arXiv subjects

Xin Zhou

Publications and source records attributed to Xin Zhou.

At least 91 records · Page 5Linked to original sources

arXiv2026

SP^2DPO: An LLM-assisted Semantic Per-Pair DPO Generalization

Direct Preference Optimization (DPO) controls the trade-off between fitting preference labels and staying close to a reference model using a single global temperature beta, implicitly treating all preference pairs as equally informative. Real-world preference corpora are heterogeneous: they mix high-signal, objective failures (for example, safety, factuality, instruction violations) with low-signal or subjective distinctions (for example, style), and also include label noise. We introduce our method, SP2DPO (Semantic Per-Pair DPO), a generalization that replaces the global temperature with an instance-specific schedule beta_i pre-decided offline from structured semantic-gap annotations (category, magnitude, confidence) produced by teacher language models. We instantiate this procedure on the UltraFeedback preference corpus (59,960 pairs), enabling large-scale construction of an auditable beta_i artifact, and incur zero training-time overhead: the inner-loop optimizer remains standard DPO with beta set per pair. We focus our empirical study on AlpacaEval 2.0, reporting both raw win rate and length-controlled win rate. Across four open-weight, instruction-tuned student backbones (4B-8B), SP2DPO is competitive with a tuned global-beta DPO baseline and improves AlpacaEval 2.0 length-controlled win rate on two of four backbones, while avoiding per-model beta sweeps. All code, annotations, and artifacts will be released.

cs.CL↗

arXiv2026

FedUTR: Federated Recommendation with Augmented Universal Textual Representation for Sparse Interaction Scenarios

Federated recommendations (FRs) have emerged as an on-device privacy-preserving paradigm, attracting considerable attention driven by rising demands for data security. Existing FRs predominantly adapt ID embeddings to represent items, making the quality of item embeddings entirely dependent on users' historical behaviors. However, we empirically observe that this pattern leads to suboptimal recommendation performance under high data sparsity scenarios, due to its strong reliance on historical interactions. To address this issue, we propose a novel method named FedUTR, which incorporates item textual representations as a complement to interaction behaviors, aiming to enhance model performance under high data sparsity. Specifically, we utilize textual modality as the universal representation to capture generic item knowledge, and design a Collaborative Information Fusion Module (CIFM) to complement each user's personalized interaction information. Besides, we introduce a Local Adaptation Module (LAM) that adaptively exploits the off-the-shelf local model to efficiently preserve client-specific personalized preferences. Moreover, we propose a variant of FedUTR, termed FedUTR-SAR, which incorporates a sparsity-aware resnet component to granularly balance universal and personalized information. The convergence analysis proves theoretical guarantees for the effectiveness of FedUTR. Extensive experiments on four real-world datasets show that our method achieves superior performance, with improvements of up to 59% across all datasets compared to the SOTA baselines.

cs.IR↗

arXiv2026

SLM-SS: Speech Language Model for Generative Speech Separation

Speech separation (SS) has advanced significantly with neural network-based methods, showing improved performance on signal-level metrics. However, these methods often struggle to maintain speech intelligibility in the separated signals, which can negatively affect the performance of downstream tasks such as speech recognition. In this work, we propose SLM-SS, a novel approach that applies speech language models to SS, aiming to enhance the intelligibility and coherence of the separated signals. We frame SS as discrete multi-codebook sequence generation, using Encoder-Decoder models to map quantized speech mixtures to target tokens. In addition to the autoregressive modeling strategy, we introduce a non-autoregressive model to improve decoding efficiency for residual tokens. Experimental results on the LibriMix dataset demonstrate that our approach shows significantly better preservation of speech intelligibility, leading to improved linguistic consistency in a variety of downstream tasks compared to existing approaches.

cs.SD↗

arXiv2026

Transportability of Regression Calibration with External Validation Studies for Measurement Error Correction

In nutritional and environmental epidemiology, exposures are impractical to measure accurately, while practical measures for these exposures are often subject to substantial measurement error. Regression calibration is among the most used measurement error correction methods with external validation studies. The use of external studies to assess the measurement error process always carries the risk of introducing estimation bias into the main study analysis. Although the transportability of regression calibration is usually assumed for practical epidemiology studies, it has not been well studied. In this work, under the measurement error process with a mixture of Berkson-like and classical-like errors, we investigate conditions under which the effect estimate from regression calibration with an external validation study is unbiased for the association between exposure and health outcome. We further examine departures from the transportability assumption, under which the regression calibration estimator is itself biased. However, we theoretically prove that, in most cases, it yields lower bias than the naive method. The derived conditions are confirmed through simulation studies and further verified in an example investigating the association between the risk of cardiovascular disease and moderate physical activity in the health professional follow-up study.

stat.ME↗

arXiv2026

Out of Distribution, Out of Luck: How Well Can LLMs Trained on Vulnerability Datasets Detect Top 25 CWE Weaknesses?

