arXiv subjects

Shimin Zhang

Publications and source records attributed to Shimin Zhang.

At least 19 recordsLinked to original sources

arXiv2026

First-principles design of main-group dimer defects in ZnO as candidate quantum defects

Zinc oxide (ZnO), a wide-band-gap semiconductor with mature growth techniques, is a promising host for optically active quantum spins. Yet, optically active quantum defects in ZnO remain largely unexplored. Here, we identify and characterize a family of double substitutional impurities in ZnO, formed by main-group donor-acceptor (DA) pairs, as candidates for optically active quantum defects. Using hybrid density functional theory (DFT), we systematically investigate double substitutional DA complexes and their defect physics, including electronic structure, thermodynamic stability, and optical properties. The proposed defects exhibit isolated defect states, strong spin localization on the acceptor site, and $C_{3v}$ symmetry. Importantly, the electronic structure of the DA pairs is largely determined by the atomic properties of their constituent atoms. We further examine their optical characteristics, including zero-phonon lines (ZPLs), radiative lifetimes, and nonradiative decay to assess their viability as color centers. Notably, among the dimers, (Si$_{Zn}$-B$_O$)$^+$ and (Ge$_{Zn}$-B$_O$)$^+$ exhibit visible optical transitions with sub-microsecond radiative lifetimes and robust charge states against optical ionization, while (Si$_{Zn}$-C$_O$)$^{2+}$ shows the smallest Huang-Rhys factor, approximately 5.6. Our results propose a new family of main-group donor-acceptor defects in ZnO as promising candidates for optically active spin defects.

cond-mat.mtrl-sci↗

arXiv2026

$\boldsymbol{f}$-OPD: Stabilizing Long-Horizon On-Policy Distillation with Freshness-Aware Control

Scaling on-policy distillation (OPD) for large language models (LLMs) confronts a fundamental tension: asynchronous execution is necessary for system efficiency, but structurally deviates from the ideal on-policy objective. To address this challenge, we theoretically decompose the objective discrepancy into rollout drift and supervision drift, capturing staleness in student rollout and teacher context, respectively. Building on this, we introduce a sample-level freshness score that quantifies the reliability of a buffered sample with respect to the on-policy objective. Guided by this signal, we further propose f-OPD, a novel framework that adaptively regulates stale-sample influence and constrains policy drift accumulated under asynchronous training. Across reasoning, tool-use, and coding-agent tasks of increasing interaction horizon, f-OPD consistently achieves task performance comparable to synchronous optimization while largely retaining the throughput advantages of asynchronous execution. Our results establish the first recipe for achieving a performance-efficiency trade-off in OPD, paving the way for long-horizon agentic post-training at scale.

cs.LG↗

arXiv2026

Understanding CDCL Solvers via Scalability Studies and Proofdoors

Over the past several decades, CDCL SAT solvers have proven remarkably effective on large industrial formulas, despite SAT being NP-complete and widely believed to be intractable. While considerable empirical research has been done on solver performance over benchmarks like the SAT competition, as well as scaling studies on random and crafted families, surprisingly little effort has gone into systematic scaling studies over industrial instances. To address this gap, we collect a large benchmark of Bounded Model Checking (BMC) instances (76,600+ across 766 families) and perform a systematic scaling study of solver performance. We observe a spectrum: some families scale linearly, others polynomially or exponentially. Building on this foundation, we study the structural parameters that have been proposed to explain this phenomenon. We first show that previously proposed parameters -- clause-variable ratio, treewidth, and community structure -- fail to discriminate between the linear and exponential regimes. By contrast, the recently proposed \emph{proofdoor} parameter explains this phenomenon well. Informally, a proofdoor is a sequence of interpolants between chunks of a formula, where each interpolant represents the solver's memoization of reasoning effort on chunks it has already analyzed. In support of the proofdoor hypothesis, we make three key contributions. First, we empirically show that CDCL solvers do compute small proofdoors for linearly-scaling BMC instances. Second, we show that for exponentially-scaling instances, sampled proofdoors scale exponentially and are typically not incrementally absorbed. Third, we show that scrambling linearly-scaling instances yields larger proofdoor sizes relative to pre-scrambling, relating poor branching order to larger proofdoor sizes and drop in solver performance.

cs.LO↗

arXiv2026

Nonlinear synthetic Schlieren methods for free-surface topography measurement using telecentric imaging

