Searcharxiv⌕ Search

arXiv subjects

Xuan Wen

Publications and source records attributed to Xuan Wen.

11 recordsLinked to original sources
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

Impact of Evaporation Barriers on Solar-Captured Dark Matter Distribution and Evaporation Mass

Evaporation determines the low-mass reach of solar-captured dark matter because that reach is controlled by the small population of particles closest to the escape threshold. We present an orbit-space calculation of the non-thermal distribution of captured dark matter in the presence of an evaporation barrier generated by a smooth in-medium attraction sourced by the solar medium. We show that the barrier not only deepens the effective potential but also reshapes the near-threshold phase-space structure, displacing the equilibrium distribution away from weakly bound, escape-prone trajectories and toward more tightly bound core-crossing orbits, thereby suppressing evaporation and lowering the evaporation mass. Although the bulk population remains near thermal equilibrium, the near-threshold tail, as reflected in the projected velocity spectrum, acquires characteristic non-thermal structure because the barrier deforms the bound orbit space and preferentially retains particles that repeatedly traverse the hot solar core. The near-threshold tail is therefore essential for determining the low-mass reach of solar dark-matter searches in the barrier regime, and our orbit-space treatment captures the relevant physics in a controlled way.

hep-ph↗
arXiv2025

Step-Audio 2 Technical Report

This paper presents Step-Audio 2, an end-to-end multi-modal large language model designed for industry-strength audio understanding and speech conversation. By integrating a latent audio encoder and reasoning-centric reinforcement learning (RL), Step-Audio 2 achieves promising performance in automatic speech recognition (ASR) and audio understanding. To facilitate genuine end-to-end speech conversation, Step-Audio 2 incorporates the generation of discrete audio tokens into language modeling, significantly enhancing its responsiveness to paralinguistic information such as speaking styles and emotions. To effectively leverage the rich textual and acoustic knowledge in real-world data, Step-Audio 2 integrates retrieval-augmented generation (RAG) and is able to call external tools such as web search to mitigate hallucination and audio search to switch timbres. Trained on millions of hours of speech and audio data, Step-Audio 2 delivers intelligence and expressiveness across diverse conversational scenarios. Evaluation results demonstrate that Step-Audio 2 achieves state-of-the-art performance on various audio understanding and conversational benchmarks compared to other open-source and commercial solutions. Please visit https://github.com/stepfun-ai/Step-Audio2 for more information.

cs.CL↗
arXiv2025

Step-Audio: Unified Understanding and Generation in Intelligent Speech Interaction

Real-time speech interaction, serving as a fundamental interface for human-machine collaboration, holds immense potential. However, current open-source models face limitations such as high costs in voice data collection, weakness in dynamic control, and limited intelligence. To address these challenges, this paper introduces Step-Audio, the first production-ready open-source solution. Key contributions include: 1) a 130B-parameter unified speech-text multi-modal model that achieves unified understanding and generation, with the Step-Audio-Chat version open-sourced; 2) a generative speech data engine that establishes an affordable voice cloning framework and produces the open-sourced lightweight Step-Audio-TTS-3B model through distillation; 3) an instruction-driven fine control system enabling dynamic adjustments across dialects, emotions, singing, and RAP; 4) an enhanced cognitive architecture augmented with tool calling and role-playing abilities to manage complex tasks effectively. Based on our new StepEval-Audio-360 evaluation benchmark, Step-Audio achieves state-of-the-art performance in human evaluations, especially in terms of instruction following. On open-source benchmarks like LLaMA Question, shows 9.3% average performance improvement, demonstrating our commitment to advancing the development of open-source multi-modal language technologies. Our code and models are available at https://github.com/stepfun-ai/Step-Audio.

cs.CL↗
arXiv2024

GrabDAE: An Innovative Framework for Unsupervised Domain Adaptation Utilizing Grab-Mask and Denoise Auto-Encoder

Unsupervised Domain Adaptation (UDA) aims to adapt a model trained on a labeled source domain to an unlabeled target domain by addressing the domain shift. Existing Unsupervised Domain Adaptation (UDA) methods often fall short in fully leveraging contextual information from the target domain, leading to suboptimal decision boundary separation during source and target domain alignment. To address this, we introduce GrabDAE, an innovative UDA framework designed to tackle domain shift in visual classification tasks. GrabDAE incorporates two key innovations: the Grab-Mask module, which blurs background information in target domain images, enabling the model to focus on essential, domain-relevant features through contrastive learning; and the Denoising Auto-Encoder (DAE), which enhances feature alignment by reconstructing features and filtering noise, ensuring a more robust adaptation to the target domain. These components empower GrabDAE to effectively handle unlabeled target domain data, significantly improving both classification accuracy and robustness. Extensive experiments on benchmark datasets, including VisDA-2017, Office-Home, and Office31, demonstrate that GrabDAE consistently surpasses state-of-the-art UDA methods, setting new performance benchmarks. By tackling UDA's critical challenges with its novel feature masking and denoising approach, GrabDAE offers both significant theoretical and practical advancements in domain adaptation.

cs.CV↗
arXiv2022

Multiple EffNet/ResNet Architectures for Melanoma Classification

Melanoma is the most malignant skin tumor and usually cancerates from normal moles, which is difficult to distinguish benign from malignant in the early stage. Therefore, many machine learning methods are trying to make auxiliary prediction. However, these methods attach more attention to the image data of suspected tumor, and focus on improving the accuracy of image classification, but ignore the significance of patient-level contextual information for disease diagnosis in actual clinical diagnosis. To make more use of patient information and improve the accuracy of diagnosis, we propose a new melanoma classification model based on EffNet and Resnet. Our model not only uses images within the same patient but also consider patient-level contextual information for better cancer prediction. The experimental results demonstrated that the proposed model achieved 0.981 ACC. Furthermore, we note that the overall ROC value of the model is 0.976 which is better than the previous state-of-the-art approaches.

