Search arXiv⌕ Search

arXiv subjects

Xu Yang

Publications and source records attributed to Xu Yang.

At least 109 records · Page 6Linked to original sources
arXiv2026

Temporal Evidence Routing with Structured Visual Evidence for TimeLogicQA

TimeLogicQA evaluates whether video question answering systems can reason over temporal relations such as event existence, ordering, persistence, boundary conditions, and overlap. We address this task with a visual evidence routing pipeline that separates perception from symbolic temporal reasoning. The system first parses each question into event targets, answer mode, candidate options, and temporal operators. It then routes videos according to duration and operator difficulty, using ordered full-frame evidence for short clips and event-focused candidate windows for long videos. A multimodal large language model produces structured visual evidence for the relevant events, while programmatic verifiers recover dense action intervals and a deterministic reducer applies operator-specific temporal rules to produce the final answer. Conservative fusion accepts an answer only when the visual evidence, temporal program, and confidence checks agree, reducing noisy answer flips. On the official test evaluation, our final system achieves an AvgAcc of 81.8.

cs.CV↗
arXiv2026

Benchmarking Inductive Biases for Multivariate Time-Series Anomaly Detection with a Robust Multi-View Channel-Graph Detector

We present a unified experiment, analysis, and benchmark study of multivariate time-series (MTS) anomaly detection. Ten family-representative detectors -- spanning statistical, reconstruction, association, frequency, and generic-transformer families -- are evaluated on five datasets (SMD, MSL, SMAP, PSM, and MSDS) under effectiveness, efficiency, robustness, and cross-dataset generalisation. All methods share the same windowing, scoring, hardware, and metric protocols. Effectiveness, ablation, and robustness use three random seeds; cross-dataset transfer uses seed~0 because each extra seed requires $250$ source-target evaluations. The benchmark yields three method-independent findings: no single-bias baseline dominates; absolute perturbation VUS-ROC is more informative than retention ratios; and MSDS behaves as an event-dense deployment workload rather than a sparse point-anomaly benchmark. Under this protocol we also introduce \ours{}, an adaptive detector family combining a NOTEARS-constrained directed channel-graph view with optional patch-attention and temporal-association views. \ours{} achieves the best macro-average VUS-ROC ($0.675$, $+5.1$~pt over the second-best LSTM-AE), ranks first overall, and reaches the top-3 on all five datasets. Its wins on MSL and MSDS are narrow, while its average and robustness gains are larger: under the same three-seed robustness protocol for every method, it obtains the strongest absolute VUS-ROC across noise, channel dropout, and time-shift perturbations. We release the MSDS preprocessing protocol, configurations, scripts, and seed-level metric dumps.

cs.LG↗
arXiv2026

AeroTSBoost: Temporal-Statistical Boosting for Real-World UAV Telemetry Anomaly Mining

Mining anomalies from unmanned aerial vehicle (UAV) state-estimation logs is challenging because failures are sparse, temporally structured, and distributed across heterogeneous PX4 telemetry streams with variable sensor availability and missing values. We present AeroTSBoost, a temporal-statistical boosting framework for real-world UAV telemetry anomaly mining. AeroTSBoost aligns multivariate flight logs, converts each window into deterministic descriptors that capture distributional shifts, quantile structure, endpoint drift, local dynamics, and lag correlation, and trains a class-balanced LightGBM detector. On UAV-SEAD, AeroTSBoost achieves the strongest AUPRC among evaluated classical, supervised tabular, neural reconstruction, recurrent, Granger-causality-based, and frequency-domain baselines. Across five seeds, it reaches $0.7516\pm0.0043$ AUPRC and $0.5342\pm0.0108$ threshold-swept event F1, improving AUPRC by 5.79 absolute points over the strongest non-AeroTSBoost baseline. Under purged chronological and leave-log-out protocols, it remains the best AUPRC method, reaching $0.6066\pm0.0193$ and $0.6388\pm0.0315$, respectively. On related ALFA fixed-wing UAV fault logs, AeroTSBoost reaches $0.9259\pm0.0076$ leave-sequence-out AUPRC, ahead of RandomForest ($0.8835\pm0.0797$) and moments-only ($0.8700\pm0.0481$). These results show that deterministic temporal-statistical representations remain highly competitive for sparse anomaly mining in operational cyber-physical telemetry.

