arXiv subjects

Wenyu Wang

Publications and source records attributed to Wenyu Wang.

At least 19 recordsLinked to original sources

arXiv2026

Direct Detection of Light Self-Interacting Dark Matter via Electronic Collective Excitations

Models of light dark matter often invoke a light mediator to facilitate interactions with the Standard Model. If sufficiently light, this mediator can induce long-range self-interactions among dark matter particles, offering a compelling resolution to small-scale structure anomalies. However, direct detection of light self-interacting dark matter (SIDM) remains challenging for conventional detectors. In this work, we investigate the sensitivity of searches for light SIDM accelerated by high-energy cosmic rays in silicon detectors. Leveraging the electronic collective excitations, we derive 90\% C.L. exclusion limits using public SENSEI and DAMIC-M ionization data. Our constraints can cover a portion of the light SIDM parameter space favored by galactic small-scale anomalies.

hep-ph↗

arXiv2026

Towards Federated Long-Tailed Graph Learning: An Energy-Guided Dual Decoupling Approach

Federated Graph Learning facilitates collaborative graph modeling across distributed clients while preserving data privacy. However, real-world data categories frequently exhibit long-tailed distributions. Such statistical scarcity severely degrades performance in two ways: it biases the global model toward majority classes, and it structurally isolates minority nodes by submerging them in heterophilic, head-dominated neighborhoods. While existing methods attempt topology-agnostic statistical compensations, they often fail under data scarcity. Instead of recovering tail nodes, they overfit the structural noise from adjacent dominant classes, leading to representation degradation. To address these limitations, we propose FedEPD, a framework built on a dual decoupling paradigm that separates topological purification from semantic recalibration. Specifically, FedEPD utilizes distribution-aware Dirichlet energy pruning to filter spatial heterophilic edges. It then overcomes Non-IID distribution shifts by extracting robust global prototypes from topologically central nodes, which are incorporated into local representations via a spatial low-pass prototype injection. Furthermore, a two stage alternating optimization strategy strictly protects majority decision boundaries while improving minority accuracy. Extensive experiments demonstrate that FedEPD achieves state-of-the-art performance across diverse long-tailed benchmarks, yielding absolute improvements of up to 4.97% in Accuracy and 5.48% in Macro-F1.

cs.AI↗

arXiv2026

SP-TransientBench: A Real-Captured Single Photon Perception Benchmark

Single-photon LiDAR (SPL) based on single-photon avalanche diode (SPAD) sensing enables time-resolved photon measurements with extreme sensitivity, offering unique potential for active 3D perception in photon-starved scenarios.However, real-world single photon perception remains fundamentally challenging due to unique measurement noise and complex multi-return transient phenomena, which jointly complicate geometric reconstruction and semantic scene understanding. Despite growing interest in SPAD-based sensing, existing studies are largely limited to simulated data or small-scale controlled captures. As a result, systematic evaluation of real-world single photon perception across depth estimation, multi-view reconstruction, and 3D semantic understanding remains underexplored. To bridge this gap, we introduce SP-TransientBench (STB), a real-captured multi-task benchmark for single photon perception. SP-TransientBenc comprises 10 diverse scenes and 10,297 views captured using a solid-state single-photon LiDAR at $256\times192$ resolution. Each view provides full time-of-flight histograms with multi-return behavior,standardized metadata, and calibrated camera poses for multi-view evaluation. We further provide 13-class 3D semantic annotations for selected scenes. By providing dedicated data splits and evaluation protocols for each task, STB enables consistent and reproducible benchmarking of real-world single photon perception across multiple 3D vision problems. The dataset and code will be released upon acceptance.

cs.CV↗

arXiv2026

Conjectural Variations in Competitive Dynamic Pricing: A Learning Foundation via Experimentation Design and Feedback Structure