eess.IV↗
arXiv2021

Ferromagnetic half-metallicity in YBaCo2O6 and spin-states driven metal-insulator transition

Cobaltates have rich spin-states and diverse properties. Using spin-state pictures and firstprinciples calculations, here we study the electronic structure and magnetism of the mixed-valent double perovskite YBaCo2O6. We find that YBaCo2O6 is in the formal intermediate-spin (IS) Co3+/low-spin (LS) Co4+ ground state. The hopping of eg electron from IS-Co3+ to LS-Co4+ via double exchange gives rise to a ferromagnetic half-metallicity, which well accounts for the recent experiments. The reduction of both magnetization and Curie temperature by oxygen vacancies is discussed, aided with Monte Carlo simulations. We also explore several other possible spin-states and their interesting electronic/magnetic properties. Moreover, we predict that a volume expansion more than 3% would tune YBaCo2O6 into the high-spin (HS) Co3+/LS Co4+ ferromagnetic state and simultaneously drive a metal-insulator transition. Therefore, spin-states are a useful parameter for tuning the material properties of cobaltates.

cond-mat.mtrl-sci↗
arXiv2021

Rotation Based Slice Error Correction Protocol for Continuous-variable Quantum Key Distribution and its Implementation with Polar Codes

Reconciliation is an essential procedure for continuous-variable quantum key distribution (CV-QKD). As the most commonly used reconciliation protocol in short-distance CV-QKD, the slice error correction (SEC) allows a system to distill more than 1 bit from each pulse. However, its quantization efficiency is greatly affected by the noisy channel with a low signal-to-noise ratio (SNR), which usually limits the secure distance to about 30 km. In this paper, an improved SEC protocol, named Rotation-based SEC (RSEC), is proposed through performing a random orthogonal rotation on the raw data before quantization, and deducing a new estimator for quantized sequences. Moreover, the RSEC protocol is implemented with polar codes. Experimental results show that the proposed protocol can reach up to a quantization efficiency of about 99\%, and maintains at around 96\% even at the relatively low SNRs $(0.5,1)$, which theoretically extends the secure distance to about 45 km. When implemented with the polar codes with block length of 16 Mb, the RSEC can achieve a reconciliation efficiency of above 95\%, which outperforms all previous SEC schemes. In terms of finite-size effects, we achieve a secret key rate of $7.83\times10^{-3}$ bits/pulse at a distance of 33.93 km (the corresponding SNR value is 1). These results indicate that the proposed protocol significantly improves the performance of SEC and is a competitive reconciliation scheme for the CV-QKD system.

quant-ph↗
arXiv2019

Novel Reconciliation Protocol Based on Spinal Code for Continuous-variable Quantum Key Distribution

Reconciliation is a crucial procedure in post-processing of continuous variable quantum key distribution (CV-QKD) system, which is used to make two distant legitimate parties share identical corrected keys. The adaptive reconciliation is necessary and important for practical systems to cope with the variable channel. Many researchers adopt the punctured LDPC codes to implement adaptive reconciliation. In this paper, a novel rateless reconciliation protocol based on spinal code is proposed, which can achieve a high-efficiency and adaptive reconciliation in a larger range of SNRs. Due to the short codes length and simple tructure, our protocol is easy to implement without the complex codes designs of fixed rate codes, e.g., LDPC codes. Meanwhile, the structure of our protocol is highly parallel, which is suitable for hardware implementation, thus it also has the potential of high-speed hardware implementation. Besides, the security of proposed protocol is proved in theory. Experiment results show that the reconciliation efficiency maintains around 95% for ranging SNRs in a larger range (0,0.5), even exceeds 96.5% at extremely low SNR (<= 0.03) by using this novel scheme. The proposed protocol makes the long-distance CV-QKD systems much easier and stable to perform a high-performance and adaptive reconciliation.

quant-ph↗
arXiv2019

Contrasting magnetism in isovalent layered LaSr3NiRuO4H4 and LaSrNiRuO4 due to distinct spin-orbital states

The recently synthesized first 4d transition-metal oxide-hydride LaSr3NiRuO4H4 with the unusual high H:O ratio surprisingly displays no magnetic order down to 1.8 K. This is in sharp contrast to the similar unusual low-valent Ni+ - Ru2+ layered oxide LaSrNiRuO4 which has a rather high ferromagnetic (FM) ordering Curie temperature TC about 250 K. In this work, using density functional calculations with aid of crystal field level diagrams and superexchange pictures, we find that the contrasting magnetism is due to the distinct spin-orbital states of the Ru2+ ions (in addition to the common Ni+ S=1/2 state but with a different orbital state): the Ru2+ S=0 state in LaSr3NiRuO4H4, but the Ru2+ S=1 state in LaSrNiRuO4. The Ru2+ S=0 state has the (xy)2(xz; yz)4 occupation due to the RuH4O2 octahedral coordination, and then the nonmagnetic Ru2+ ions dilute the S=1/2 Ni+ sublattice which consequently has a very weak antiferromagnetic (AF) superexchange and thus accounts for no presence of magnetic order down to 1.8 K in LaSr3NiRuO4H4. In strong contrast, the Ru2+ S=1 state in LaSrNiRuO4 has the (3z2-r2)2(xz; yz)3(xy)1 occupation due to the planar square RuO4 coordination, and then the multi-orbital FM superexchange between the S=1/2 Ni+ and S=1 Ru2+ ions gives rise to the high TC in LaSrNiRuO4. This work highlights the importance of spin-orbital states in determining the distinct magnetism.

cond-mat.str-el↗