arXiv subjects

Xu Yang

Publications and source records attributed to Xu Yang.

At least 181 records · Page 10Linked to original sources

arXiv2025

Accelerated Kaczmarz methods via randomized sketch techniques for solving consistent linear systems

Motivated by the randomized sketch to solve a variety of problems in scientific computation, we improve both the maximal weighted residual Kaczmarz method and the randomized block average Kaczmarz method using two new randomized sketch techniques. Besides, convergence analyses of the proposed methods are provided. Furthermore, we establish an upper bound for the discrepancy between the numerical solutions obtained via the proposed methods and those derived from the original approaches. Numerical experiments demonstrate that the new methods perform better than the existing ones in terms of the running time with the same accuracy.

math.NA↗

arXiv2025

MSLoRA: Multi-Scale Low-Rank Adaptation via Attention Reweighting

We introduce MSLoRA, a backbone-agnostic, parameter-efficient adapter that reweights feature responses rather than re-tuning the underlying backbone. Existing low-rank adaptation methods are mostly confined to vision transformers (ViTs) and struggle to generalize across architectures. MSLoRA unifies adaptation for both convolutional neural networks (CNNs) and ViTs by combining a low-rank linear projection with a multi-scale nonlinear transformation that jointly modulates spatial and channel attention. The two components are fused through pointwise multiplication and a residual connection, yielding a lightweight module that shifts feature attention while keeping pretrained weights frozen. Extensive experiments demonstrate that MSLoRA consistently improves transfer performance on classification, detection, and segmentation tasks with roughly less than 5\% of backbone parameters. The design further enables stable optimization, fast convergence, and strong cross-architecture generalization. By reweighting rather than re-tuning, MSLoRA provides a simple and universal approach for efficient adaptation of frozen vision backbones.

cs.CV↗

arXiv2025

Efficient and Effective In-context Demonstration Selection with Coreset

In-context learning (ICL) has emerged as a powerful paradigm for Large Visual Language Models (LVLMs), enabling them to leverage a few examples directly from input contexts. However, the effectiveness of this approach is heavily reliant on the selection of demonstrations, a process that is NP-hard. Traditional strategies, including random, similarity-based sampling and infoscore-based sampling, often lead to inefficiencies or suboptimal performance, struggling to balance both efficiency and effectiveness in demonstration selection. In this paper, we propose a novel demonstration selection framework named Coreset-based Dual Retrieval (CoDR). We show that samples within a diverse subset achieve a higher expected mutual information. To implement this, we introduce a cluster-pruning method to construct a diverse coreset that aligns more effectively with the query while maintaining diversity. Additionally, we develop a dual retrieval mechanism that enhances the selection process by achieving global demonstration selection while preserving efficiency. Experimental results demonstrate that our method significantly improves the ICL performance compared to the existing strategies, providing a robust solution for effective and efficient demonstration selection.

cs.CV↗

arXiv2025

On the Color Discrepancy of Spanning Trees in Random and Randomly Perturbed Graphs

In this work, we study the color discrepancy of spanning trees in random graphs. We show that for the Erdős-Rényi random graph $G(n,p)$ with $p$ above the connectivity threshold, the following holds with high probability: in every 2-edge-coloring of the graph, there exists a spanning tree with a linear number of leaves such that one color class contains more than $\frac{1 + \varepsilon}{2}n $ of the tree's edges. Here, $\varepsilon>0$ is a small absolute constant independent of $p$. We also extend this line of research to randomly perturbed dense graphs, showing that adding a few random edges to a dense graph typically creates a spanning tree with a large color discrepancy under any 2-edge-coloring.

math.CO↗

arXiv2025

Adversarial Graph Fusion for Incomplete Multi-view Semi-supervised Learning with Tensorial Imputation