We study competitive dynamic pricing among multiple sellers, motivated by the rise of large-scale experimentation and algorithmic pricing in retail and online marketplaces. Sellers repeatedly set prices using simple learning rules and observe their own realized demand, while possibly observing only a subset of rivals' prices, even though demand depends on all sellers' prices and is subject to random shocks. Each seller runs local price experiments, such as switchback-style designs, and updates a focal price using a linear demand estimate fitted to its own demand data and the competitor prices it observes. Under certain conditions on demand, the resulting dynamics converge to a Conjectural Variations (CV) equilibrium, a classic static equilibrium notion in which each seller best responds under a conjecture that rivals' prices co-move systematically to changes in its own price. Unlike standard CV models that treat conjectures as behavioral primitives, we show that these conjectures arise endogenously from the interaction between the feedback structure and the correlation structure of experimentation. When a seller does not observe some rivals' prices, correlated experimentation induces an omitted-variable bias in demand estimation. We show that this bias determines the conjectures that govern the long-run equilibrium. Notably, when this learning bias vanishes, for example under full price feedback or independent experimentation of unobserved rivals, the learning dynamics converge to the standard Nash equilibrium. We provide simple sufficient conditions on demand for convergence in standard models and establish a finite-sample guarantee, showing that the mean squared price error decays at a rate of $\widetilde O (T^{-1/2})$.

cs.GT↗

arXiv2026

CoLLM: Continuous Adaptation for SLO-Aware LLM Serving on Shared GPU Clusters

As Large Language Models (LLMs) are increasingly adopted in edge intelligence to power domain-specific applications and personalized services, the quality and efficiency of the LLM post-training phase-including fine-tuning and inference, have become critical due to constrained resources. Although recent advances in federated parameter-efficient fine-tuning (FL PEFT) and low-latency inference have improved individual task performance, fine-tuning and inference are still handled as isolated workloads, which overlooks their interdependence and results in redundant deployments and delayed improvement in inference quality. To address these limitations, we introduce a new co-execution framework and instantiate it with CoLLM, a system that unifies FL PEFT and inference on shared edge replicas and model parameters. CoLLM addresses key challenges at both replica and cluster levels through: (1) an intra-replica model sharing mechanism that enables real-time model parameter reuse via unmerged inference and shadow adapter strategies; and (2) a two-timescale inter-replica coordination algorithm that adaptively balances fine-tuning and inference workloads to jointly optimize long-term model quality gains and short-term inference efficiency. Extensive evaluation across diverse LLMs and real-world traces show that CoLLM consistently outperforms state-of-the-art LLM systems, achieving up to 3x higher goodput, demonstrating its effectiveness in enabling seamless LLM post-training for edge intelligence.

cs.DC↗

arXiv2026

Generalized Category Discovery in Federated Graph Learning

Federated Graph Learning (FGL) enables collaborative learning over distributed graph data, yet existing approaches largely rely on a closed-world assumption, limiting their applicability in dynamic environments where novel categories continuously emerge. To bridge this gap, we target the practical scenario of Federated Graph Generalized Category Discovery (FGGCD), aiming to collaboratively discover novel categories across decentralized graph clients while retaining knowledge of known categories. We observe that FGGCD introduces two fundamental challenges: (1) the Neighborhood Absorption Effect, where structural fragmentation leads to biased neighborhood aggregation, causing novel nodes to be misclassified as known categories; and (2) Global Semantic Inconsistency, where the aforementioned local biases propagate to the server and are amplified by heterogeneous subgraph distributions, hindering cross-client knowledge integration. To address these issues, we propose GCD-FGL, an FGL framework for GCD that integrates a client-side Topology-Reliable Semantic Alignment and Discovery process to mitigate the neighborhood absorption effect, and a server-side Hierarchical Prototype Alignment strategy to resolve global semantic inconsistency. Extensive experiments on five real-world graph datasets demonstrate that GCD-FGL consistently outperforms state-of-the-art baselines, achieving an average absolute gain of +4.86 in HRScore.

cs.LG↗

arXiv2026

FineServe: A Fine-Grained Dataset and Characterization of Global LLM Serving Workloads

Large language models (LLMs) are increasingly deployed as always-on online services, making efficient LLM serving a critical systems challenge. Achieving low latency and high throughput under volatile demand requires deep understanding of real-world serving workloads, yet existing studies often rely on proxy traces or coarse-grained characterizations that fail to capture the heterogeneity of modern multi-model LLM platforms. We present FineServe, an in-the-wild, multi-model LLM serving workload dataset collected from a global commercial marketplace, enabling fine-grained characterization of real-world serving dynamics across heterogeneous models and tasks. Leveraging FineServe, we conduct a comprehensive analysis of arrival dynamics and token behavior, revealing fundamentally different fluctuation regimes across model architectures, scales and task intents. Building on these insights, we develop the FineServe workload generator, which composes fine-grained model-aware workloads into configurable mixtures tailored for benchmarking multi-model serving platforms. By exposing these fine-grained workload dynamics, FineServe provides a realistic foundation for evaluating routing, scheduling, and capacity-planning strategies in LLM serving systems. FineServe is available at https://github.com/hihiztc1/FineServe.

