arXiv subjects

Xiang Gao

Publications and source records attributed to Xiang Gao.

At least 37 records · Page 2Linked to original sources

arXiv2026

ReMP: Low-Downtime Runtime Model-Parallelism Reconfiguration for LLM Serving

Current large language model (LLM) inference systems universally deploy ultra-large-scale models using a combination of Tensor Parallelism (TP) and Pipeline Parallelism (PP). However, existing systems treat the model parallelism topology as a static configuration that cannot be flexibly adjusted at runtime. This rigid design creates a fundamental contradiction with the dynamically changing inference workloads in real-world scenarios. State-of-the-art systems lack online reconfiguration capabilities and can only switch configurations by restarting the service, resulting in several minutes of service interruption, KV cache loss, and prohibitive recomputation overhead. To address this problem, this paper presents ReMP, a runtime model parallelism reconfiguration framework that supports low downtime. ReMP achieves dynamic adjustment through three key techniques: (1) decoupling the model parallelism topology from runtime state to avoid full service reconstruction; (2) designing a two-dimensional KV cache migration mechanism to preserve reusable cache states after TP/PP changes; and (3) implementing end-to-end online reconfiguration. Experiments demonstrate that ReMP can complete most topology switches within 1-7 seconds on models ranging from 7B to 70B parameters, achieving speedups of tens to over a hundred times compared to the restart approach. Moreover, ReMP significantly outperforms fixed configurations under dynamic workloads, delivering superior performance in terms of TTFT, TPOT, and output throughput.

cs.DC↗

arXiv2026

ClockRoPE: Random Fourier Rotations for Temporal Routine Modeling

Rotary Position Embedding (RoPE) has been widely adopted in transformer-based large language models. However, its log-linear frequency schedule, originally designed to produce long-term attention decay, limits its adoption in domains with more complex distance-correlation patterns, such as temporal periodicity in sequential recommendation. We investigate the expressiveness of general query/key rotations and find that any normalized continuous positive-definite attention modulation function can be approximated by random rotations induced by its own Fourier transform, which we term Random Fourier Rotations. Building on this theory, we propose ClockRoPE for routine modeling in sequential recommendation, where rotation frequencies are derived from periodic attention modulation functions. In online A/B tests, ClockRoPE demonstrates consistent improvements in valued engagement metrics, and has been successfully deployed in production-scale generative retrieval system at a major video-sharing platform.

cs.LG↗

arXiv2026

DREAM Technical Report

Industrial recommender systems commonly use cascaded retrieval, ranking, and re-ranking pipelines. Although efficient, these pipelines fragment information and objectives across modules, rely on rigid rules, and have limited awareness of real-time intent, leaving session-level shifts among browsing, comparison, and purchase insufficiently addressed. We present DREAM (Developing Recommender Engine with Agentic Methods), an autonomous optimization control architecture that adds a perception-aware, orchestrable, and auditable policy layer atop existing pipelines without replacing them. DREAM has two core components. First, a three-tier Intent Engine fuses on-device signals into structured L0/L1/L2 intent representations; its edge-cloud trigger chain reduces reporting volume to approximately 8.7%. Second, a Meta Engine uses a MetaModel for layered M1-to-M2-to-M3 reasoning: intent summarization, strategy planning informed by Strategy Memory, and parameter translation. It dispatches the resulting parameters through a unified outlet with safety guardrails. A Reward Dual Loop continuously optimizes both components by combining offline simulation for strategy-space exploration with online feedback for outcome calibration, forming a cycle of generation, execution, evaluation, and experience accumulation. Large-scale A/B tests on Taobao's homepage feed show that re-ranking control alone improves IPV by 2.06%, Core IPV by 2.39%, and GMV by 0.88%. Extending control to fine ranking raises these gains to 2.71%, 3.06%, and 1.31%, respectively, while consistently improving PV by more than 1%. These gains require neither replacement of pipeline models nor compromise of serving stability, supporting agentic meta-control as a viable paradigm for industrial recommendation.

cs.IR↗

arXiv2026

CABS+: Efficient and Scalable Model Merging via Conflict-Aware Sparsification and Adaptive Weight Allocation

