons, providing a natural explanation for the colossal negative magnetoresistance observed in many of the EuA$_2$X$_2$ material systems.

cond-mat.str-el↗

arXiv2025

MenTeR: A fully-automated Multi-agenT workflow for end-to-end RF/Analog Circuits Netlist Design

RF/Analog design is essential for bridging digital technologies with real-world signals, ensuring the functionality and reliability of a wide range of electronic systems. However, analog design procedures are often intricate, time-consuming and reliant on expert intuition, and hinder the time and cost efficiency of circuit development. To overcome the limitations of the manual circuit design, we introduce MenTeR - a multiagent workflow integrated into an end-to-end analog design framework. By employing multiple specialized AI agents that collaboratively address different aspects of the design process, such as specification understanding, circuit optimization, and test bench validation, MenTeR reduces the dependency on frequent trial-and-error-style intervention. MenTeR not only accelerates the design cycle time but also facilitates a broader exploration of the design space, demonstrating robust capabilities in handling real-world analog systems. We believe that MenTeR lays the groundwork for future "RF/Analog Copilots" that can collaborate seamlessly with human designers.

cs.AI↗

arXiv2025

New Lower Bounds for Stochastic Non-Convex Optimization through Divergence Decomposition

We study fundamental limits of first-order stochastic optimization in a range of nonconvex settings, including L-smooth functions satisfying Quasar-Convexity (QC), Quadratic Growth (QG), and Restricted Secant Inequalities (RSI). While the convergence properties of standard algorithms are well-understood in deterministic regimes, significantly fewer results address the stochastic case, where only unbiased and noisy gradients are available. We establish new lower bounds on the number of noisy gradient queries to minimize these classes of functions, also showing that they are tight (up to a logarithmic factor) in all the relevant quantities characterizing each class. Our approach reformulates the optimization task as a function identification problem, leveraging divergence decomposition arguments to construct a challenging subclass that leads to sharp lower bounds. Furthermore, we present a specialized algorithm in the one-dimensional setting that achieves faster rates, suggesting that certain dimensional thresholds are intrinsic to the complexity of non-convex stochastic optimization.

stat.ML↗

arXiv2024

Nonlinear and nonreciprocal transport effects in untwinned thin films of ferromagnetic Weyl metal SrRuO$_3$

The identification of distinct charge transport features, deriving from nontrivial bulk band and surface states, has been a challenging subject in the field of topological systems. In topological Dirac and Weyl semimetals, nontrivial conical bands with Fermi-arc surface states give rise to negative longitudinal magnetoresistance due to chiral anomaly effect and unusual thickness dependent quantum oscillation from Weyl-orbit effect, which were demonstrated recently in experiments. In this work, we report the experimental observations of large nonlinear and nonreciprocal transport effects for both longitudinal and transverse channels in an untwinned Weyl metal of SrRuO$_3$ thin film grown on a SrTiO$_{3}$ substrate. From rigorous measurements with bias current applied along various directions with respect to the crystalline principal axes, the magnitude of nonlinear Hall signals from the transverse channel exhibits a simple sin$α$ dependence at low temperatures, where $α$ is the angle between bias current direction and orthorhombic [001]$_{\rm o}$, reaching a maximum when current is along orthorhombic [1-10]$_{\rm o}$. On the contrary, the magnitude of nonlinear and nonreciprocal signals in the longitudinal channel attains a maximum for bias current along [001]$_{\rm o}$, and it vanishes for bias current along [1-10]$_{\rm o}$. The observed $α$-dependent nonlinear and nonreciprocal signals in longitudinal and transverse channels reveal a magnetic Weyl phase with an effective Berry curvature dipole along [1-10]$_{\rm o}$ from surface states, accompanied by 1D chiral edge modes along [001]$_{\rm o}$.

cond-mat.mes-hall↗

arXiv2023

Ab-initio study of the energy competition between Γand K valleys in bilayer transition metal dichalcogenides

Moiré engineering in two-dimensional van der Waals bilayer crystals has emerged as a flexible platform for controlling strongly correlated electron systems. The competition between valleys for the band extremum energy position in the parent layers is crucial in deciding the qualitative nature of the moiré Hamiltonian since it controls the physics of the moiré minibands. Here we use density functional theory to examine the competition between K and $Γ$ for the valence band maximum in homo- and hetero-bilayers formed from the transition metal dichalcogenides (TMD), MX\{_2} where M=Mo,W and X=S,Se,Te. We shed light on how the competition is influenced by interlayer separation, which can be modified by applying pressure, by external gate-defined electric fields, and by transition metal atom d-orbital correlations. Our findings are related to several recent experiments, and contribute to the development of design rules for moiré materials.

