arXiv subjects

Yuqing Zhou

Publications and source records attributed to Yuqing Zhou.

17 recordsLinked to original sources

arXiv2026

Influence of magnetic fields on the performance of spin-orbit torque magnetic random-access memory

Spin-orbit torque magnetic random-access memory (SOT-MRAM) offers high speed, ultrahigh endurance, and compatibility with advanced semiconductor processes, making it a promising candidate for next-generation nonvolatile memory. However, intrinsic bias fields in magnetic tunnel junctions (MTJs), originating from reference-layer stray fields and interlayer coupling, cause asymmetric critical switching currents and increased energy consumption. Existing compensation approaches usually introduce additional magnetic layers into the MTJ stack, which increases fabrication complexity and limits wafer-scale integration. Here, we propose a bias-compensation strategy without modifying the MTJ stack by engineering local stray magnetic fields through magnetic filling materials in vertical interconnect access (VIA) channels during the back-end-of-line process. Micromagnetic simulations show that the proposed magnetic filling layer can provide the required auxiliary field for deterministic switching and significantly suppress write-current asymmetry. By optimizing the MTJ position relative to the magnetic filling structure, the write-current bias ratio is reduced from 21.6% in the conventional design to 1.3%. The approach is also applicable to in-plane magnetic anisotropy SOT-MTJs, reducing the bias ratio from 19.8% to -0.2%. Scaling analysis further demonstrates that the compensation effect remains effective when the device size is reduced to 20% of the original dimension (MTJ diameter approximately 10 nm), indicating its potential for high-density SOT-MRAM integration.

physics.app-ph↗

arXiv2026

Deep Learning with Magnetic Parameter Constraints for Short-Term Prediction of Solar Active Region Vector Magnetic Fields

Forecasting the dynamic evolution of solar magnetic fields is a critical technique for enabling space weather warnings. Addressing the limitations of existing methods in predicting all vector magnetic field components and in maintaining consistency with solar surface magnetic-field-related quantities, this study proposes a deep learning prediction method that integrates dynamic masks of active regions with multiple magnetic parameter constraints. By constructing a three-channel representation of vector magnetic fields, applying dynamic masks to enhance attention to strong-field regions, and incorporating multi-parameter magnetic parameter constraints, we developed an end-to-end short-term (12-hour) predictive model of solar vector magnetic field evolution. Using SDO/SHARP vector magnetogram data, the model predicts and analyses field evolution across all components. Quantitative evaluations demonstrate that our approach achieves horizon-averaged structural similarity index measure (SSIM) of 0.912 (per-hour range: 0.909--0.916) and correlation coefficient (CC) of 0.998 for the radial component Br (root-mean-square error (RMSE) 13.0--21.0 G); the horizontal components achieve Bphi SSIM 0.760--0.800 (CC 0.910--0.945, RMSE 38.5--50.0 G) and Btheta SSIM 0.728--0.750 (CC 0.895--0.920, RMSE 38.5--49.0 G). The model maintains unsigned magnetic flux prediction errors at 7.82% (95% confidence interval (CI): +/-0.11%). These results demonstrate strong image-domain performance together with consistency under the magnetic-parameter diagnostics used here, suggesting initial potential for supporting future space weather forecasting efforts.

astro-ph.SR↗

arXiv2026

Confidence-Orchestrated Self-Evolution against Uncertain LLM Feedback

Self-evolving large language models (LLMs) learn by generating their own training tasks and solutions, reducing reliance on human-curated supervision. However, in many reasoning domains, the model must also validate generated tasks and judge generated answers to obtain training signals. This creates a training-signal challenge: erroneous self-judgments become erroneous gradient updates. Existing approaches either rely on external verifiers, which limits generality, or treat noisy self-generated feedback as supervision. We propose COSE (Confidence-Orchestrated Self-Evolution), which uses the LLM's intrinsic confidence as a lightweight uncertainty signal to modulate learning. COSE introduces confidence-weighted PPO updates and confidence-prioritized replay. Across 19 held-out benchmarks and four Qwen/Llama backbones (0.6B--4B), COSE consistently improves over base models and achieves the best average performance in general reasoning and mathematics, while remaining competitive on code. Code and data are available at https://anonymous.4open.science/r/COSE_-B5C2.

cs.AI↗

arXiv2026

LSTM-PINN for Steady-State Electrothermal Transport: Preserving Multi-Field Consis tency in Strongly Coupled Heat and Fluid Flow