Free-surface synthetic Schlieren (FS-SS) is a high-resolution, refraction-based optical technique for measuring the instantaneous elevation of a liquid interface. Under the assumptions of small amplitude, small slope, and small paraxial angle, the method yields a linear relationship between the gradient of the surface elevation and the apparent displacement field of a refracted pattern imaged through the surface. Here, we propose three new, nonlinear extensions of the FS-SS method that are specifically dedicated to telecentric imaging. Paraxial distortions are eliminated with a telecentric lens, thereby simplifying the optical model. This allows us to derive nonlinear surface reconstruction models that reach beyond the usual limits of small slope and small wave-magnitudes. We implement these nonlinear surface reconstruction algorithms and compare them to the original, linear reconstruction algorithm in three different experiments, using a solid glass lens, spreading oil drops and nonlinear Faraday waves. At the price of a few iterations, we can realise nonlinear surface reconstructions that are more precise, in particular when we reach high slopes or high amplitude regimes. We share a library that encodes these nonlinear surface reconstruction algorithms.

physics.flu-dyn↗

arXiv2025

ReLaX: Reasoning with Latent Exploration for Large Reasoning Models

Reinforcement Learning with Verifiable Rewards (RLVR) has recently demonstrated remarkable potential in enhancing the reasoning capability of Large Reasoning Models (LRMs). However, RLVR often drives the policy toward over-determinism, resulting in ineffective exploration and premature policy convergence. While promoting token-level diversity has shown promise in mitigating entropy collapse, we argue that the latent dynamics underlying token generation encode a far richer computational structure for steering policy optimization toward a more effective exploration-exploitation tradeoff. To enable tractable analysis and intervention of the latent dynamics of LRMs, we leverage Koopman operator theory to obtain a linearized representation of their hidden state dynamics. This enables us to introduce Dynamic Spectral Dispersion (DSD), a new metric to quantify the heterogeneity of the model's latent dynamics, serving as a direct indicator of policy exploration. Building upon these foundations, we propose Reasoning with Latent eXploration (ReLaX), a framework that explicitly incorporates latent dynamics to regulate exploration and exploitation during policy optimization. Comprehensive experiments across a wide range of multimodal and text-only reasoning benchmarks show that ReLaX consistently incentivizes reasoning capability and outperforms existing token-level methods. Our project is available at https://github.com/ZhangShimin1/ReLaX.

cs.LG↗

arXiv2025

IML-Spikeformer: Input-aware Multi-Level Spiking Transformer for Speech Processing

Spiking Neural Networks (SNNs), inspired by biological neural mechanisms, represent a promising neuromorphic computing paradigm that offers energy-efficient alternatives to traditional Artificial Neural Networks (ANNs). Despite proven effectiveness, SNN architectures have struggled to achieve competitive performance on large-scale speech processing tasks. Two key challenges hinder progress: (1) the high computational overhead during training caused by multi-timestep spike firing, and (2) the absence of large-scale SNN architectures tailored to speech processing tasks. To overcome the issues, we introduce Input-aware Multi-Level Spikeformer, i.e. IML-Spikeformer, a spiking Transformer architecture specifically designed for large-scale speech processing. Central to our design is the Input-aware Multi-Level Spike (IMLS) mechanism, which simulates multi-timestep spike firing within a single timestep using an adaptive, input-aware thresholding scheme. IML-Spikeformer further integrates a Re-parameterized Spiking Self-Attention (RepSSA) module with a Hierarchical Decay Mask (HDM), forming the HD-RepSSA module. This module enhances the precision of attention maps and enables modeling of multi-scale temporal dependencies in speech signals. Experiments demonstrate that IML-Spikeformer achieves word error rates of 6.0\% on AiShell-1 and 3.4\% on Librispeech-960, comparable to conventional ANN transformers while reducing theoretical inference energy consumption by 4.64$\times$ and 4.32$\times$ respectively. IML-Spikeformer marks an advance of scalable SNN architectures for large-scale speech processing in both task performance and energy efficiency. Our source code and model checkpoints are publicly available at github.com/Pooookeman/IML-Spikeformer.

cs.MM↗

arXiv2025

Artificial Symmetry Breaking by Self-Interaction Error

Symmetry is a cornerstone of quantum mechanics and materials theory, underpinning the classification of electronic states and the emergence of complex phenomena such as magnetism and superconductivity. While symmetry breaking in density functional theory can reveal strong electron correlation, it may also arise spuriously from self-interaction error (SIE), an intrinsic flaw in many approximate exchange-correlation functionals. In this work, we present clear evidence that SIE alone can induce artificial symmetry breaking, even in the absence of strong correlation. Using a family of one-electron, multi-nuclear-center systems $ \mathrm{H}^+_{n \times \frac{+2}{n}}(R) $, we show that typical semilocal density functionals exhibit symmetry-breaking localization as system size increases, deviating from the exact, symmetry-preserving Hartree-Fock solution. We further demonstrate that this localization error contrasts with the well-known delocalization error of semilocal density functionals and design a semilocal density functional that avoids the artifact. Finally, we illustrate the real-world relevance of this effect in the \ch{Ti_{Zn}v_O} defect in ZnO, where a semilocal density functional breaks the $C_{3v}$ symmetry while a hybrid density functional preserves it. These findings highlight the need for improved functional design to prevent spurious symmetry breaking in both model and real materials.