cs.CE↗
arXiv2026

Rethinking LLM Ensembling from the Perspective of Mixture Models

Model ensembling is a well-established technique for improving the performance of machine learning models. Conventionally, this involves averaging the output distributions of multiple models and selecting the most probable label. This idea has been naturally extended to large language models (LLMs), yielding improved performance but incurring substantial computational cost. This inefficiency stems from directly applying conventional ensemble implementation to LLMs, which require a separate forward pass for each model to explicitly compute the ensemble distribution. In this paper, we propose the Mixture-model-like Ensemble (ME). By reinterpreting the ensemble as a mixture model, ME stochastically selects a single model at each step to generate the next token, thereby avoiding the need to explicitly compute the full ensemble distribution. ME is mathematically equivalent to sampling from the ensemble distribution, but requires invoking only one model, making it 1.78x-2.68x faster than conventional ensembling. Furthermore, this perspective connects LLM ensembling and token-level routing methods, suggesting that LLM ensembling is a special case of routing methods. Our findings open new avenues for efficient LLM ensembling and motivate further exploration of token-level routing strategies for LLMs. Our code is available at https://github.com/Kamichanw/Mixture-model-like-Ensemble.

cs.LG↗
arXiv2026

Adversarial Orthogonal Disentanglement for LVLM Hallucination Mitigation

Large Vision-Language Models (LVLMs) have advanced multimodal understanding, yet their reliability is limited by hallucination, where generated content conflicts with visual facts. Existing mitigation methods either rely on costly external interventions, such as instruction tuning and retrieval, or use internal mechanisms that remain limited by flawed attention weights and entangled hidden representations. We propose Adversarial Orthogonal Disentanglement (AOD), a latent geometric framework for mitigating LVLM hallucinations. AOD learns a hallucination-related direction through a minimax objective: a classifier concentrates hallucination signals into the projected component, while an adversary removes them from the orthogonal residual space via a Gradient Reversal Layer. The learned direction enables a training-free dual-forward-pass contrastive decoding strategy that suppresses hallucinations while preserving general capabilities. Experiments on three LVLMs across four hallucination and four utility benchmarks show that AOD consistently outperforms strong baselines. It improves POPE accuracy by over 6\% on average, boosts AMBER by 6\%, and maintains strong performance on utility tasks such as MMMU. Further analysis shows robust transfer across datasets, suggesting that AOD captures general hallucination-related biases rather than dataset-specific artifacts. Our source code and datasets are available at https://github.com/Hunter-Wrynn/AOD.

cs.CV↗
arXiv2026

Ultrafast Magneto-optical Fingerprints of Altermagnetism in MnTe

Recently identified altermagnets exhibit a distinctive dual-space nature: they possess spin-split electronic bands akin to ferromagnets in momentum space while maintaining the fully compensated magnetization of antiferromagnets in real space. This inherent duality, originating from the same crystal symmetry, gives rise to various intriguing physical phenomena unique to altermagnets. Consequently, a robust and efficient experimental signature capable of revealing this dual character is critically needed. The magneto-optical Kerr and Voigt effects, given their high sensitivity to ferromagnetism and antiferromagnetism, respectively, are ideally suited to probe this duality. Here, using time-resolved pump-probe magneto-optical measurements, we report the coexistence of pronounced Kerr and Voigt effects in the altermagnet MnTe. Combining the magnetization measurement and first-principles calculations, we demonstrate that the Kerr effect originates from the intrinsic Berry curvature of altermagnetism distribution in momentum space, while the Voigt effect arises from an anisotropic permittivity induced by the in-plane Néel order in real space, directly revealing the dual-space nature of altermagnets. Furthermore, the transient Kerr signal exhibits faster relaxation dynamics than the transient Voigt signal, underscoring their distinct origins in Berry curvature and Néel order, respectively. These findings establish transient magneto-optical responses as distinctive fingerprints of altermagnetism and position altermagnets as promising platforms for manipulating magneto-optical phenomena in ultrafast spin optoelectronics.

cond-mat.mtrl-sci↗
arXiv2026

$π$MPC: A Parallel-in-horizon and Construction-free NMPC Solver

The alternating direction method of multipliers (ADMM) has gained increasing popularity in embedded model predictive control (MPC) due to its code simplicity and pain-free parameter selection. However, existing ADMM solvers either target general quadratic programming (QP) problems or exploit sparse MPC formulations via Riccati recursions, which are inherently sequential and therefore difficult to parallelize for long prediction horizons. This technical note proposes a novel \textit{parallel-in-horizon} and \textit{construction-free} nonlinear MPC algorithm, termed $π$MPC, which combines a new variable-splitting scheme with a velocity-based system representation in the ADMM framework, enabling horizon-wise parallel execution while operating directly on system matrices without explicit MPC-to-QP construction. Numerical experiments and accompanying code are provided to validate the effectiveness of the proposed method.