Steady-state electrothermal systems involve strongly coupled heat transfer, fluid flow, and electric-potential transport, creating severe numerical challenges for standard physics-informed neural networks (PINNs) due to stark disparities in gradient scales and residual stiffnesses across the physical fields. To resolve these multiphysics bottlenecks, we introduce a Long Short-Term Memory PINN (LSTM-PINN) framework that utilizes a depth-recursive memory mechanism to preserve long-range spatial feature dependencies and maintain strict cross-field consistency. The proposed architecture is rigorously evaluated against conventional and attention-based networks across a unified five-field formulation encompassing four complex convective and drag regimes: Boussinesq electrothermal flow, drift-potential gauge-constrained transport, strong buoyancy-coupled convection, and Brinkman--Forchheimer drift. Quantitative and visual analyses demonstrate that LSTM-PINN successfully suppresses non-physical artifacts and structural distortions, yielding the highest thermodynamic fidelity and consistently outperforming state-of-the-art baselines in global error metrics. Ultimately, this memory-enhanced approach provides a highly robust and accurate computational baseline for capturing localized boundary layers and complex energy-momentum feedback in advanced electrothermal energy systems.

physics.comp-ph↗

arXiv2026

Residual Attention Physics-Informed Neural Networks for Robust Multiphysics Simulation of Steady-State Electrothermal Energy Systems

Efficient thermal management and precise field prediction are critical for the design of advanced energy systems, including electrohydrodynamic transport, microfluidic energy harvesters, and electrically driven thermal regulators. However, the steady-state simulation of these electrothermal coupled multiphysics systems remains challenging for physics-informed neural computation due to strong nonlinear field coupling, temperature-dependent coefficient variability, and complex interface dynamics. This study proposes a Residual Attention Physics-Informed Neural Network (RA-PINN) framework for the unified solution of coupled velocity, pressure, electric-potential, and temperature fields. By integrating a unified five-field operator formulation with residual-connected feature propagation and attention-guided channel modulation, the proposed architecture effectively captures localized coupling structures and steep gradients. We evaluate RA-PINN across four representative energy-relevant benchmarks: constant-coefficient coupling, indirect pressure-gauge constraints, temperature-dependent transport, and oblique-interface consistency. Comparative analysis against Pure-MLP, LSTM-PINN, and pLSTM-PINN demonstrates that RA-PINN achieves superior accuracy, yielding the lowest MSE, RMSE, and relative $L_2$ errors across all scenarios. Notably, RA-PINN maintains high structural fidelity in interface-dominated and variable-coefficient settings where conventional PINN backbones often fail. These results establish RA-PINN as a robust and accurate computational framework for the high-fidelity modeling and optimization of complex electrothermal multiphysics in sustainable energy applications.

cs.LG↗

arXiv2026

WISE-Flow: Workflow-Induced Structured Experience for Self-Evolving Conversational Service Agents

Large language model (LLM)-based agents are widely deployed in user-facing services but remain error-prone in new tasks, tend to repeat the same failure patterns, and show substantial run-to-run variability. Fixing failures via environment-specific training or manual patching is costly and hard to scale. To enable self-evolving agents in user-facing service environments, we propose WISE-Flow, a workflow-centric framework that converts historical service interactions into reusable procedural experience by inducing workflows with prerequisite-augmented action blocks. At deployment, WISE-Flow aligns the agent's execution trajectory to retrieved workflows and performs prerequisite-aware feasibility reasoning to achieve state-grounded next actions. Experiments on ToolSandbox and $\tau^2$-bench show consistent improvement across base models.

cs.CL↗

arXiv2025

A fine-grained look at causal effects in causal spaces

The notion of causal effect is fundamental across many scientific disciplines. Traditionally, quantitative researchers have studied causal effects at the level of variables; for example, how a certain drug dose (W) causally affects a patient's blood pressure (Y). However, in many modern data domains, the raw variables-such as pixels in an image or tokens in a language model-do not have the semantic structure needed to formulate meaningful causal questions. In this paper, we offer a more fine-grained perspective by studying causal effects at the level of events, drawing inspiration from probability theory, where core notions such as independence are first given for events and sigma-algebras, before random variables enter the picture. Within the measure-theoretic framework of causal spaces, a recently introduced axiomatisation of causality, we first introduce several binary definitions that determine whether a causal effect is present, as well as proving some properties of them linking causal effect to (in)dependence under an intervention measure. Further, we provide quantifying measures that capture the strength and nature of causal effects on events, and show that we can recover the common measures of treatment effect as special cases.

stat.ME↗

arXiv2025

JW-Flare: Accurate Solar Flare Forecasting Method Based on Multimodal Large Language Models