cs.AI↗

arXiv2026

Controllable Singing Style Conversion with Boundary-Aware Information Bottleneck

This paper presents the submission of the S4 team to the Singing Voice Conversion Challenge 2025 (SVCC2025)-a novel singing style conversion system that advances fine-grained style conversion and control within in-domain settings. To address the critical challenges of style leakage, dynamic rendering, and high-fidelity generation with limited data, we introduce three key innovations: a boundary-aware Whisper bottleneck that pools phoneme-span representations to suppress residual source style while preserving linguistic content; an explicit frame-level technique matrix, enhanced by targeted F0 processing during inference, for stable and distinct dynamic style rendering; and a perceptually motivated high-frequency band completion strategy that leverages an auxiliary standard 48kHz SVC model to augment the high-frequency spectrum, thereby overcoming data scarcity without overfitting. In the official SVCC2025 subjective evaluation, our system achieves the best naturalness performance among all submissions while maintaining competitive results in speaker similarity and technique control, despite using significantly less extra singing data than other top-performing systems. Audio samples are available online.

cs.SD↗

arXiv2026

Fractional Programming for Stochastic Precoding over Generalized Fading Channels

This paper seeks an efficient algorithm for stochastic precoding to maximize the long-term average weighted sum rates throughout a multiple-input multiple-output (MIMO) network. Unlike many existing works that assume a particular probability distribution model for fading channels (which is typically Gaussian), our approach merely relies on the first and second moments of fading channels. For the stochastic precoding problem, a naive idea is to directly apply the fractional programming (FP) method to the data rate inside the expectation; it does not work well because the auxiliary variables introduced by FP are then difficult to decide. To address the above issue, we propose using a lower bound to approximate the expectation of data rate. This lower bound stems from a nontrivial use of the matrix FP, and outperforms the existing lower bounds in that it accounts for generalized fading channels whose first and second moments are known. The resulting approximate problem can be efficiently solved in closed form in an iterative fashion. Furthermore, for large-scale MIMO, we improve the efficiency of the proposed algorithm by eliminating the large matrix inverse. Simulations show that the proposed stochastic precoding method outperforms the benchmark methods in both Gaussian and non-Gaussian fading channel cases.

cs.IT↗

arXiv2026

Causal Discovery for Cross-Sectional Data Based on Super-Structure and Divide-and-Conquer

This paper tackles a critical bottleneck in Super-Structure-based divide-and-conquer causal discovery: the high computational cost of constructing accurate Super-Structures--particularly when conditional independence (CI) tests are expensive and domain knowledge is unavailable. We propose a novel, lightweight framework that relaxes the strict requirements on Super-Structure construction while preserving the algorithmic benefits of divide-and-conquer. By integrating weakly constrained Super-Structures with efficient graph partitioning and merging strategies, our approach substantially lowers CI test overhead without sacrificing accuracy. We instantiate the framework in a concrete causal discovery algorithm and rigorously evaluate its components on synthetic data. Comprehensive experiments on Gaussian Bayesian networks, including magic-NIAB, ECOLI70, and magic-IRRI, demonstrate that our method matches or closely approximates the structural accuracy of PC and FCI while drastically reducing the number of CI tests. Further validation on the real-world China Health and Retirement Longitudinal Study (CHARLS) dataset confirms its practical applicability. Our results establish that accurate, scalable causal discovery is achievable even under minimal assumptions about the initial Super-Structure, opening new avenues for applying divide-and-conquer methods to large-scale, knowledge-scarce domains such as biomedical and social science research.

cs.LG↗

arXiv2025

The Lepton Flavor Changing Decays and One-loop Muon Anomalous Magnetic Moment in the Extended Mirror Twin Higgs Models