math.OC↗
arXiv2026

FT-Dojo: Towards Autonomous LLM Fine-Tuning with Language Agents

Fine-tuning large language models for vertical domains remains labor-intensive, requiring practitioners to curate data, configure training, and iteratively diagnose model behavior. Despite growing interest in autonomous machine learning and language agents, end-to-end LLM fine-tuning has not been systematically studied as an interactive agent task. We introduce FT-Dojo, an interactive benchmark environment for autonomous LLM fine-tuning, comprising 13 tasks across 5 domains. Rather than a new collection of static datasets, FT-Dojo standardizes a task interface, shared raw-data repository, sandboxed execution environment, structured feedback protocol, and held-out evaluation procedure. We further develop FT-Agent, a fine-tuning-oriented autonomous framework that uses structured iteration planning, fail-fast validation, and multi-level feedback analysis to refine data and training strategies. Experiments show that FT-Agent provides a strong initial baseline, achieving the best performance on 10 out of 13 tasks, with additional controlled comparisons against frontier agents, open-source planning backbones, and multi-run statistics supporting the main findings. Case studies show that agents can recover from failures through cumulative learning, while still exposing limitations in causal diagnosis and long-horizon planning. The implementation is available at https://github.com/microsoft/rd-agent.

cs.AI↗
arXiv2026

KIO-planner: Attention-Guided Single-Stage Motion Planning with Dual Mapping for UAV Navigation

Autonomous UAV flight in confined, wall-dense environments requires low-latency and reliable motion planning under strict safety constraints. Traditional optimization-based planners suffer from mapping latency and easily fall into local minima when navigating through dense structural obstacles. Meanwhile, existing end-to-end learning methods struggle to extract fine-grained geometric features from raw depth images and lack hard kinodynamic constraints, leading to unpredictable collisions near walls. To address these issues, we propose KIO-planner, an attention-guided single-stage trajectory planning framework. First, we integrate a Convolutional Block Attention Module (CBAM) into the perception backbone to adaptively focus on critical structural edges and traversable space. Second, we introduce a novel Dual Mapping mechanism--comprising physical bounds activation and a deterministic Geometric Safety Shield in the depth-pixel space--to enforce kinodynamic feasibility and collision-free flight without global map fusion. Extensive high-fidelity simulated experiments demonstrate that KIO-planner enables highly agile navigation at speeds up to 3.0 m/s. Compared to the state-of-the-art baseline, KIO-planner achieves lower inference latency (approximately 24 ms) and generates significantly smoother trajectories, reducing control cost by 28.4%. Most notably, our Dual Mapping substantially increases the worst-case safety margin, measured by minimum distance to obstacles, from 0.48 m to 0.76 m, ensuring fast, smooth, and safer navigation in highly constrained environments.

cs.RO↗
arXiv2026

Bidirectional Plasma Jets Driven by Magnetic Reconnection: Observations by GST and SDO

Using high-resolution photospheric and chromospheric observations taken by the Goode Solar Telescope (GST), we studied two groups of bidirectional plasma jets occurring in active region NOAA 13110. Supplementary observations are also provided by Solar Dynamics Observatory's (SDO) Atmospheric Imaging Assembly (AIA) and Helioseismic and Magnetic Imager (HMI). From the photospheric observations and magnetograms, the two successive bidirectional plasma jets were initially located in the vicinity of the polarity inversion lines and at the intersection of the umbra and penumbra of the sunspot, followed by magnetic flux emergence and cancellation. As the cool filamentary threads are continuously emerging from the lower chromosphere and interact with overlying horizontal magnetic loops, it leads to the bidirectional jets, erupting outflow plasmoids, and heating coronal magnetic loops. We find that the bidirectional jets extended from the central excitation location in opposite directions, at the speed of about dozens of km s$^{-1}$. For the first jet, the initial brightening first appears in 304 angstroms, about 30 s earlier than the H$α$ observations, indicating that magnetic reconnection takes place in the transition region. While the initial reconnection for the second jet occurs simultaneously in H$α$ and 304 angstroms, showing the recurrent eruptions. These observations confirm that the bidirectional plasma jets can be generated by magnetic reconnection between the rising filamentary threads or material and the overlying horizontal magnetic loops. Our results provide new insights into the generation of the bidirectional plasma jets and reconnection-based coronal heating.