cond-mat.mtrl-sci↗

arXiv2022

Topological phase transition in commensurate multi-frequency Floquet Su-Schrieffer-Heeger model

Recently, Floquet systems have attracted a great deal of interest as they offer unprecedented ability to engineer topological states through the tuning of an external time-periodic drive. Consequentially, seeking new driving protocols that allow for more exotic topological phases and transitions becomes imperative for the Floquet engineer. In this paper, we study the Su-Schrieffer-Heeger model driven by two time-dependent periodic sources with commensurate frequencies and an amplitude modulation. Imposing more than one driving frequency allows us to realize even more exotic topological phases resulting from new couplings appearing in the Fourier space representation. Moreover, we find an experimentally practical method for sweeping the system through a topological phase transition by varying the amplitude mixture of the commensurate sources. We employ the local Chern marker, a real space representation of the Chern number, to simulate topological phase diagrams of the two-drive Floquet Hamiltonian in a variety of driving cases.

cond-mat.mes-hall↗

arXiv2022

Mechanism of the Resistivity Switching Induced by the Joule Heating in Crystalline NbO$_2$

Recently the memristive electrical transport properties in NbO$_2$ have attracted much attention for their promising application to the neuromorphic computation. At the center of debates is whether the metal-to-insulator transition (MIT) originates from the structural distortion (Peierls) or the electron correlation (Mott). With inputs from experiments and first principles calculations, we develop a thermodynamical model rooted in the scenario of the MIT driven by a $2^{nd}$ order Peierls instability. We find that the temperature dependence of the electrical conductivity can be accurately fit by the band gap varying with temperature due to the gradual weakening of the Nb-Nb dimers. The resistivity switching can consequently be understood by dimer-free metallic domains induced by local Joule heating. In solving the heat equation, we find that the steady state can not be reached if the applied voltage exceeds a threshold, resulting in the chaotic behavior observed in the high voltage and current states. With the Ginzburg-Landau theory and the Joule heating equation, the evolution of the metallic domains under bias voltage can be simulated and directly verified by experiments.

cond-mat.mes-hall↗

arXiv2022

The thickness dependence of quantum oscillations in ferromagnetic Weyl metal SrRuO$_{3}$

Quantum oscillations in resistivity and magnetization at high magnetic fields are a macroscopic fingerprint of the energy quantization due to the cyclotron motion of quasiparticles. In a thin Weyl semimetal, a unique thickness dependent Weyl-orbit quantum oscillation was proposed to exist, originating from a nonlocal cyclotron orbit via the electron tunneling between the top and bottom Fermi-arc surface states. Here, untwinned and high crystalline Weyl metal SrRuO$_3$ thin films with different thicknesses were grown on miscut SrTiO$_3$ (001) substrates. Magneto-transport measurements were carried out in magnetic fields up to 35 T, and quantum oscillations with different frequencies were observed and compared to the calculated band structure. In particular, we discovered a frequency $F \approx$ 30 T at low temperatures and above 3 T that corresponds to a small Fermi pocket with a light effective mass. Its oscillation amplitude appears to be at maximum for film thicknesses in a range of 10 to 20 nm, and the phase of the oscillation exhibits a systematic change with the film thickness. After isolating the well separated frequencies, the constructed Landau fan diagram shows an unusual concave downward curvature in the 1/$μ_0H_n$-$n$ curve, where $n$ is the Landau level index. Based on the rigorous analysis of the thickness and field-orientation dependence of the quantum oscillations, the oscillation with $F \approx$ 30 T is attributed to be of surface origin, which is related to the Fermi-arc surface state originating from non-overlapping Weyl nodes projected on the film's surface plane. Those findings can be understood within the framework of the Weyl-orbit quantum oscillation effect with non-adiabatic corrections.

cond-mat.mes-hall↗

arXiv2022

Surface state evolution induced by magnetic order in axion insulator candidate EuIn2As2

Gapping of Dirac surface states through time reversal symmetry breaking may realize the axion insulator state in condensed matter. Despite tremendous efforts, only a few material systems fall in this category of intrinsic magnetic topological insulators (TI). Recent theoretical calculations proposed the antiferromagnetic EuIn$_2$As$_2$ to be a topologically non-trivial magnetic insulator with gapped surface states. Here we use scanning tunneling microscopy and spectroscopy (STM/STS) complemented with density-functional theory (DFT) calculations and modelling to probe the surface electronic states in EuIn$_2$As$_2$. We find a spin-orbit induced bulk gap of ~120 meV located only a few meV above the Fermi energy, within which topological surface states reside. Temperature dependent measurements provide evidence of the partial gapping (~40 meV) of the surface states at low temperatures below the AFM order, which decreases with increasing temperature but remains finite above $T_N$

cond-mat.str-el↗

arXiv2022

Physics-Guided Problem Decomposition for Scaling Deep Learning of High-dimensional Eigen-Solvers: The Case of Schrödinger's Equation