cond-mat.mtrl-sci↗

arXiv2025

Deep Spin Defects in Zinc Oxide for High-Fidelity Single-Shot Readout

Wide-bandgap oxides such as ZnO are favorable hosts for spin defect qubits due to their dilute nuclear spin background and potential for ultra-high purity. Yet, a deep-level defect qubit with robust optical and spin properties has not been identified in this material. Here, using first-principles calculations, we predict that the molybdenum-vacancy complex, Mo_Zn-V_O, exhibits the essential characteristics of an optically addressable spin qubit: a spin-triplet ground state, visible-range optical transitions with high quantum yield, and an unusually small Huang-Rhys factor (~5, compared to 10-30 in known ZnO defects). We further find long spin coherence times (T_2 ~ 4 ms) when both nuclear and impurity spin baths are considered, with paramagnetic impurities setting a threshold concentration of 0.035 ppm. Importantly, the combination of strong spin-orbit coupling and the absence of Jahn-Teller distortion supports spin-selective intersystem crossing and high-fidelity single-shot readout at elevated temperatures and across wide magnetic field ranges. By identifying ZnO as a host for deep-level defect qubits, our work points toward a pathway to scalable, integrable oxide-based quantum technologies and broadens the material foundation for solid-state quantum information science.

cond-mat.mtrl-sci↗

arXiv2024

Excited-State Dynamics and Optically Detected Magnetic Resonance of Solid-State Spin Defects from First Principles

Optically detected magnetic resonance (ODMR) is an efficient and reliable method that enables initialization and readout of spin states through spin-photon interface. In general, high quantum efficiency and large spin-dependent photoluminescence contrast are desirable for reliable quantum information readout. However, reliable prediction of the ODMR contrast from first-principles requires accurate description of complex spin polarization mechanisms of spin defects. These mechanisms often include multiple radiative and nonradiative processes in particular intersystem crossing (ISC)among multiple excited electronic states. In this work we present our implementation of the first-principles ODMR contrast, by solving kinetic master equation with calculated rates from \textit{ab initio} electronic structure methods then benchmark the implementation on the case of the negatively-charged nitrogen vacancy center in diamond. We show the importance of correct description of multi-reference electronic states and pseudo Jahn-Teller effect for quantitatively, even qualitatively correct prediction of spin-orbit coupling (SOC) and the rate of ISC. We present the complete calculation of SOC for different ISC processes that align with both group theory and experimental observations. Moreover, we provide a comprehensive picture of excitation and relaxation dynamics, including previously unexplored internal conversion processes. We show good agreement between our first-principles calculations and the experimental ODMR contrast under magnetic field. We then demonstrate reliable predictions of magnetic field direction, pump power, and microwave frequency dependency, as important parameters for ODMR experiments. Our work provides a predictive computational platform for spin polarization and optical readout of solid-state quantum defects from first principles.

cond-mat.mtrl-sci↗

arXiv2024

Single nuclear spin detection and control in a van der Waals material

Optically active spin defects in solids are leading candidates for quantum sensing and quantum networking. Recently, single spin defects were discovered in hexagonal boron nitride (hBN), a layered van der Waals (vdW) material. Due to its two-dimensional structure, hBN allows spin defects to be positioned closer to target samples than in three-dimensional crystals, making it ideal for atomic-scale quantum sensing, including nuclear magnetic resonance (NMR) of single molecules. However, the chemical structures of these defects remain unknown, and detecting a single nuclear spin with an hBN spin defect has been elusive. In this study, we created single spin defects in hBN using $^{13}$C ion implantation and identified three distinct defect types based on hyperfine interactions. We observed both S=1 and S=1/2 spin states within a single hBN spin defect. We demonstrated atomic-scale NMR and coherent control of individual nuclear spins in a vdW material, with a $\pi$-gate fidelity up to 99.75% at room temperature. By comparing experimental results with density-functional theory calculations, we propose chemical structures for these spin defects. Our work advances the understanding of single spin defects in hBN and provides a pathway to enhance quantum sensing using hBN spin defects with nuclear spins as quantum memories.

