Search arXiv⌕ Search

arXiv subjects

Zhe Wang

Publications and source records attributed to Zhe Wang.

At least 127 records · Page 7Linked to original sources
arXiv2026

GloPath: An Entity-Centric Foundation Model for Glomerular Lesion Assessment and Clinicopathological Insights

Glomerular pathology is central to the diagnosis and prognosis of renal diseases, yet the heterogeneity of glomerular morphology and fine-grained lesion patterns remain challenging for current AI approaches. We present GloPath, an entity-centric foundation model trained on over one million glomeruli extracted from 14,049 renal biopsy specimens using multi-scale and multi-view self-supervised learning. GloPath addresses two major challenges in nephropathology: glomerular lesion assessment and clinicopathological insights discovery. For lesion assessment, GloPath was benchmarked across three independent cohorts on 52 tasks, including lesion recognition, grading, few-shot classification, and cross-modality diagnosis-outperforming state-of-the-art methods in 42 tasks (80.8%). In the large-scale real-world study, it achieved an ROC-AUC of 91.51% for lesion recognition, demonstrating strong robustness in routine clinical settings. For clinicopathological insights, GloPath systematically revealed statistically significant associations between glomerular morphological parameters and clinical indicators across 224 morphology-clinical variable pairs, demonstrating its capacity to connect tissue-level pathology with patient-level outcomes. Together, these results position GloPath as a scalable and interpretable platform for glomerular lesion assessment and clinicopathological discovery, representing a step toward clinically translatable AI in renal pathology.

cs.CV↗
arXiv2026

Addressing general measurements in quantum Monte Carlo

Quantum Monte Carlo is one of the most promising approaches for dealing with large-scale quantum many-body systems. It has played an extremely important role in understanding strongly correlated physics. However, two fundamental problems, namely the sign problem and general measurement issues, have seriously hampered its scope of application. We propose a universal scheme to tackle the problems of general measurement. The target observables are expressed as the ratio of two types of partition functions $\langle \mathrm{O} \rangle=\bar{Z}/Z$, where $\bar{Z}=\mathrm{tr} (\mathrm{Oe^{-βH}})$ and $Z=\mathrm{tr} (\mathrm{e^{-βH}})$. These two partition functions can be estimated separately within the reweight-annealing frame, and then be connected by an easily solvable reference point. We have successfully applied this scheme to XXZ model and transverse field Ising model, from 1D to 2D systems, from two-body to multi-body correlations and even non-local disorder operators, and from equal-time to imaginary-time correlations. The reweighting path is not limited to physical parameters, but also works for space and time. Essentially, this scheme solves the long-standing problem of calculating the overlap between different distribution functions in mathematical statistics, which can be widely used in statistical problems, such as quantum many-body computation, big data and machine learning.

cond-mat.str-el↗
arXiv2026

Magnetic dipole-dipole transition for scintillation quenching

A magnetic dipole-dipole interaction is proposed as a scintillation quenching mechanism. The interaction rate follows $R^{-6}$ as the electric dipole-dipole interaction in F$\mathrm{\ddot{o}}$ster resonance energy transfer theory. The proposed mechanism causes a long-range resonance energy transfer, and the resonance condition is that the spins of donor and acceptor electrons both flip, and the energy level differences are the same. The new mechanism is distinct to the known spin-orbit coupling induced intersystem crossing, and it can enhance the overall intersystem crossing rate. When oxygen or organic molecules including heavy elements are dissolved in a liquid scintillator, these requirements are possible to be satisfied. The proposal in the paper adds a new approach for scintillation quenching in liquid scintillators.

physics.ins-det↗
arXiv2026

HIME: Mitigating Object Hallucinations in LVLMs via Hallucination Insensitivity Model Editing

Large Vision-Language Models (LVLMs) have demonstrated impressive multimodal understanding capabilities, yet they remain prone to object hallucination, where models describe non-existent objects or attribute incorrect factual information, raising serious concerns for reliable real-world deployment. While fine-tuning is a commonly adopted mitigation strategy, its high computational cost and practical difficulty motivate the need for training-free alternatives, among which model editing has recently emerged as a promising direction. However, indiscriminate editing risks disrupting the rich implicit knowledge encoded in pre-trained LVLMs, leading to a fundamental question: how much intervention is necessary at each layer to suppress hallucinations while preserving pre-trained knowledge? To address this question, we present a systematic analysis of LVLM decoders built on three widely used large language model backbones-Qwen, LLaMA, and Vicuna-revealing clear layer-wise differences in susceptibility to object hallucination. Building on these insights, we introduce the Hallucination Insensitivity Score (HIS), a principled metric that quantifies each layer's sensitivity to hallucination and provides guidance for targeted intervention. Leveraging HIS, we propose Hallucination Insensitivity Model Editing (HIME), a simple yet effective layer-adaptive weight editing approach that selectively modifies latent features to suppress hallucinations while preserving pre-trained knowledge. Extensive experiments demonstrate that HIME reduces hallucinations by an average of 61.8% across open-ended generation benchmarks, including CHAIR, MME, and GPT-4V-aided evaluation, without introducing additional parameters, inference-time latency, or computational overhead.