astro-ph.SR↗
arXiv2026

Depletion-mode N-polar AlN-based high electron mobility transistors with improved on/off ratios

We report N-polar AlN-based high-electron mobility transistors (HEMTs) with a GaN channel thickness of 5.2 nm on N-polar AlN on sapphire. The threshold voltage is around -2.4 to -3.0 V with saturation currents over 240 mA/mm and on/off ratios as high as 10,000, much higher than previously reported N-polar AlN-based HEMTs. The high on/off ratio is attributed to the use of an abrupt AlN/GaN heterostructure with a dedicated AlN transition layer, together with improved gate leakage. The high frequency properties as well as the on-resistance of ~20 Ohm mm are all limited by the 2000 Ohm/square sheet resistance of the channel layer.

cond-mat.mtrl-sci↗
arXiv2026

Landscape-Aware Bandit Hyper-Heuristics for Online Operator Selection in UAV Inspection Routing

UAV multi-site inspection often reduces to choosing a high-quality visiting order after target sites have been extracted from a map. This paper develops LA-BHH, a landscape-aware bandit hyper-heuristic that learns an operator-selection policy online for this routing layer. LA-BHH treats 2-opt, swap, relocate, and Or-opt moves as low-level arms, builds context from static landscape descriptors and online search-state features, and updates a LinUCB controller from improvement rewards during the same run. Experimental results on 45 generated Euclidean TSP instances show that LA-BHH achieves the best mean final gap and convergence AUC, with 0.0223 and 0.0389 respectively. It reduces final gap by 17.6\% over UCB-HH, 22.6\% over Random-HH, and 68.2\% over nearest-neighbor construction. Ablation results further show that contextual credit assignment, 2-opt repair, and stagnation-aware state use are the main contributors.

cs.CE↗
arXiv2026

Agent^2 RL-Bench: Can LLM Agents Engineer Agentic RL Post-Training?

We introduce Agent2 RL-Bench, a compact diagnostic benchmark for evaluating agentic RL post-training, which tests whether LLM agents can autonomously design, implement, debug, and execute post-training pipelines that improve foundation models. RL post-training increasingly drives model alignment and specialization, yet existing benchmarks are largely static, rewarding supervised fine-tuning or script generation without assessing an agent's ability to close an interactive RL loop. Agent2 RL-Bench provides a unified agent-facing interface: each run starts from an isolated workspace containing a base model, task data, instructions, and a grading API, and agents must iterate within a fixed budget by training models and submitting artifacts for evaluation. The benchmark spans six tasks across three levels, from static rule-based training to judge-based optimization and closed-loop online RL with trajectory collection. Two diagnostic skills, namely runtime recording and post-hoc summarization, enable structured analysis of agent behavior, facilitating smooth and effective iteration of the benchmark's evaluation framework. Across five agent systems and six driver LLMs, agents show intelligent behavior but clear limitations: one RL-oriented run improves ALFWorld from 4.85 to 93.28 via SFT warm-up and GRPO with online rollouts, yet DeepSearchQA remains difficult, most successful routes rely on supervised pipelines, and interactive outcomes show large single-run differences across agent stacks. Overall, Agent2 RL-Bench shows that current agents can sometimes engineer online RL, but stable agent-driven RL post-training remains rare under fixed budgets. It also demonstrates that our benchmark provides a strong and effective evaluation framework for future research in this direction. Code is available at https://github.com/microsoft/RD-Agent/blob/main/rdagent/scenarios/rl/autorl_bench/README.md

cs.AI↗
arXiv2026

WASHH: An Anchor-Aware Whale-Guided Selection Hyper-Heuristic for Continuous Optimization and SVC Configuration

Learning-assisted algorithm design often has to make reliable search decisions under small evaluation budgets, where committing to a single metaheuristic can be unreliable. We propose WASHH, a Whale-guided Adaptive Selection Hyper-Heuristic for continuous black-box optimization. WASHH uses WOA as the main exploitation backbone, but treats PSO-style memory, GWO-style leader averaging, DE-style variation, local coordinate search, and anchor-guided refinement as selectable search behaviors. An online reward controller allocates evaluations according to observed improvements, while anchor refinement exploits inexpensive reference configurations such as box centers or default model settings without bypassing black-box evaluation. On ten 30-dimensional benchmark functions with 10 independent runs and 12,000 evaluations, WASHH achieves the best average rank, 1.10, and is best or tied best on all ten functions. It strictly improves over WOA on eight functions and ties WOA at the numerical optimum on Rastrigin and Griewank. We further study SVC hyperparameter configuration for breast cancer diagnosis under a 300-evaluation budget. WASHH obtains the lowest mean validation log loss among the compared optimizers, suggesting that anchor-aware selection hyper-heuristics are a practical lightweight direction for LEAD systems.