quant-ph↗

arXiv2024

TC-LIF: A Two-Compartment Spiking Neuron Model for Long-Term Sequential Modelling

The identification of sensory cues associated with potential opportunities and dangers is frequently complicated by unrelated events that separate useful cues by long delays. As a result, it remains a challenging task for state-of-the-art spiking neural networks (SNNs) to establish long-term temporal dependency between distant cues. To address this challenge, we propose a novel biologically inspired Two-Compartment Leaky Integrate-and-Fire spiking neuron model, dubbed TC-LIF. The proposed model incorporates carefully designed somatic and dendritic compartments that are tailored to facilitate learning long-term temporal dependencies. Furthermore, a theoretical analysis is provided to validate the effectiveness of TC-LIF in propagating error gradients over an extended temporal duration. Our experimental results, on a diverse range of temporal classification tasks, demonstrate superior temporal classification capability, rapid training convergence, and high energy efficiency of the proposed TC-LIF model. Therefore, this work opens up a myriad of opportunities for solving challenging temporal processing tasks on emerging neuromorphic computing systems. Our code is publicly available at https://github.com/ZhangShimin1/TC-LIF.

cs.NE↗

arXiv2023

Four years of multi-modal odometry and mapping on the rail vehicles

Precise, seamless, and efficient train localization as well as long-term railway environment monitoring is the essential property towards reliability, availability, maintainability, and safety (RAMS) engineering for railroad systems. Simultaneous localization and mapping (SLAM) is right at the core of solving the two problems concurrently. In this end, we propose a high-performance and versatile multi-modal framework in this paper, targeted for the odometry and mapping task for various rail vehicles. Our system is built atop an inertial-centric state estimator that tightly couples light detection and ranging (LiDAR), visual, optionally satellite navigation and map-based localization information with the convenience and extendibility of loosely coupled methods. The inertial sensors IMU and wheel encoder are treated as the primary sensor, which achieves the observations from subsystems to constrain the accelerometer and gyroscope biases. Compared to point-only LiDAR-inertial methods, our approach leverages more geometry information by introducing both track plane and electric power pillars into state estimation. The Visual-inertial subsystem also utilizes the environmental structure information by employing both lines and points. Besides, the method is capable of handling sensor failures by automatic reconfiguration bypassing failure modules. Our proposed method has been extensively tested in the long-during railway environments over four years, including general-speed, high-speed and metro, both passenger and freight traffic are investigated. Further, we aim to share, in an open way, the experience, problems, and successes of our group with the robotics community so that those that work in such environments can avoid these errors. In this view, we open source some of the datasets to benefit the research community.

cs.RO↗

arXiv2023

Long Short-term Memory with Two-Compartment Spiking Neuron

The identification of sensory cues associated with potential opportunities and dangers is frequently complicated by unrelated events that separate useful cues by long delays. As a result, it remains a challenging task for state-of-the-art spiking neural networks (SNNs) to identify long-term temporal dependencies since bridging the temporal gap necessitates an extended memory capacity. To address this challenge, we propose a novel biologically inspired Long Short-Term Memory Leaky Integrate-and-Fire spiking neuron model, dubbed LSTM-LIF. Our model incorporates carefully designed somatic and dendritic compartments that are tailored to retain short- and long-term memories. The theoretical analysis further confirms its effectiveness in addressing the notorious vanishing gradient problem. Our experimental results, on a diverse range of temporal classification tasks, demonstrate superior temporal classification capability, rapid training convergence, strong network generalizability, and high energy efficiency of the proposed LSTM-LIF model. This work, therefore, opens up a myriad of opportunities for resolving challenging temporal processing tasks on emerging neuromorphic computing machines.

cs.NE↗

arXiv2023

Effect of Environmental Screening and Strain on Optoelectronic Properties of Two-Dimensional Quantum Defects

Point defects in hexagonal boron nitride (hBN) are promising candidates as single-photon emitters (SPEs) in nanophotonics and quantum information applications. The precise control of SPEs requires in-depth understanding of their optoelectronic properties. However, how the surrounding environment of host materials, including number of layers, substrates, and strain, influences SPEs has not been fully understood. In this work, we study the dielectric screening effect due to the number of layers and substrates, and the strain effect on the optical properties of carbon dimer and nitrogen vacancy defects in hBN from first-principles many-body perturbation theory. We report that the environmental screening causes lowering of the GW gap and exciton binding energy, leading to nearly constant optical excitation energy and exciton radiative lifetime. We explain the results with an analytical model starting from the BSE Hamiltonian with Wannier basis. We also show that optical properties of quantum defects are largely tunable by strain with highly anisotropic response, in good agreement with experimental measurements. Our work clarifies the effect of environmental screening and strain on optoelectronic properties of quantum defects in two-dimensional insulators, facilitating future applications of SPEs and spin qubits in low-dimensional systems.