Solar flares, the most powerful explosive phenomena in the solar system, may pose significant hazards to spaceborne satellites and ground-based infrastructure. Despite decades of intensive research, reliable flare prediction remains a challenging task. Large Language Models, as a milestone in artificial intelligence, exhibit exceptional general knowledge and next-token prediction capabilities. Here we introduce JW-Flare, the first Multimodal Large Language Models (MLLMs) explicitly trained for solar flare forecasting through fine-tuning on textual physic parameters of solar active regions and magnetic field images. This method demonstrates state-of-the-art (SOTA) performance for large flares prediction on the test dataset. It effectively identifies all 79 X-class flares from 18,949 test samples, yielding a True Skill Statistic (TSS) of 0.95 and a True Positive Rate (TPR) of 1.00, outperforming traditional predictive models. We further investigate the capability origins of JW-Flare through explainability experiments, revealing that solar physics knowledge acquired during pre-training contributes to flare forecasting performance. Additionally, we evaluate models of different parameter scales, confirming the Scaling_Law of Large Language Models in domain-specific applications, such as solar physics. This study marks a substantial advance in both the scale and accuracy of solar flare forecasting and opens a promising avenue for AI-driven methodologies in broader scientific domains.

astro-ph.SR↗

arXiv2024

Fighting Spurious Correlations in Text Classification via a Causal Learning Perspective

In text classification tasks, models often rely on spurious correlations for predictions, incorrectly associating irrelevant features with the target labels. This issue limits the robustness and generalization of models, especially when faced with out-of-distribution data where such spurious correlations no longer hold. To address this challenge, we propose the Causally Calibrated Robust Classifier (CCR), which aims to reduce models' reliance on spurious correlations and improve model robustness. Our approach integrates a causal feature selection method based on counterfactual reasoning, along with an unbiased inverse propensity weighting (IPW) loss function. By focusing on selecting causal features, we ensure that the model relies less on spurious features during prediction. We theoretically justify our approach and empirically show that CCR achieves state-of-the-art performance among methods without group labels, and in some cases, it can compete with the models that utilize group labels.

cs.LG↗

arXiv2024

Navigating the Shortcut Maze: A Comprehensive Analysis of Shortcut Learning in Text Classification by Language Models

Language models (LMs), despite their advances, often depend on spurious correlations, undermining their accuracy and generalizability. This study addresses the overlooked impact of subtler, more complex shortcuts that compromise model reliability beyond oversimplified shortcuts. We introduce a comprehensive benchmark that categorizes shortcuts into occurrence, style, and concept, aiming to explore the nuanced ways in which these shortcuts influence the performance of LMs. Through extensive experiments across traditional LMs, large language models, and state-of-the-art robust models, our research systematically investigates models' resilience and susceptibilities to sophisticated shortcuts. Our benchmark and code can be found at: https://github.com/yuqing-zhou/shortcut-learning-in-text-classification.

cs.CL↗

arXiv2024

Shell topology optimization based on level set method

This paper proposes a level set-based method for optimizing shell structures with large design changes in shape and topology. Conventional shell optimization methods, whether parametric or nonparametric, often only allow limited design changes in shape. In the proposed method, the shell structure is defined as the isosurface of a level set function. The level set function is iteratively updated based on the shape sensitivity on the surface mesh. Therefore, the proposed method can represent an arbitrary manifold surface while dealing with topological changes, for example, from a spherical surface to a toroidal surface. We applied the proposed method to the mean compliance minimization problems of 3D shell structural designs for dome, bending plate and cantilever beam examples to demonstrate its efficacy of the proposed method.

cs.CE↗

arXiv2022

Inference on the Best Policies with Many Covariates

Understanding the impact of the most effective policies or treatments on a response variable of interest is desirable in many empirical works in economics, statistics and other disciplines. Due to the widespread winner's curse phenomenon, conventional statistical inference assuming that the top policies are chosen independent of the random sample may lead to overly optimistic evaluations of the best policies. In recent years, given the increased availability of large datasets, such an issue can be further complicated when researchers include many covariates to estimate the policy or treatment effects in an attempt to control for potential confounders. In this manuscript, to simultaneously address the above-mentioned issues, we propose a resampling-based procedure that not only lifts the winner's curse in evaluating the best policies observed in a random sample, but also is robust to the presence of many covariates. The proposed inference procedure yields accurate point estimates and valid frequentist confidence intervals that achieve the exact nominal level as the sample size goes to infinity for multiple best policy effect sizes. We illustrate the finite-sample performance of our approach through Monte Carlo experiments and two empirical studies, evaluating the most effective policies in charitable giving and the most beneficial group of workers in the National Supported Work program.