cs.NE↗
arXiv2026

Covering Human Action Space for Computer Use: Data Synthesis and Benchmark

Computer-use agents (CUAs) automate on-screen work, as illustrated by GPT-5.4 and Claude. Yet their reliability on complex, low-frequency interactions is still poor, limiting user trust. Our analysis of failure cases from advanced models suggests a long-tail pattern in GUI operations, where a relatively small fraction of complex and diverse interactions accounts for a disproportionate share of task failures. We hypothesize that this issue largely stems from the scarcity of data for complex interactions. To address this problem, we propose a new benchmark CUActSpot for evaluating models' capabilities on complex interactions across five modalities: GUI, text, table, canvas, and natural image, as well as a variety of actions (click, drag, draw, etc.), covering a broader range of interaction types than prior click-centric benchmarks that focus mainly on GUI widgets. We also design a renderer-based data-synthesis pipeline: scenes are automatically generated for each modality, screenshots and element coordinates are recorded, and an LLM produces matching instructions and action traces. After training on this corpus, our Phi-Ground-Any-4B outperforms open-source models with fewer than 32B parameters. We will release our benchmark, data, code, and models at https://github.com/microsoft/Phi-Ground.git

cs.CV↗
arXiv2026

Low-Cost Labels, Reliable Choices: Rollout-Calibrated Hyper-Heuristics for Job Shop Scheduling

Learning-assisted hyper-heuristics can select among dispatching rules while preserving the feasibility and interpretability of constructive Job Shop Scheduling Problem (JSSP) heuristics. Their main computational cost lies in label generation rather than model fitting, since each supervised label usually requires rolling out candidate rules from a partial schedule. We study this label-cost problem together with a reliability problem: a learned selector should not switch away from a strong default rule unless the predicted gain is credible. The proposed selector uses regret-normalized rollout labels, a contextual KNN uncertainty estimate, and a gate that acts only when the predicted improvement exceeds an uncertainty-adjusted margin. We also vary rollout depth and breadth to measure the cost-quality trade-off. On synthetic JSSP instances, the gated selector achieves the lowest mean RPD among learned selectors, remains close to the best fixed dispatching rule, and reduces Random-HH mean RPD by more than an order of magnitude.

cs.AI↗
arXiv2026

XPERT: Expert Knowledge Transfer for Effective Training of Language Models

Mixture-of-Experts (MoE) language models organize knowledge into explicitly routed expert modules, making expert-level representations traceable and analyzable. By analyzing expert activation patterns in MoE large language models (LLMs), we find that a subset of experts is consistently activated across diverse knowledge domains. These common experts encode cross-domain, generalizable knowledge that is closely related to model generalization, naturally raising the question of how such identifiable expert knowledge can be practically reused. Motivated by this observation, we propose XPERT, a framework that extracts, consolidates, and reuses expert knowledge from pre-trained MoE LLMs to support more effective training of language models across different model scales. XPERT identifies cross-domain experts via inference-only analysis, refines their representations through tensor decomposition, and adapts the extracted knowledge to reuse in downstream models. Experiments on language understanding and dialogue generation benchmarks show that models benefiting from reused expert knowledge achieve consistently stronger performance and faster convergence compared to strong baselines. These results highlight MoE LLMs as structured and reusable knowledge sources, and demonstrate the value of expert-level knowledge reuse for improving model training.

cs.CL↗
arXiv2026

Understanding Performance Collapse in Layer-Pruned Large Language Models via Decision Representation Transitions

Layer pruning efficiently reduces Large Language Model (LLM) computational costs but often triggers sudden performance collapse. Existing representation-based analyses struggle to explain this mechanism. We propose studying pruning through decision representation. Focusing on multiple-choice tasks, we introduce two metrics, Decision Margin and Option Frequency, and an Iterative Pruning method to analyze layer-wise decision dynamics. Our findings reveal a sharp decision transition that partitions the network into two stages: a Silent Phase, where the model cannot yet predict the correct answer, and a Decisive Phase, where the correct prediction emerges. We also find that pruning the Decisive Phase has minimal impact, whereas pruning the Silent Phase triggers immediate performance collapse, highlighting its extreme sensitivity to structural changes. Therefore, we conclude that pruning-induced collapse stems from disrupting the Silent Phase, which prevents the critical decision transition from occurring.

cs.CL↗