View missing remains a significant challenge in graph-based multi-view semi-supervised learning, hindering their real-world applications. To address this issue, traditional methods introduce a missing indicator matrix and focus on mining partial structure among existing samples in each view for label propagation (LP). However, we argue that these disregarded missing samples sometimes induce discontinuous local structures, i.e., sub-clusters, breaking the fundamental smoothness assumption in LP. Consequently, such a Sub-Cluster Problem (SCP) would distort graph fusion and degrade classification performance. To alleviate SCP, we propose a novel incomplete multi-view semi-supervised learning method, termed AGF-TI. Firstly, we design an adversarial graph fusion scheme to learn a robust consensus graph against the distorted local structure through a min-max framework. By stacking all similarity matrices into a tensor, we further recover the incomplete structure from the high-order consistency information based on the low-rank tensor learning. Additionally, the anchor-based strategy is incorporated to reduce the computational complexity. An efficient alternative optimization algorithm combining a reduced gradient descent method is developed to solve the formulated objective, with theoretical convergence. Extensive experimental results on various datasets validate the superiority of our proposed AGF-TI as compared to state-of-the-art methods. Code is available at https://github.com/ZhangqiJiang07/AGF_TI.

cs.LG↗

arXiv2025

Frozen Gaussian approximation for the fractional Schrödinger equation

We develop a refined Frozen Gaussian approximation (FGA) for the fractional Schrödinger equation in the semi-classical regime, where the solution exhibits rapid oscillations as the scaled Planck constant $\varepsilon$ becomes small. Our approach utilizes an integral representation based on asymptotic analysis, offering a highly efficient computational framework for high-frequency wave function evolution. Crucially, we introduce the momentum space representation of the FGA and a regularization parameter $δ$ to address singularities in the higher-order derivatives of the Hamiltonian flow coefficients, which are typically assumed to be second-order differentiable or smooth in conventional analysis. We rigorously prove convergence of the method to the true solution and provide numerical experiments that demonstrate its precision and robust convergence behavior.

math.NA↗

arXiv2025

Antarctic Infrared Binocular Telescope. I. System Overview, Laboratory Testing, and On-Sky Performance Evaluation

Infrared time-domain surveys remain significantly underdeveloped compared with their optical counterparts. We have developed the Antarctic Infrared Binocular Telescope (AIRBT) to study the dynamic infrared sky at Dome A, Antarctica, taking advantage of the superb infrared observational conditions at this site. AIRBT consists of two identical 15 cm f/3 optical tube assemblies and two cost-effective indium gallium arsenide (InGaAs) cameras equipped with J and H filters, respectively. The cameras have 640 x 512 pixels with a size of 15 micrometers, providing a scale of 6.9 arcseconds per pixel and a field of view of 1.22 x 0.97 square degrees. We characterize the performance of the InGaAs cameras, including bias, readout noise, dark current, nonlinearity, and photon transfer curve. Our analysis highlights the distinct behaviors of InGaAs cameras compared with charge-coupled devices (CCDs). The bias and readout noise show temperature dependence, and the noise measured from the photon transfer curves has additional components that increase with exposure time. On-sky tests were conducted in October 2022 including system calibration, limiting depth, and photometric precision. For a single 3-second exposure, we achieved 5-sigma limiting magnitudes of 11.2 mag (Vega system) in J band and 9.7 mag in H band. The best photometric precision reached 20 millimagnitudes at the bright end, which could be further improved to sub-percent levels through image stacking. AIRBT was installed at Dome A in January 2023, and scientific observations began as soon as darkness set in.

astro-ph.IM↗

arXiv2025

Lingxi: Repository-Level Issue Resolution Framework Enhanced by Procedural Knowledge Guided Scaling

Driven by the advancements of Large Language Models (LLMs), LLM-powered agents are making significant improvements in software engineering tasks, yet struggle with complex, repository-level issue resolution. Existing agent-based methods have two key limitations. First, they lack of procedural knowledge (i.e., how an issue is fixed step-by-step and rationales behind it) to learn and leverage for issue resolution. Second, they rely on massive computational power to blindly explore the solution space. % To address those limitations, we propose Lingxi, an issue resolution framework that leverages procedural knowledge extracted from historical issue-fixing data to guide agents in solving repository-level issues. \ourTool first constructs this knowledge offline through a hierarchical abstraction mechanism, enabling agents to learn the how and why behind a fix, not just the final solution. During online application, it employs a knowledge-driven scaling method that leverages the procedural knowledge of similar issues to intelligently analyze the target issue from multiple perspectives, in sharp contrast to undirected, brute-force exploration. % Lingxi successfully resolves 74.6\% of bugs on the SWE-bench Verified benchmark in Past@1 setting, outperforming five state-of-the-art techniques by a significant margin (5.4\% to 14.9\%). Our comprehensive ablation study confirmed that the success of Lingxi comes directly from its use of procedural knowledge. Without it, the performance gains from scaling alone is negligible. Our qualitative study further shows that the ``design patterns $\&$ coding practices'' is the most critical knowledge aspect, and that the roles of different knowledge aspects switch across different stages (i.e., analysis, planning, and fixing).