Given their ability to effectively learn non-linear mappings and perform fast inference, deep neural networks (NNs) have been proposed as a viable alternative to traditional simulation-driven approaches for solving high-dimensional eigenvalue equations (HDEs), which are the foundation for many scientific applications. Unfortunately, for the learned models in these scientific applications to achieve generalization, a large, diverse, and preferably annotated dataset is typically needed and is computationally expensive to obtain. Furthermore, the learned models tend to be memory- and compute-intensive primarily due to the size of the output layer. While generalization, especially extrapolation, with scarce data has been attempted by imposing physical constraints in the form of physics loss, the problem of model scalability has remained. In this paper, we alleviate the compute bottleneck in the output layer by using physics knowledge to decompose the complex regression task of predicting the high-dimensional eigenvectors into multiple simpler sub-tasks, each of which are learned by a simple "expert" network. We call the resulting architecture of specialized experts Physics-Guided Mixture-of-Experts (PG-MoE). We demonstrate the efficacy of such physics-guided problem decomposition for the case of the Schrödinger's Equation in Quantum Mechanics. Our proposed PG-MoE model predicts the ground-state solution, i.e., the eigenvector that corresponds to the smallest possible eigenvalue. The model is 150x smaller than the network trained to learn the complex task while being competitive in generalization. To improve the generalization of the PG-MoE, we also employ a physics-guided loss function based on variational energy, which by quantum mechanics principles is minimized iff the output is the ground-state solution.

cs.LG↗

arXiv2021

CoPhy-PGNN: Learning Physics-guided Neural Networks with Competing Loss Functions for Solving Eigenvalue Problems

Physics-guided Neural Networks (PGNNs) represent an emerging class of neural networks that are trained using physics-guided (PG) loss functions (capturing violations in network outputs with known physics), along with the supervision contained in data. Existing work in PGNNs has demonstrated the efficacy of adding single PG loss functions in the neural network objectives, using constant trade-off parameters, to ensure better generalizability. However, in the presence of multiple PG functions with competing gradient directions, there is a need to adaptively tune the contribution of different PG loss functions during the course of training to arrive at generalizable solutions. We demonstrate the presence of competing PG losses in the generic neural network problem of solving for the lowest (or highest) eigenvector of a physics-based eigenvalue equation, which is commonly encountered in many scientific problems. We present a novel approach to handle competing PG losses and demonstrate its efficacy in learning generalizable solutions in two motivating applications of quantum mechanics and electromagnetic propagation. All the code and data used in this work is available at https://github.com/jayroxis/Cophy-PGNN.

cs.LG↗

arXiv2021

Physics-Informed Machine Learning for Optical Modes in Composites

We demonstrate that embedding physics-driven constraints into machine learning process can dramatically improve accuracy and generalizability of the resulting model. Physics-informed learning is illustrated on the example of analysis of optical modes propagating through a spatially periodic composite. The approach presented can be readily utilized in other situations mapped onto an eigenvalue problem, a known bottleneck of computational electrodynamics. Physics-informed learning can be used to improve machine-learning-driven design, optimization, and characterization, in particular in situations where exact solutions are scarce or are slow to come up with.

physics.comp-ph↗

arXiv2021

Quantum-statistical transport phenomena in memristive computing architectures

The advent of reliable, nanoscale memristive components is promising for next generation compute-in-memory paradigms, however, the intrinsic variability in these devices has prevented widespread adoption. Here we show coherent electron wave functions play a pivotal role in the nanoscale transport properties of these emerging, non-volatile memories. By characterizing both filamentary and non-filamentary memristive devices as disordered Anderson systems, the switching characteristics and intrinsic variability arise directly from the universality of electron transport in disordered media. Our framework suggests localization phenomena in nanoscale, solid-state memristive systems are directly linked to circuit level performance. We discuss how quantum conductance fluctuations in the active layer set a lower bound on device variability. This finding implies there is a fundamental quantum limit on the reliability of memristive devices, and electron coherence will play a decisive role in surpassing or maintaining Moore's Law with these systems.

cond-mat.mes-hall↗