cs.CV↗
arXiv2026

A Geometric Probe of the Accuracy-Robustness Trade-off: Sharp Boundaries in Symmetry-Breaking Dimensional Expansion

The trade-off between clean accuracy and adversarial robustness is a pervasive phenomenon in deep learning, yet its geometric origin remains elusive. In this work, we utilize Symmetry-Breaking Dimensional Expansion (SBDE) as a controlled probe to investigate the mechanism underlying this trade-off. SBDE expands input images by inserting constant-valued pixels, which breaks translational symmetry and consistently improves clean accuracy (e.g., from $90.47\%$ to $95.63\%$ on CIFAR-10 with ResNet-18) by reducing parameter degeneracy. However, this accuracy gain comes at the cost of reduced robustness against iterative white-box attacks. By employing a test-time \emph{mask projection} that resets the inserted auxiliary pixels to their training values, we demonstrate that the vulnerability stems almost entirely from the inserted dimensions. The projection effectively neutralizes the attacks and restores robustness, revealing that the model achieves high accuracy by creating \emph{sharp boundaries} (steep loss gradients) specifically along the auxiliary axes. Our findings provide a concrete geometric explanation for the accuracy-robustness paradox: the optimization landscape deepens the basin of attraction to improve accuracy but inevitably erects steep walls along the auxiliary degrees of freedom, creating a fragile sensitivity to off-manifold perturbations.

cs.LG↗
arXiv2026

Stable-MoE: Lyapunov-based Token Routing for Distributed Mixture-of-Experts Training over Edge Networks

The sparse activation mechanism of mixture of experts (MoE) model empowers edge intelligence with enhanced training efficiency and reduced computational resource consumption. However, traditional token routing in distributed MoE training faces significant challenges in resource-constrained edge networks characterized by heterogeneous computing capabilities and stochastic token arrivals, which inevitably suffer from workload backlog, resource inefficiency, and performance degradation. To address this issue, we propose a novel Lyapunov-based token routing framework for distributed MoE training over resource-heterogeneous edge networks, termed Stable-MoE. Specifically, we formulate a stochastic optimization problem to maximize both system throughput and gating consistency via optimizing the token routing strategy and computational resource allocation, while ensuring long-term stability of both token and energy queues at the edge devices. Using the Lyapunov optimization, we transform the intractable long-term optimization problem into tractable per-slot subproblems by enabling online decision-making of token routing and computation frequency utilization without the knowledge of future system states. Experimental results on the SVHN and CIFAR-100 datasets demonstrate that Stable-MoE outperforms the baselines with at least 40% and 5% gains in system throughput and test accuracy, respectively.

cs.DC↗
arXiv2026

Low-Latency Federated Fine-Tuning for Large Language Models Over Wireless Networks

Recently, federated large language models (LLMs) have drawn significant attention thanks to coupled capabilities of LLMs and federated learning (FL) that address privacy concerns in collaborative fine-tuning. However, due to large-scale parameters of LLMs, existing federated LLM fine-tuning frameworks incur significant challenges in resource-constrained clients characterized by heterogeneous computing capabilities and random wireless channels. To address this issue, we propose a joint client-specific pruning and bandwidth allocation (JCPBA) framework for federated LLMs to improve the fine-tuning efficiency over the wireless networks. Specifically, we formulate a fine-tuning latency minimization problem by jointly optimizing pruning rates and bandwidth allocations. Furthermore, we solve this optimization problem using a block coordinate descent method. Extensive experiments on the datasets of Yahoo Answers and GSM8K demonstrate that the proposed framework significantly reduces wall-clock fine-tuning time compared with state-of-the-art baselines and gains equal or lower test loss at the cost of lower computation and communication overhead.

cs.DC↗
arXiv2026

Non-commutative Dynamic Approaches to the Kibble-Zurek Scaling Limit with an Initial Gapless Order