stat.ME↗

arXiv2022

Cellular Topology Optimization on Differentiable Voronoi Diagrams

Cellular structures manifest their outstanding mechanical properties in many biological systems. One key challenge for designing and optimizing these geometrically complicated structures lies in devising an effective geometric representation to characterize the system's spatially varying cellular evolution driven by objective sensitivities. A conventional discrete cellular structure, e.g., a Voronoi diagram, whose representation relies on discrete Voronoi cells and faces, lacks its differentiability to facilitate large-scale, gradient-based topology optimizations. We propose a topology optimization algorithm based on a differentiable and generalized Voronoi representation that can evolve the cellular structure as a continuous field. The central piece of our method is a hybrid particle-grid representation to encode the previously discrete Voronoi diagram into a continuous density field defined in a Euclidean space. Based on this differentiable representation, we further extend it to tackle anisotropic cells, free boundaries, and functionally-graded cellular structures. Our differentiable Voronoi diagram enables the integration of an effective cellular representation into the state-of-the-art topology optimization pipelines, which defines a novel design space for cellular structures to explore design options effectively that were impractical for previous approaches. We showcase the efficacy of our approach by optimizing cellular structures with up to thousands of anisotropic cells, including femur bone and Odonata wing.

cs.GR↗

arXiv2019

Anisotropic multicomponent topology optimization for additive manufacturing with build orientation design and stress-constrained interfaces

This paper presents a multicomponent topology optimization method for designing structures assembled from additively-manufactured components, considering anisotropic material behavior for each component due to its build orientation, distinct material behavior and stress constraint at component interfaces (i.e., joints). Based upon the multicomponent topology optimization (MTO) framework, the simultaneous optimization of structural topology, its partitioning, and the build orientations of each component is achieved, which maximizes an assembly-level structural stiffness performance subject to maximum stress constraints at component interfaces. The build orientations of each component are modeled by its orientation tensor that avoids numerical instability experienced by the conventional angular representation. A new joint model is introduced at component interfaces, which enables the identification of the interface location, the specification of a distinct material tensor, and imposing maximum stress constraints during optimization. Both 2D and 3D numerical examples are presented to illustrate the effect of the build orientation anisotropy and the component interface behavior on the resulting multicomponent assemblies.

cs.CE↗

arXiv2019

Inverse design of fluid flow structure with Turing pattern

Microchannel reactors are critical in biological plus energy-related applications and require meticulous design of hundreds-to-thousands of fluid flow channels. Such systems commonly comprise intricate space-filling microstructures to control the fluid flow distribution for the reaction process. Traditional flow channel design schemes are intuition-based or utilize analytical rule-based optimization strategies that are oversimplified for large-scale domains of arbitrary geometry. Here, a gradient-based optimization method is proposed, where effective porous media and fluid velocity vector design information is exploited and linked to explicit microchannel parameterizations. Reaction-diffusion equations are then utilized to generate space-filling Turing pattern microchannel flow structures from the porous media field. With this computationally efficient and broadly applicable technique, precise control of fluid flow distribution is demonstrated across large numbers (on the order of hundreds) of microchannels.

cs.CE↗

arXiv2019

Topology optimization for 3D thin-walled structures with adaptive meshing

This paper presents a density-based topology optimization method for designing 3D thin-walled structures with adaptive meshing. Uniform wall thickness is achieved by simultaneously constraining the minimum and maximum feature sizes using Helmholtz partial differential equations (PDE). The PDE-based constraints do not require information about neighbor cells and therefore can readily be integrated with an adaptive meshing scheme. This effectively enables the 3D topology optimization of thin-walled structures with a desktop PC, by significantly reducing computation in large void regions that appear during optimization. The uniform feature size constraint, when applied to 3D structures, can produce thin-walled geometries with branches and holes, which have previously been difficult to obtain via topology optimization. The resulting thin-walled structures can provide valuable insights for designing thin-walled lightweight structures made of stamping, investment casting and composite manufacturing.

cs.CE↗

arXiv2011

On the Performance of P2P Network: An Assortment Method

P2P systems have grown dramatically in recent years. The most popular type of P2P systems are file sharing networks, which are used to share various types of content over the Internet. Due to the increase in popularity of P2P systems, the network performance of these systems has become a very important issue in the design and realization of these networks. Hence, the performance of the P2P has been improved. This paper will suggest the following methods for the improvement of the P2P systems: Method-1: Improve the P2P routing by using a sandwich technique. Method-2: Improve the search performance by introducing a new search based on the super peer. Method-3: Improving the search by introducing a ranking algorithm based on the knowledge database. The system demonstrates that the methods introduced here have the improved efficiency compared to the previous methodologies. So, the results show that the performance of the P2P systems have been improved by using the above methods, hence the traffic can be reduced.

physics.data-an↗