cond-mat.mtrl-sci↗

arXiv2023

TEA-PSE 3.0: Tencent-Ethereal-Audio-Lab Personalized Speech Enhancement System For ICASSP 2023 DNS Challenge

This paper introduces the Unbeatable Team's submission to the ICASSP 2023 Deep Noise Suppression (DNS) Challenge. We expand our previous work, TEA-PSE, to its upgraded version -- TEA-PSE 3.0. Specifically, TEA-PSE 3.0 incorporates a residual LSTM after squeezed temporal convolution network (S-TCN) to enhance sequence modeling capabilities. Additionally, the local-global representation (LGR) structure is introduced to boost speaker information extraction, and multi-STFT resolution loss is used to effectively capture the time-frequency characteristics of the speech signals. Moreover, retraining methods are employed based on the freeze training strategy to fine-tune the system. According to the official results, TEA-PSE 3.0 ranks 1st in both ICASSP 2023 DNS-Challenge track 1 and track 2.

eess.AS↗

arXiv2023

Two-step Band-split Neural Network Approach for Full-band Residual Echo Suppression

This paper describes a Two-step Band-split Neural Network (TBNN) approach for full-band acoustic echo cancellation. Specifically, after linear filtering, we split the full-band signal into wide-band (16KHz) and high-band (16-48KHz) for residual echo removal with lower modeling difficulty. The wide-band signal is processed by an updated gated convolutional recurrent network (GCRN) with U$^2$ encoder while the high-band signal is processed by a high-band post-filter net with lower complexity. Our approach submitted to ICASSP 2023 AEC Challenge has achieved an overall mean opinion score (MOS) of 4.344 and a word accuracy (WAcc) ratio of 0.795, leading to the 2$^{nd}$ (tied) in the ranking of the non-personalized track.

eess.AS↗

arXiv2022

Personalized Acoustic Echo Cancellation for Full-duplex Communications

Deep neural networks (DNNs) have shown promising results for acoustic echo cancellation (AEC). But the DNN-based AEC models let through all near-end speakers including the interfering speech. In light of recent studies on personalized speech enhancement, we investigate the feasibility of personalized acoustic echo cancellation (PAEC) in this paper for full-duplex communications, where background noise and interfering speakers may coexist with acoustic echoes. Specifically, we first propose a novel backbone neural network termed as gated temporal convolutional neural network (GTCNN) that outperforms state-of-the-art AEC models in performance. Speaker embeddings like d-vectors are further adopted as auxiliary information to guide the GTCNN to focus on the target speaker. A special case in PAEC is that speech snippets of both parties on the call are enrolled. Experimental results show that auxiliary information from either the near-end speaker or the far-end speaker can improve the DNN-based AEC performance. Nevertheless, there is still much room for improvement in the utilization of the finite-dimensional speaker embeddings.

cs.SD↗

arXiv2022

Multi-Task Deep Residual Echo Suppression with Echo-aware Loss

This paper introduces the NWPU Team's entry to the ICASSP 2022 AEC Challenge. We take a hybrid approach that cascades a linear AEC with a neural post-filter. The former is used to deal with the linear echo components while the latter suppresses the residual non-linear echo components. We use gated convolutional F-T-LSTM neural network (GFTNN) as the backbone and shape the post-filter by a multi-task learning (MTL) framework, where a voice activity detection (VAD) module is adopted as an auxiliary task along with echo suppression, with the aim to avoid over suppression that may cause speech distortion. Moreover, we adopt an echo-aware loss function, where the mean square error (MSE) loss can be optimized particularly for every time-frequency bin (TF-bin) according to the signal-to-echo ratio (SER), leading to further suppression on the echo. Extensive ablation study shows that the time delay estimation (TDE) module in neural post-filter leads to better perceptual quality, and an adaptive filter with better convergence will bring consistent performance gain for the post-filter. Besides, we find that using the linear echo as the input of our neural post-filter is a better choice than using the reference signal directly. In the ICASSP 2022 AEC-Challenge, our approach has ranked the 1st place on word accuracy (WAcc) (0.817) and the 3rd place on both mean opinion score (MOS) (4.502) and the final score (0.864).

cs.SD↗