Nonequilibrium many-body physics is one of the core problems in modern physics, while the dynamical scaling from a gapless phase to the critical point is a most important challenge with very few knowledge so far. In the driven dynamics with a tuning rate $R$ across the quantum critical point (QCP) of a system with size $L$, the finite-time scaling shows that the square of the order parameter $m^2$ obeys a simple scaling relation $m^2\propto R^{2β/νr}$ in the Kibble-Zurek (KZ) scaling limit with $RL^r\gg1$. Here, by studying the driven critical dynamics from a gapless ordered phase in the bilayer Heisenberg model, we unveil that the approaches to the scaling region dominated by the KZ scaling limit with $RL^r\gg1$ are {\it non-commutative}: this scaling region is inaccessible for large $R$ and finite medium $L$, while merely accessible for large $L$ and moderately finite $R$. We attribute this to the memory effect induced by the finite-size correction in the gapless ordered phase. This non-commutative property makes $m^2$ still strongly depends on the system size and deviates from $m^2\propto R^{2β/νr}$ even for large $R$. We further show that a similar correction applies to the imaginary-time relaxation dynamics. Our results establish an essential extension of nonequilibrium scaling theory with a gapless ordered initial state.

cond-mat.str-el↗
arXiv2026

A Novel 4-D Dataset Paradigm for Studying Complete Ligand-Protein Dissociation Dynamics

The kinetics and dynamics of drug-protein binding and dissociation are crucial to understanding drug absorption and metabolism. Despite advances in artificial intelligence (AI) tools for drug-protein interaction studies, existing training datasets remain limited to static structures or quasi-static conformations. This paper proposes a novel computational approach for rapidly generating drug-protein dissociation trajectories and presents the inaugural dynamically time-resolved 4-D (t, x, y, z) trajectory database DD-13M. This dataset captures over 26,000 complete dissociation processes for 565 ligand-protein complexes, providing nearly 13 million frames of all-atom simulation trajectories. A deep equivariant generative model, UnbindingFlow, was trained using the DD-13M dataset. This model has the capacity to produce dissociation trajectories for novel targets whilst accurately predicting their rate constants (koff). DD-13M introduces a new type of training dataset for AI models, establishing a de novo paradigm for studying the dynamics of drug-protein interactions.

physics.comp-ph↗
arXiv2026

MDL: A Unified Multi-Distribution Learner in Large-scale Industrial Recommendation through Tokenization

Industrial recommender systems increasingly adopt multi-scenario learning (MSL) and multi-task learning (MTL) to handle diverse user interactions and contexts, but existing approaches suffer from two critical drawbacks: (1) underutilization of large-scale model parameters due to limited interaction with complex feature modules, and (2) difficulty in jointly modeling scenario and task information in a unified framework. To address these challenges, we propose a unified \textbf{M}ulti-\textbf{D}istribution \textbf{L}earning (MDL) framework, inspired by the "prompting" paradigm in large language models (LLMs). MDL treats scenario and task information as specialized tokens rather than auxiliary inputs or gating signals. Specifically, we introduce a unified information tokenization module that transforms features, scenarios, and tasks into a unified tokenized format. To facilitate deep interaction, we design three synergistic mechanisms: (1) feature token self-attention for rich feature interactions, (2) domain-feature attention for scenario/task-adaptive feature activation, and (3) domain-fused aggregation for joint distribution prediction. By stacking these interactions, MDL enables scenario and task information to "prompt" and activate the model's vast parameter space in a bottom-up, layer-wise manner. Extensive experiments on real-world industrial datasets demonstrate that MDL significantly outperforms state-of-the-art MSL and MTL baselines. Online A/B testing on Douyin Search platform over one month yields +0.0626\% improvement in LT30 and -0.3267\% reduction in change query rate. MDL has been fully deployed in production, serving hundreds of millions of users daily.

cs.IR↗
arXiv2026

SWE-Bench Mobile: Can Large Language Model Agents Develop Industry-Level Mobile Applications?

Can large language model agents develop industry-level mobile applications? We introduce \textbf{SWE-Bench Mobile}, a benchmark for evaluating coding agents on realistic software engineering tasks derived from a production iOS codebase. Unlike existing benchmarks that focus on isolated problems or bug fixes, SWE-Bench Mobile captures the full complexity of industrial development: multi-modal inputs (PRDs and Figma designs), a large-scale mixed Swift/Objective-C codebase, and comprehensive test suites. We evaluate 22 agent-model configurations across four coding agents -- three commercial (Cursor, Codex, Claude Code) and one open-source (OpenCode) -- and find that even the best configurations achieve only 12\% task success rate. Our analysis reveals that (1) agent design matters as much as model capability -- the same model shows up to 6$\times$ performance gap across agents, (2) commercial agents consistently outperform open-source alternatives, and (3) simple ``Defensive Programming'' prompts outperform complex ones by 7.4\%. These findings highlight a significant gap between current agent capabilities and industrial requirements, while providing actionable insights for practitioners and researchers. We release SWE-Bench Mobile as a \textit{hosted benchmark challenge} to prevent data contamination and ensure fair evaluation. The public leaderboard and development toolkit are available at https://swebenchmobile.com.