Automated vulnerability detection research has made substantial progress, yet its real-world impact remains limited. Prior work found that current vulnerability datasets suffer from issues including label inaccuracy rates of 20%-71%, extensive duplication, and poor coverage of critical Common Weakness Enumeration (CWE). These issues create a significant generalization gap where models achieve misleading In-Distribution (ID) accuracies (testing on splits from the same dataset) by exploiting spurious correlations rather than learning true vulnerability patterns. To address these limitations, we present a three-part solution. First, we introduce BenchVul, which is a manually curated and balanced test dataset covering the MITRE Top 25 Most Dangerous CWEs, to enable fair model evaluation. Second, we construct a high-quality training dataset, TitanVul, comprising 38,548 functions by aggregating seven public sources and applying deduplication and validation using a novel multi-agent LLM pipeline. Third, we propose a Realistic Vulnerability Generation (RVG) pipeline, which synthesizes context-aware vulnerability examples for underrepresented but critical CWE types through simulated development workflows. Our evaluation reveals that In-Distribution (ID) performance does not reliably predict Out-of-Distribution (OOD) performance on BenchVul. For example, a model trained on BigVul achieves the highest 0.703 ID accuracy but fails on BenchVul's real-world samples (0.493 OOD accuracy). Conversely, a model trained on our TitanVul achieves the highest OOD performance on both the real-world (0.881) and synthesized (0.785) portions of BenchVul, improving upon the next-best performing dataset by 5.3% and 11.8% respectively, despite a modest ID score (0.590). Augmenting TitanVul with our RVG further boosts this leading OOD performance, improving accuracy on real-world data by 5.8% (to 0.932).

cs.CR↗

arXiv2026

TransFR: Transferable Federated Recommendation with Adapter Tuning on Pre-trained Language Models

Federated recommendations (FRs), facilitating multiple local clients to collectively learn a global model without disclosing user private data, have emerged as a prevalent on-device service. In conventional FRs, a dominant paradigm is to utilize discrete identities to represent clients and items, which are then mapped to domain-specific embeddings to participate in model training. Despite considerable performance, we reveal three inherent limitations that can not be ignored in federated settings, i.e., non-transferability across domains, ineffectiveness in cold-start settings, and potential privacy violations during federated training. To this end, we propose a transferable federated recommendation model, TransFR, which delicately incorporates the general capabilities empowered by pre-trained models and the personalized abilities by fine-tuning local private data. Specifically, it first learns domain-agnostic representations of items by exploiting pre-trained models with public textual corpora. To tailor for FR tasks, we further introduce efficient federated adapter-tuning and test-time adaptation mechanisms, which facilitate personalized local adapters for each client by fitting their private data distributions. We theoretically prove the advantages of incorporating adapter tuning in FRs regarding both effectiveness and privacy. Through extensive experiments, we show that our TransFR model surpasses several state-of-the-art FRs on transferability.

cs.IR↗

arXiv2026

Single-orientation Crystalline Domains of Active Brownian Particles Lead to Collective Motions

Active Brownian particles, even without attractive and anisotropic inter-particle interactions, can form a high-density phase featuring structure-ordered domains as well as collective motion regions under thermal noise. However, the mechanism, particularly the relationship between the motion and structure, remains unclear. In this study, we show that the motion-correlation regions are spatially coincident with the single-orientation crystalline domains. Each domain translates or rotates as a whole due to the net active force or torque acting upon it, allowing relative motions between these crystalline domains. The particles at domain boundaries usually have the active forces pointing inward, which helps to stabilize these structure-ordered domains and their corresponding collective motion regions.