cs.SE↗

arXiv2025

LayerCake: Token-Aware Contrastive Decoding within Large Language Model Layers

Large language models (LLMs) excel at natural language understanding and generation but remain vulnerable to factual errors, limiting their reliability in knowledge-intensive tasks. While decoding-time strategies provide a promising efficient solution without training, existing methods typically treat token-level and layer-level signals in isolation, overlooking the joint dynamics between them. In this work, we introduce a token-aware, layer-localized contrastive decoding method that aligns specific token types with their most influential transformer layers to improve factual generation. Through empirical attention analysis, we identify two key patterns: punctuation tokens receive dominant attention in early layers, while conceptual tokens govern semantic reasoning in intermediate layers. By selectively suppressing attention to these token types at their respective depths, we achieve the induction of controlled factual degradation and derive contrastive signals to guide the final factual decoding. Our method requires no additional training or model modification, and experiments demonstrate that our method consistently improves factuality across multiple LLMs and various benchmarks.

cs.AI↗

arXiv2025

Integrating Sequential Hypothesis Testing into Adversarial Games: A Sun Zi-Inspired Framework

This paper investigates the interplay between sequential hypothesis testing (SHT) and adversarial decision-making in partially observable games, focusing on the deceptive strategies of red and blue teams. Inspired by Sun Zi's The Art of War and its emphasis on deception, we develop a novel framework to both deceive adversaries and counter their deceptive tactics. We model this interaction as a Stackelberg game where the blue team, as the follower, optimizes its controls to achieve its goals while misleading the red team into forming incorrect beliefs on its intentions. The red team, as the leader, strategically constructs and instills false beliefs through the blue team's envisioned SHT to manipulate the blue team's behavior and reveal its true objectives. The blue team's optimization problem balances the fulfillment of its primary objectives and the level of misdirection, while the red team coaxes the blue team into behaving consistently with its actual intentions. We derive a semi-explicit solution for the blue team's control problem within a linear-quadratic framework, and illustrate how the red team leverages leaked information from the blue team to counteract deception. Numerical experiments validate the model, showcasing the effectiveness of deception-driven strategies in adversarial systems. These findings integrate ancient strategic insights with modern control and game theory, providing a foundation for further exploration in adversarial decision-making, such as cybersecurity, autonomous systems, and financial markets.

math.OC↗

arXiv2025

R&D-Agent: An LLM-Agent Framework Towards Autonomous Data Science

Recent advances in AI and ML have transformed data science, yet increasing complexity and expertise requirements continue to hinder progress. Although crowd-sourcing platforms alleviate some challenges, high-level machine learning engineering (MLE) tasks remain labor-intensive and iterative. We introduce R&D-Agent, a comprehensive, decoupled, and extensible framework that formalizes the MLE process. R&D-Agent defines the MLE workflow into two phases and six components, turning agent design for MLE from ad-hoc craftsmanship into a principled, testable process. Although several existing agents report promising gains on their chosen components, they can mostly be summarized as a partial optimization from our framework's simple baseline. Inspired by human experts, we designed efficient and effective agents within this framework that achieve state-of-the-art performance. Evaluated on MLE-Bench, the agent built on R&D-Agent ranks as the top-performing machine learning engineering agent, achieving 35.1% any medal rate, demonstrating the ability of the framework to speed up innovation and improve accuracy across a wide range of data science applications. We have open-sourced R&D-Agent on GitHub: https://github.com/microsoft/RD-Agent.