cs.SE↗
arXiv2026

MSN: A Memory-based Sparse Activation Scaling Framework for Large-scale Industrial Recommendation

Scaling deep learning recommendation models is an effective way to improve model expressiveness. Existing approaches often incur substantial computational overhead, making them difficult to deploy in large-scale industrial systems under strict latency constraints. Recent sparse activation scaling methods, such as Sparse Mixture-of-Experts, reduce computation by activating only a subset of parameters, but still suffer from high memory access costs and limited personalization capacity due to the large size and small number of experts. To address these challenges, we propose MSN, a memory-based sparse activation scaling framework for recommendation models. MSN dynamically retrieves personalized representations from a large parameterized memory and integrates them into downstream feature interaction modules via a memory gating mechanism, enabling fine-grained personalization with low computational overhead. To enable further expansion of the memory capacity while keeping both computational and memory access costs under control, MSN adopts a Product-Key Memory (PKM) mechanism, which factorizes the memory retrieval complexity from linear time to sub-linear complexity. In addition, normalization and over-parameterization techniques are introduced to maintain balanced memory utilization and prevent memory retrieval collapse. We further design customized Sparse-Gather operator and adopt the AirTopK operator to improve training and inference efficiency in industrial settings. Extensive experiments demonstrate that MSN consistently improves recommendation performance while maintaining high efficiency. Moreover, MSN has been successfully deployed in the Douyin Search Ranking System, achieving significant gains over deployed state-of-the-art models in both offline evaluation metrics and large-scale online A/B test.

cs.IR↗
arXiv2026

Phi-Former: A Pairwise Hierarchical Approach for Compound-Protein Interactions Prediction

Drug discovery remains time-consuming, labor-intensive, and expensive, often requiring years and substantial investment per drug candidate. Predicting compound-protein interactions (CPIs) is a critical component in this process, enabling the identification of molecular interactions between drug candidates and target proteins. Recent deep learning methods have successfully modeled CPIs at the atomic level, achieving improved efficiency and accuracy over traditional energy-based approaches. However, these models do not always align with chemical realities, as molecular fragments (motifs or functional groups) typically serve as the primary units of biological recognition and binding. In this paper, we propose Phi-former, a pairwise hierarchical interaction representation learning method that addresses this gap by incorporating the biological role of motifs in CPIs. Phi-former represents compounds and proteins hierarchically and employs a pairwise pre-training framework to model interactions systematically across atom-atom, motif-motif, and atom-motif levels, reflecting how biological systems recognize molecular partners. We design intra-level and inter-level learning pipelines that make different interaction levels mutually beneficial. Experimental results demonstrate that Phi-former achieves superior performance on CPI-related tasks. A case study shows that our method accurately identifies specific atoms or motifs activated in CPIs, providing interpretable model explanations. These insights may guide rational drug design and support precision medicine applications.

cs.AI↗
arXiv2026

Detecting underlying symmetry-protected topological phases via strange correlators and edge engineering

The vast majority of symmetry-protected topological (SPT) states are difficult to detect, which often leads to their misidentification as ordinary or topologically trivial phases. In this work, we propose a general framework for detecting these hidden topological states. We distinguish the ordinary matter state from the topological phase by exploiting the boundary effects in space (via surface behaviors on engineered edge) and time (via strange correlators) according to the principle of bulk-edge correspondence. As a concrete example, we study the dimerized spin-1/2 Heisenberg model on a square lattice using quantum Monte Carlo simulations, focusing on its paramagnetic dimer phase and edge states. The dimer phase has been widely regarded as topologically trivial due to its gapped edge state on conventional edges. However, the model can also be viewed as two-dimensional antiferromagnetically (AF) coupled usual ladders, which suggests an SPT state adiabatically connected to the one-dimensional Haldane phase. We resolve this puzzle and demonstrate that the dimer phase is indeed a quasi-one-dimensional SPT state by measuring generalized strange correlators introduced in this work and by showing that the nontrivial gapless edge state on a zigzag edge is ferromagnetically ordered, resulting from effective ferromagnetic interactions between degenerate spinons liberated on each side of the cut. Furthermore, we show that the ordered edge state gives rise to an extraordinary surface critical behavior at the (2+1)-dimensional O(3) bulk critical points of the model, which contradicts theoretical predictions based on classical-quantum mapping. Overall, we establish a standard detection method for uncovering topological phases that masquerade as ordinary states of matter.