cond-mat.soft↗

arXiv2026

Low-Dimensional Federated Knowledge Graph Embedding via Knowledge Distillation

Federated Knowledge Graph Embedding (FKGE) aims to facilitate collaborative learning of entity and relation embeddings from distributed Knowledge Graphs (KGs) across multiple clients, while preserving data privacy. Training FKGE models with higher dimensions is typically favored due to their potential for achieving superior performance. However, high-dimensional embeddings present significant challenges in terms of storage resource and inference speed. Unlike traditional KG embedding methods, FKGE involves multiple client-server communication rounds, where communication efficiency is critical. Existing embedding compression methods for traditional KGs may not be directly applicable to FKGE as they often require multiple model trainings which potentially incur substantial communication costs. In this paper, we propose a light-weight component based on Knowledge Distillation (KD) which is titled FedKD and tailored specifically for FKGE methods. During client-side local training, FedKD facilitates the low-dimensional student model to mimic the score distribution of triples from the high-dimensional teacher model using KL divergence loss. Unlike traditional KD way, FedKD adaptively learns a temperature to scale the score of positive triples and separately adjusts the scores of corresponding negative triples using a predefined temperature, thereby mitigating teacher over-confidence issue. Furthermore, we dynamically adjust the weight of KD loss to optimize the training process. Extensive experiments on three datasets support the effectiveness of FedKD.

cs.AI↗

arXiv2025

USE: A Unified Model for Universal Sound Separation and Extraction

Sound separation (SS) and target sound extraction (TSE) are fundamental techniques for addressing complex acoustic scenarios. While existing SS methods struggle with determining the unknown number of sound sources, TSE approaches require precisely specified clues to achieve optimal performance. This paper proposes a unified framework that synergistically combines SS and TSE to overcome their individual limitations. Our architecture employs two complementary components: 1) An Encoder-Decoder Attractor (EDA) network that automatically infers both the source count and corresponding acoustic clues for SS, and 2) A multi-modal fusion network that precisely interprets diverse user-provided clues (acoustic, semantic, or visual) for TSE. Through joint training with cross-task consistency constraints, we establish a unified latent space that bridges both paradigms. During inference, the system adaptively operates in either fully autonomous SS mode or clue-driven TSE mode. Experiments demonstrate remarkable performance in both tasks, with notable improvements of 1.4 dB SDR improvement in SS compared to baseline and 86\% TSE accuracy.

eess.AS↗

arXiv2025

VLA-AN: An Efficient and Onboard Vision-Language-Action Framework for Aerial Navigation in Complex Environments

This paper proposes VLA-AN, an efficient and onboard Vision-Language-Action (VLA) framework dedicated to autonomous drone navigation in complex environments. VLA-AN addresses four major limitations of existing large aerial navigation models: the data domain gap, insufficient temporal navigation with reasoning, safety issues with generative action policies, and onboard deployment constraints. First, we construct a high-fidelity dataset utilizing 3D Gaussian Splatting (3D-GS) to effectively bridge the domain gap. Second, we introduce a progressive three-stage training framework that sequentially reinforces scene comprehension, core flight skills, and complex navigation capabilities. Third, we design a lightweight, real-time action module coupled with geometric safety correction. This module ensures fast, collision-free, and stable command generation, mitigating the safety risks inherent in stochastic generative policies. Finally, through deep optimization of the onboard deployment pipeline, VLA-AN achieves a robust real-time 8.3x improvement in inference throughput on resource-constrained UAVs. Extensive experiments demonstrate that VLA-AN significantly improves spatial grounding, scene reasoning, and long-horizon navigation, achieving a maximum single-task success rate of 98.1%, and providing an efficient, practical solution for realizing full-chain closed-loop autonomy in lightweight aerial robots.

cs.RO↗

arXiv2025

Step-GUI Technical Report

Recent advances in multimodal large language models unlock unprecedented opportunities for GUI automation. However, a fundamental challenge remains: how to efficiently acquire high-quality training data while maintaining annotation reliability? We introduce a self-evolving training pipeline powered by the Calibrated Step Reward System, which converts model-generated trajectories into reliable training signals through trajectory-level calibration, achieving >90% annotation accuracy with 10-100x lower cost. Leveraging this pipeline, we introduce Step-GUI, a family of models (4B/8B) that achieves state-of-the-art GUI performance (8B: 80.2% AndroidWorld, 48.5% OSWorld, 62.6% ScreenShot-Pro) while maintaining robust general capabilities. As GUI agent capabilities improve, practical deployment demands standardized interfaces across heterogeneous devices while protecting user privacy. To this end, we propose GUI-MCP, the first Model Context Protocol for GUI automation with hierarchical architecture that combines low-level atomic operations and high-level task delegation to local specialist models, enabling high-privacy execution where sensitive data stays on-device. Finally, to assess whether agents can handle authentic everyday usage, we introduce AndroidDaily, a benchmark grounded in real-world mobile usage patterns with 3146 static actions and 235 end-to-end tasks across high-frequency daily scenarios (8B: static 89.91%, end-to-end 52.50%). Our work advances the development of practical GUI agents and demonstrates strong potential for real-world deployment in everyday digital interactions.