Mirror Twin Higgs(MTH) models always contain heavy gauge bosons and extra Higgses. Besides, to accommodate tiny neutrino masses via seesaw mechanism, new heavy neutrinos can also be introduced in MTH extension models. Such new particles and interactions may lead to new contributions to the lepton flavor violating (LFV) processes, including $\ell_i \to \ell_jγ$ and $\ell_i \to \ell_j\ell_k\ell_l$. We find that current experimental data can stringently constrain the parameter spaces and certain LFV processes can possibly be tested by the next generation colliders. One-loop contributions of the new particles to the muon anomalous magnetic momentumare also calculated. Such contributions can still not solve the discrepancy between the experiments and the prediction of the standard model.

hep-ph↗

arXiv2025

Search for Ultralight Dark Matter with Quantum Magnetometry in the Earth's Cavity

Ultralight dark matter candidates, such as axions and dark photons, are leading dark matter candidates. They may couple feebly to photons, sourcing oscillating electromagnetic signals in the Earth's conducting cavity formed between the ground and the ionosphere, providing detectable magnetic field signatures at wavelengths above the Earth's size. We carry out a project aiming to search for new physics using an unshielded high-sensitivity atomic magnetometer, termed the Geomagnetic Probe for nEw physiCS (GPEX). In this work, we report our first search for axion and dark photon dark matter, conducted in the desert of XiaoDushan in Gansu Province, China. Analysis of the collection of one-hour data shows no robust evidence for axion- or dark photon-induced magnetic signals. Correspondingly, we set the constraints on the axion-photon coupling with $g_{aγγ} < 7\times10^{-10}\, \mathrm{GeV^{-1}}$ and the dark photon kinetic-mixing parameter $ε< 2\times10^{-6}$ in the mass range $3.5 \times 10^{-16}\, \mathrm{eV} \sim 1.8 \times 10^{-14}\, \mathrm{eV}$. Our findings demonstrate the feasibility of using ground-based quantum magnetic sensors for ultralight dark matter searches. Future networks of such detectors operating over extended periods could improve the sensitivity by about three orders of magnitude.

hep-ph↗

arXiv2025

Eigenstate solutions of the Fermi-Hubbard model via symmetry-enhanced variational quantum eigensolver

The Variational Quantum Eigensolver (VQE), as a hybrid quantum-classical algorithm, is an important tool for effective quantum computing in the current noisy intermediate-scale quantum (NISQ) era. However, the traditional hardware-efficient ansatz without taking into account symmetries requires more computational resources to explore the unnecessary regions in the Hilbert space. The conventional Subspace-Search VQE (SSVQE) algorithm, which can calculate excited states, is also unable to effectively handle degenerate states since the loss function only contains the expectation value of the Hamiltonian. In this study, the energy eigenstates of the one-dimensional Fermi-Hubbard model with two lattice sites and the two-dimensional Hubbard model with four lattice sites are calculated. By incorporating symmetries into the quantum circuits and loss function, we find that both the ground state and excited state calculations are improved greatly compared to the case without symmetries. The enhancement in excited state calculations is particularly significant. This is because quantum circuits that conserve the particle number are used, and appropriate penalty terms are added to the loss function, enabling the optimization process to correctly identify degenerate states. The results are verified through repeated simulations.

cond-mat.str-el↗

arXiv2025

FabasedVC: Enhancing Voice Conversion with Text Modality Fusion and Phoneme-Level SSL Features

In voice conversion (VC), it is crucial to preserve complete semantic information while accurately modeling the target speaker's timbre and prosody. This paper proposes FabasedVC to achieve VC with enhanced similarity in timbre, prosody, and duration to the target speaker, as well as improved content integrity. It is an end-to-end VITS-based VC system that integrates relevant textual modality information, phoneme-level self-supervised learning (SSL) features, and a duration predictor. Specifically, we employ a text feature encoder to encode attributes such as text, phonemes, tones and BERT features. We then process the frame-level SSL features into phoneme-level features using two methods: average pooling and attention mechanism based on each phoneme's duration. Moreover, a duration predictor is incorporated to better align the speech rate and prosody of the target speaker. Experimental results demonstrate that our method outperforms competing systems in terms of naturalness, similarity, and content integrity.

cs.SD↗

arXiv2025

Latent Zoning Network: A Unified Principle for Generative Modeling, Representation Learning, and Classification