cs.AI↗

arXiv2025

CoT Vectors: Transferring and Probing the Reasoning Mechanisms of LLMs

Chain-of-Thought (CoT) prompting has emerged as a powerful approach to enhancing the reasoning capabilities of Large Language Models (LLMs). However, existing implementations, such as in-context learning and fine-tuning, remain costly and inefficient. To improve CoT reasoning at a lower cost, and inspired by the task vector paradigm, we introduce CoT Vectors, compact representations that encode task-general, multi-step reasoning knowledge. Through experiments with Extracted CoT Vectors, we observe pronounced layer-wise instability, manifesting as a U-shaped performance curve that reflects a systematic three-stage reasoning process in LLMs. To address this limitation, we propose Learnable CoT Vectors, optimized under a teacher-student framework to provide more stable and robust guidance. Extensive evaluations across diverse benchmarks and models demonstrate that CoT Vectors not only outperform existing baselines but also achieve performance comparable to parameter-efficient fine-tuning methods, while requiring fewer trainable parameters. Moreover, by treating CoT Vectors as a probe, we uncover how their effectiveness varies due to latent space structure, information density, acquisition mechanisms, and pre-training differences, offering new insights into the functional organization of multi-step reasoning in LLMs. The source code will be released.

cs.CL↗

arXiv2025

CATMark: A Context-Aware Thresholding Framework for Robust Cross-Task Watermarking in Large Language Models

Watermarking algorithms for Large Language Models (LLMs) effectively identify machine-generated content by embedding and detecting hidden statistical features in text. However, such embedding leads to a decline in text quality, especially in low-entropy scenarios where performance needs improvement. Existing methods that rely on entropy thresholds often require significant computational resources for tuning and demonstrate poor adaptability to unknown or cross-task generation scenarios. We propose \textbf{C}ontext-\textbf{A}ware \textbf{T}hreshold watermarking ($\myalgo$), a novel framework that dynamically adjusts watermarking intensity based on real-time semantic context. $\myalgo$ partitions text generation into semantic states using logits clustering, establishing context-aware entropy thresholds that preserve fidelity in structured content while embedding robust watermarks. Crucially, it requires no pre-defined thresholds or task-specific tuning. Experiments show $\myalgo$ improves text quality in cross-tasks without sacrificing detection accuracy.

cs.CR↗

arXiv2025

R&D-Agent-Quant: A Multi-Agent Framework for Data-Centric Factors and Model Joint Optimization

Financial markets pose fundamental challenges for asset return prediction due to their high dimensionality, non-stationarity, and persistent volatility. Despite advances in large language models and multi-agent systems, current quantitative research pipelines suffer from limited automation, weak interpretability, and fragmented coordination across key components such as factor mining and model innovation. In this paper, we propose R&D-Agent for Quantitative Finance, in short RD-Agent(Q), the first data-centric multi-agent framework designed to automate the full-stack research and development of quantitative strategies via coordinated factor-model co-optimization. RD-Agent(Q) decomposes the quant process into two iterative stages: a Research stage that dynamically sets goal-aligned prompts, formulates hypotheses based on domain priors, and maps them to concrete tasks, and a Development stage that employs a code-generation agent, Co-STEER, to implement task-specific code, which is then executed in real-market backtests. The two stages are connected through a feedback stage that thoroughly evaluates experimental outcomes and informs subsequent iterations, with a multi-armed bandit scheduler for adaptive direction selection. Empirically, RD-Agent(Q) achieves up to 2X higher annualized returns than classical factor libraries using 70% fewer factors, and outperforms state-of-the-art deep time-series models on real markets. Its joint factor-model optimization delivers a strong balance between predictive accuracy and strategy robustness. Our code is available at: https://github.com/microsoft/RD-Agent.

q-fin.CP↗

arXiv2025

Large Language Model Assisted Automated Algorithm Generation and Evolution via Meta-black-box optimization

Meta-black-box optimization has been significantly advanced through the use of large language models (LLMs), yet in fancy on constrained evolutionary optimization. In this work, AwesomeDE is proposed that leverages LLMs as the strategy of meta-optimizer to generate update rules for constrained evolutionary algorithm without human intervention. On the meanwhile, $RTO^2H$ framework is introduced for standardize prompt design of LLMs. The meta-optimizer is trained on a diverse set of constrained optimization problems. Key components, including prompt design and iterative refinement, are systematically analyzed to determine their impact on design quality. Experimental results demonstrate that the proposed approach outperforms existing methods in terms of computational efficiency and solution accuracy. Furthermore, AwesomeDE is shown to generalize well across distinct problem domains, suggesting its potential for broad applicability. This research contributes to the field by providing a scalable and data-driven methodology for automated constrained algorithm design, while also highlighting limitations and directions for future work.