cs.CV↗

arXiv2025

Flying in Clutter on Monocular RGB by Learning in 3D Radiance Fields with Domain Adaptation

Modern autonomous navigation systems predominantly rely on lidar and depth cameras. However, a fundamental question remains: Can flying robots navigate in clutter using solely monocular RGB images? Given the prohibitive costs of real-world data collection, learning policies in simulation offers a promising path. Yet, deploying such policies directly in the physical world is hindered by the significant sim-to-real perception gap. Thus, we propose a framework that couples the photorealism of 3D Gaussian Splatting (3DGS) environments with Adversarial Domain Adaptation. By training in high-fidelity simulation while explicitly minimizing feature discrepancy, our method ensures the policy relies on domain-invariant cues. Experimental results demonstrate that our policy achieves robust zero-shot transfer to the physical world, enabling safe and agile flight in unstructured environments with varying illumination.

cs.RO↗

arXiv2025

The Milky Way Imaging Scroll Painting Survey: Data Release 1

We present the first data release (DR1) of the Milky Way Imaging Scroll Painting (MWISP) survey, a mapping in the J=(1-0) transition lines of 12CO, 13CO, and C18O toward the northern Galactic plane during 2011-2022. The MWISP survey was conducted using the PMO 13.7 m telescope at a spatial resolution of approximately 50" and a velocity resolution of 0.16 km/s at 115 GHz. DR1 fully covered 2310 square degrees within the Galactic longitude (l) and latitude (b) range of 9.75 deg =< l =< 229.75 deg and |b| =< 5.25 deg. The surveyed area was divided into cell units of 30'x30' for practical purposes and On-The-Fly (OTF) mapping was performed toward each target cell unit. The data were regridded into a regular 3D datacube in l-b-V_LSR with a pixel size of 30" in l-b axes and 0.16 km/s in theV_LSR axis. The median rms noise is 0.47 K, 0.25 K, and 0.25 K for 12CO, 13CO, and C18O, respectively. The equivalent 3 sigma sensitivity in 12CO luminosity is approximately 0.23 K km/s, making MWISP the most sensitive survey of its kind. In this paper, we describe the survey data, including the calibration, data cleaning, data mosaic processes, and the data products. The final mosaicked data cubes contain about 3.33x10^7 spectra (pixels) for each CO isotopologue line. Color composite images, made from the intensities of the isotopologue lines, and some concise descriptions are provided. We constructed a molecular cloud catalog based on the mosaicked 12CO data cube using the clustering algorithm DBSCAN, detecting 103,517 molecular clouds, 10,790 of which exhibit 13CO emission and 304 of which show C18O emission. Based on the histogram of voxel brightness temperature, we estimated a total 12CO flux of 7.69+/-0.38x10^7 K km/s arcmin^2, 82% of which is captured by the DBSCAN algorithm. The data, together with the cloud sample, provide unique information on molecular gas in the northern Milky Way.

astro-ph.GA↗

arXiv2025

Perfect continuous-variable quantum microcombs

Quantum microcombs generated in high-Q microresonators provide compact, multiplexed sources of entangled modes for continuous-variable (CV) quantum information processing. While deterministic generation of CV states via Kerr-induced two-mode squeezing has been demonstrated, achieving spectrally uniform squeezing remains challenging because of asymmetry and anomalies in the dispersion profile. Here we overcome these limitations by combining a microresonator with an engineered mode spectrum and optimized pump conditions. We realize a CV quantum microcomb comprising 14 independent two-mode squeezed states, each exhibiting more than 4 dB of raw squeezing (up to 4.3 dB) across a 0.7 THz bandwidth. This uniform, high-performance quantum resource represents a key step toward scalable, integrated CV quantum technologies operating beyond classical limits.