Generative modeling, representation learning, and classification are three core problems in machine learning (ML), yet their state-of-the-art (SoTA) solutions remain largely disjoint. In this paper, we ask: Can a unified principle address all three? Such unification could simplify ML pipelines and foster greater synergy across tasks. We introduce Latent Zoning Network (LZN) as a step toward this goal. At its core, LZN creates a shared Gaussian latent space that encodes information across all tasks. Each data type (e.g., images, text, labels) is equipped with an encoder that maps samples to disjoint latent zones, and a decoder that maps latents back to data. ML tasks are expressed as compositions of these encoders and decoders: for example, label-conditional image generation uses a label encoder and image decoder; image embedding uses an image encoder; classification uses an image encoder and label decoder. We demonstrate the promise of LZN in three increasingly complex scenarios: (1) LZN can enhance existing models (image generation): When combined with the SoTA Rectified Flow model, LZN improves FID on CIFAR10 from 2.76 to 2.59-without modifying the training objective. (2) LZN can solve tasks independently (representation learning): LZN can implement unsupervised representation learning without auxiliary loss functions, outperforming the seminal MoCo and SimCLR methods by 9.3% and 0.2%, respectively, on downstream linear classification on ImageNet. (3) LZN can solve multiple tasks simultaneously (joint generation and classification): With image and label encoders/decoders, LZN performs both tasks jointly by design, improving FID and achieving SoTA classification accuracy on CIFAR10. The code and trained models are available at https://github.com/microsoft/latent-zoning-networks. The project website is at https://zinanlin.me/blogs/latent_zoning_networks.html.

cs.LG↗

arXiv2025

Stratos: An End-to-End Distillation Pipeline for Customized LLMs under Distributed Cloud Environments

The growing industrial demand for customized and cost-efficient large language models (LLMs) is fueled by the rise of vertical, domain-specific tasks and the need to optimize performance under constraints such as latency and budget. Knowledge distillation, as an efficient model compression and transfer technique, offers a feasible solution. However, existing distillation frameworks often require manual intervention and struggle to meet such complex user-defined distillation requirements. To bridge this gap, we propose Stratos, an end-to-end LLM distillation pipeline that automates server and model selection, knowledge distillation, and deployment in distributed cloud environments. Given user-defined constraints on model performance and system budget, Stratos automatically selects Pareto-optimal servers, dynamically matches teacher-student pairs, and adapts distillation strategies based on task complexity to optimize cloud hosting. Experiments show that Stratos produces a student model that achieves four times the accuracy of its GPT-4o teacher baseline on a rare, domain-specific Mahjong reasoning task with reverse synthetic data and knowledge injection. Moreover, it achieves reduced latency and cost without compromising accuracy. These results highlight its promise for vertical-domain LLM deployment.

cs.LG↗

arXiv2025

Dark Photon Oscillations in Waveguide

Dark photons, which can kinetically mix with ordinary photons, represent the simplest extension to the standard model. Detecting their oscillations with visible photons could provide crucial insights into the nature of dark matter and fundamental interactions beyond the standard model. We propose a novel laboratory-based approach to detect dark photon oscillations using a laser in an Optical Time-domain Relectometry (OTDR) setup. The laser light propagating through the optical fiber undergoes oscillations with the dark photon, leading to measurable changes in the power flow. These oscillations can precisely measured, leveraging its high sensitivity and efficiency in detecting small variations in the optical signal. This approach could provide a new avenue for probing dark photon oscillations in the laboratory and greatly improve the current experimental sensitivity to dark photon in a wide mass range.

hep-ph↗

arXiv2025

Leveraging Geometric Visual Illusions as Perceptual Inductive Biases for Vision Models

Contemporary deep learning models have achieved impressive performance in image classification by primarily leveraging statistical regularities within large datasets, but they rarely incorporate structured insights drawn directly from perceptual psychology. To explore the potential of perceptually motivated inductive biases, we propose integrating classic geometric visual illusions well-studied phenomena from human perception into standard image-classification training pipelines. Specifically, we introduce a synthetic, parametric geometric-illusion dataset and evaluate three multi-source learning strategies that combine illusion recognition tasks with ImageNet classification objectives. Our experiments reveal two key conceptual insights: (i) incorporating geometric illusions as auxiliary supervision systematically improves generalization, especially in visually challenging cases involving intricate contours and fine textures; and (ii) perceptually driven inductive biases, even when derived from synthetic stimuli traditionally considered unrelated to natural image recognition, can enhance the structural sensitivity of both CNN and transformer-based architectures. These results demonstrate a novel integration of perceptual science and machine learning and suggest new directions for embedding perceptual priors into vision model design.

cs.CV↗