cs.NE↗

arXiv2025

Deep Reinforcement Learning-Assisted Component Auto-Configuration of Differential Evolution Algorithm for Constrained Optimization: A Foundation Model

Despite significant efforts to manually design high-performance evolutionary algorithms, their adaptability remains limited due to the dynamic and ever-evolving nature of real-world problems. The "no free lunch" theorem highlights that no single algorithm performs optimally across all problems. While online adaptation methods have been proposed, they often suffer from inefficiency, weak convergence, and limited generalization on constrained optimization problems (COPs). To address these challenges, we introduce a novel framework for automated component configuration in Differential Evolution (DE) algorithm to address COPs, powered by Deep Reinforcement Learning (DRL). Specifically, we propose SuperDE, a foundation model that dynamically configures DE's evolutionary components based on real-time evolution. Trained offline through meta-learning across a wide variety of COPs, SuperDE is capable of recommending optimal per-generation configurations for unseen problems in a zero-shot manner. Utilizing a Double Deep Q-Network (DDQN), SuperDE adapts its configuration strategies in response to the evolving population states during optimization. Experimental results demonstrate that SuperDE significantly outperforms existing state-of-the-art algorithms on benchmark test suites, achieving superior generalization and optimization performance.

cs.NE↗

arXiv2025

Modeling and Control of AWOISV: A Filtered Tube-Based MPC Approach for Simultaneous Tracking of Lateral Position and Heading Angle

An all-wheel omni-directional independent steering vehicle (AWOISV) is a specialized all-wheel independent steering vehicle with each wheel capable of steering up to 90°, enabling unique maneuvers like yaw and diagonal movement. This paper introduces a theoretical steering radius angle and sideslip angle ($ θ_R $-$β_R $) representation, based on the position of the instantaneous center of rotation relative to the wheel rotation center, defining the motion modes and switching criteria for AWOISVs. A generalized $ v$-$β$-$r $ dynamic model is developed with forward velocity $v$, sideslip angle $β$, and yaw rate $r$ as states, and $θ_R$ and $β_R$ as control inputs. This model decouples longitudinal and lateral motions into forward and rotational motions, allowing seamless transitions across all motion modes under specific conditions. A filtered tube-based linear time-varying MPC (FT-LTVMPC) strategy is proposed, achieving simultaneous tracking of lateral position and arbitrary heading angles, with robustness to model inaccuracies and parameter uncertainties. Co-simulation and hardware-in-loop (HIL) experiments confirm that FT-LTVMPC enables high-precision control of both position and heading while ensuring excellent real-time performance.

cs.RO↗

arXiv2025

Adaptively Clustering Neighbor Elements for Image-Text Generation

We propose a novel Transformer-based image-to-text generation model termed as \textbf{ACF} that adaptively clusters vision patches into object regions and language words into phrases to implicitly learn object-phrase alignments for better visual-text coherence. To achieve this, we design a novel self-attention layer that applies self-attention over the elements in a local cluster window instead of the whole sequence. The window size is softly decided by a clustering matrix that is calculated by the current input data and thus this process is adaptive. By stacking these revised self-attention layers to construct ACF, the small clusters in the lower layers can be grouped into a bigger cluster, \eg vision/language. ACF clusters small objects/phrases into bigger ones. In this gradual clustering process, a parsing tree is generated which embeds the hierarchical knowledge of the input sequence. As a result, by using ACF to build the vision encoder and language decoder, the hierarchical object-phrase alignments are embedded and then transferred from vision to language domains in two popular image-to-text tasks: Image captioning and Visual Question Answering. The experiment results demonstrate the effectiveness of ACF, which outperforms most SOTA captioning and VQA models and achieves comparable scores compared with some large-scale pre-trained models. Our code is available \href{https://github.com/ZihuaEvan/ACFModel/}{[here]}.

cs.CV↗