cond-mat.str-el↗
arXiv2026

Statistics Approximation-Enabled Distributed Beamforming for Cell-Free Massive MIMO

We study a distributed beamforming approach for cell-free massive multiple-input multiple-output networks, referred to as Global Statistics & Local Instantaneous information-based minimum mean-square error (GSLI-MMSE). The scenario with multi-antenna access points (APs) is considered over three different channel models: correlated Rician fading with fixed or random line-of-sight (LoS) phase-shifts, and correlated Rayleigh fading. With the aid of matrix inversion derivations, we can construct the conventional MMSE combining from the perspective of each AP, where global instantaneous information is involved. Then, for an arbitrary AP, we apply the statistics approximation methodology to approximate instantaneous terms related to other APs by channel statistics to construct the distributed combining scheme at each AP with local instantaneous information and global statistics. With the aid of uplink-downlink duality, we derive the respective GSLI-MMSE precoding schemes. Numerical results showcase that the proposed GSLI-MMSE scheme demonstrates performance comparable to the optimal centralized MMSE scheme, under the stable LoS conditions, e.g., with static users having Rician fading with a fixed LoS path.

eess.SP↗
arXiv2026

Low-Complexity Distributed Combining Design for Near-Field Cell-Free XL-MIMO Systems

In this paper, we investigate the low-complexity distributed combining scheme design for near-field cell-free extremely large-scale multiple-input-multiple-output (CF XL-MIMO) systems. Firstly, we construct the uplink spectral efficiency (SE) performance analysis framework for CF XL-MIMO systems over centralized and distributed processing schemes. Notably, we derive the centralized minimum mean-square error (CMMSE) and local minimum mean-square error (LMMSE) combining schemes over arbitrary channel estimators. Then, focusing on the CMMSE and LMMSE combining schemes, we propose five low-complexity distributed combining schemes based on the matrix approximation methodology or the symmetric successive over relaxation (SSOR) algorithm. More specifically, we propose two matrix approximation methodology-aided combining schemes: Global Statistics \& Local Instantaneous information-based MMSE (GSLI-MMSE) and Statistics matrix Inversion-based LMMSE (SI-LMMSE). These two schemes are derived by approximating the global instantaneous information in the CMMSE combining and the local instantaneous information in the LMMSE combining with the global and local statistics information by asymptotic analysis and matrix expectation approximation, respectively. Moreover, by applying the low-complexity SSOR algorithm to iteratively solve the matrix inversion in the LMMSE combining, we derive three distributed SSOR-based LMMSE combining schemes, distinguished from the applied information and initial values.

eess.SP↗
arXiv2026

Training-Free Self-Correction for Multimodal Masked Diffusion Models

Masked diffusion models have emerged as a powerful framework for text and multimodal generation. However, their sampling procedure updates multiple tokens simultaneously and treats generated tokens as immutable, which may lead to error accumulation when early mistakes cannot be revised. In this work, we revisit existing self-correction methods and identify limitations stemming from additional training requirements or reliance on misaligned likelihood estimates. We propose a training-free self-correction framework that exploits the inductive biases of pre-trained masked diffusion models. Without modifying model parameters or introducing auxiliary evaluators, our method significantly improves generation quality on text-to-image generation and multimodal understanding tasks with reduced sampling steps. Moreover, the proposed framework generalizes across different masked diffusion architectures, highlighting its robustness and practical applicability. Code can be found in https://github.com/huge123/FreeCorrection.

stat.ML↗
arXiv2026

Detecting (emergent) continuous symmetry of criticality via subsystem's entanglement spectrum

The (emergent) symmetry of a critical point is one of the most important information to identify the universality class and effective field theory, which is fundamental for various critical theories. However, the underlying symmetry so far can only be conjectured indirectly from the dimension of the order parameters in symmetry-breaking phases, and its correctness requires further verifications to avoid overlooking hidden order parameters, which by itself is also a difficult task. In this work, we propose an unbiased way to numerically identify the underlying (emergent) symmetry of a critical point in quantum many-body systems, without prior knowledge about its low-energy effective field theory. Through calculating the reduced density matrix in a very small subsystem of the total system numerically, the Anderson tower of states in the entanglement spectrum clearly reflects the underlying (emergent) symmetry of the criticality. It is attributed to the fact that the entanglement spectrum can observe the broken symmetry of the entanglement ground-state after cooling from the critical point along an extra temperature axis.

quant-ph↗