quant-ph↗

arXiv2025

Boost of critical current density near quantum critical points in FeSe-Based superconductors with two superconducting domes

Recent studies have identified two superconducting domes in FeSe-based superconductors. It was discovered that each dome is accompanied by a distinct nematic quantum critical point (QCP): one associated with a pure nematic QCP, and the other with a nematic QCP entangled with antiferromagnetism (AFM). In this study, we delve into the evolution of the critical current density ($J_{\rm{c}}$) with doping in FeSe${_{1-x}}$(Te/S)${_{x}}$ single crystals, focusing on the behavior within the two superconducting domes. Surprisingly, three maxima of $J_{\rm{c}}$ were found in the two superconducting domes, with two sharp peaks in $J_{\rm{c}}$ observed precisely at the endpoints of the nematic phases, at $x$(Te) $\sim$ 0.5 for Te-doped and $x$(S) $\sim$ 0.17 for S-doped FeSe. The mechanisms of vortex pinning and the influence of quantum critical fluctuations have been extensively explored, emphasizing the contribution of quantum critical fluctuations in modulating $J_{\rm{c}}$. Additionally, an increase in $J_{\rm{c}}$ was also noted near FeSe$_{0.1}$Te$_{0.9}$, where its origin has been explored and discussed. This finding provides crucial clues about the existence of an ordered phase endpoint beneath the superconducting dome, offering an initial basis for further investigation into the potential presence of a QCP beneath it.

cond-mat.supr-con↗

arXiv2025

Increase of critical current density in FeSe superconductor by strain effect

Conventional $J_c$-enhancement methods like doping and irradiation often introduce extrinsic elements or defects, altering intrinsic properties. Here, we report a significant $J_c$ enhancement in FeSe single crystals through compressive strain applied using a glass-fiber-reinforced plastic substrate with anisotropic thermal contraction during cooling. Under zero field at 2 K, $J_{\text{c}}$ increases by a factor of $\sim$4 from $\sim 2.3 \times 10^{4}$ to $\sim 8.7 \times 10^{4}$ A cm$^{-2}$; at 5 T, it achieves an order-of-magnitude enhancement, rising from $\sim 1.0 \times 10^{3}$ to $\sim 1.0 \times 10^{4}$ A cm$^{-2}$. Analysis based on the Dew-Hughes model of the $f_{\text{p}}$(h) relationship shows that strain strengthens vortex pinning, and shifts the pinning mechanism from point-like pinning to combined point and surface pinnings. This work offers an effective method to enhance FeSe's current-carrying limitation, deepens understanding of iron-based superconductors' pinning mechanisms, and highlights strain engineering's potential for optimizing superconducting performance.

cond-mat.supr-con↗

arXiv2025

Quantitative calibration of a TWPA applied to an optomechanical platform

In the last decade, the microwave quantum electronics toolbox has been enriched with quantum-limited detection devices such as Traveling Wave Parametric Amplifiers (TWPAs). The extreme sensitivity they provide is not only mandatory for some physics applications within quantum information processing, but is also the key element that will determine the detection limit of quantum sensing setups. In the framework of microwave optomechanical systems, an unprecedented range of small motions and forces is accessible, for which a specific quantitative calibration becomes necessary. We report on near quantum-limited measurements performed with an aluminum drumhead mechanical device within the temperature range 4 mK - 400 mK. The whole setup is carefully calibrated, especially taking into account the power-dependence of microwave absorption in the superconducting optomechanical cavity. This effect is commonly attributed to Two-Level-Systems (TLSs) present in the metal oxide. We demonstrate that a similar feature exists in the TWPA, and can be phenomenologically fit with adapted expressions. If not taken into account, the error on the signal strength can be as large as a factor of about 2, which is unacceptable for quantitative experiments. The power and temperature dependence is studied over the full parameter range, leading to an absolute definition of phonon population (i.e. Brownian motion amplitude), with an uncertainty +- 20 % limited by sources of noise internal to the optomechanical element.

quant-ph↗

arXiv2025

Pairs of Embedded Spheres with Pinched Prescribed Mean Curvature

Assume $h$ is a positive function on the unit three-sphere which satisfies the pinching condition $h < h_0 \approx 0.547$. We prove the existence of at least two embedded two-spheres with prescribed mean curvature $h$. The same result holds for sign-changing functions $h$ satisfying $\vert h\vert < h_0$ under a mild assumption on the zero set.

math.DG↗