Model merging has recently attracted significant attention as a promising paradigm for constructing unified multi-task models without requiring additional retraining. However, parameter conflicts and knowledge interference across tasks often degrade merged-model performance. Prior work introduced Conflict-Aware and Balanced Sparsification (CABS), which reduces parameter interference through structured pruning and sequential masking. However, CABS relies on grid search to determine scaling coefficients, resulting in exponential time complexity, while its optimization objective can be dominated by high-performance tasks, leading to suboptimal overall performance. To address these limitations, we extend CABS and propose CABS+. Specifically, Adaptive Weight Allocation (AWA) optimizes merging coefficients via a gradient-free search scheme to reduce time complexity, while an asymmetric fitness function promotes more comprehensive performance gains across tasks. Moreover, we conduct a systematic empirical study of key factors influencing model merging performance and propose Relative Synergy Score (RSS) to quantify model mergeability and guide model selection. We compare CABS+ with state-of-the-art model merging methods, including CABS, AdaMerging, and WUDIMerging, across 27 datasets and 5 models covering large language, small-scale language, and vision models. Extensive experiments verify the effectiveness and efficiency of CABS+. Compared with AdaMerging and WUDIMerging, CABS+ improves overall performance by 16.97% and 12.93%, respectively, exhibits stronger stability and robustness across varying task numbers and model architectures, uses less than 25% of the GPU memory required by AdaMerging, and achieves nearly a 4x speedup in merging time over WUDIMerging.

cs.AI↗

arXiv2026

TEASR: Training-Efficient Any-Step Diffusion Transformer for Real-World Image Super-Resolution

Diffusion models excel in Real-World Image Super-Resolution (Real-ISR) due to their powerful generative priors but suffer from slow iterative sampling. Although existing one-step distillation methods accelerate inference, they typically require auxiliary teacher models that inflate training memory and restrict scalability to large-scale architectures. Furthermore, these fixed-step models lack the flexibility to trade off speed for quality. In this paper, we propose TEASR, a training-efficient any-step diffusion framework for Real-ISR that enables both one-step and multi-step restoration within a unified model. Our key idea is to perform self-adversarial distillation within a single diffusion model, eliminating the need for auxiliary teachers or discriminators. Specifically, we propose a timestep-aware rectification strategy that stabilizes one-step generation across noise levels. These two designs further enables the distillation of 20B-parameter diffusion models on a single GPU, significantly improving training efficiency. Moreover, we introduce a dual-branch diffusion transformer with decoupled timestep condition to separate the current noise state and the denoising target to enhance sampling quality. Extensive experiments demonstrate that TEASR supports seamless any-step sampling and consistently outperforms state-of-the-art methods across multiple datasets.

cs.CV↗

arXiv2026

Observational Policy Ranking for SMB Financial Guidance from Multi-Action Accounting Logs

Small and medium-sized businesses need timely financial guidance, yet historical accounting logs record self-selected and often co-occurring business changes rather than randomized recommendations. We formulate this setting as observational policy ranking: from pre-decision financial information, a policy selects one of 34 ledger-derived business-change categories for a target financial KPI. Using 85,078 company-month observations from 7,505 firms, we introduce Covariate-Adjusted Residual Policy Learning (CAR-PL), an action-wise R-learner that operates directly on multi-hot logs and regularizes selection by observational support. We compare CAR-PL with an uplift T-Learner, a conservative contextual value model, a zero-shot LLM, and non-personalized references on company-disjoint held-out firms under a shared model-assisted scoring rule. CAR-PL has the highest Gross Profit point estimate (0.084), the T-Learner has the highest Revenue point estimate (0.085), and the contextual value model has the highest Quick Ratio point estimate (0.062). CAR-PL and the T-Learner are not statistically separated on either growth KPI in matched company-clustered comparisons, while CAR-PL selects 33-34 categories and produces less concentrated selections across the catalog. Outcome-model-only scoring retains the same KPI-level point-estimate leader or top pair, and category rankings remain similar when the all-zero treatment reference is replaced by the most common training co-action pattern. These findings support objective-specific ranking of SMB financial guidance from multi-action accounting logs.

cs.LG↗

arXiv2026

Lattice-QCD validation of hadron mass and trace-anomaly decomposition sum rules

We present the first lattice-QCD validation of multiple sum rules associated with quark-gluon decomposition of hadron mass by computing all components from first principles. We achieve this through nonperturbative renormalization of the QCD energy-momentum tensor, including its trace, in a gradient-flow scheme, followed by continuum extrapolations, two-loop matching to the $\overline{\mathrm{MS}}$ scheme, and zero-flow-time extrapolations. These ingredients enable a direct and simultaneous verification, in a common renormalization scheme and scale, of multiple energy-density-based and trace-based mass decomposition sum rules proposed in the literature. We demonstrate the framework for the $η_c$ and $J/ψ$ charmonia using three fine lattice spacings with a physical strange-quark and near-physical up- and down-quark masses. We present the first lattice-QCD results for the gravitational form factor $\bar{C}$. We find sizable gluonic contributions to charmonia masses at the hadronic scale, $\sim 15\%$ in the Lorcé and Metz-Pasquini-Rodini decompositions. The trace-anomaly contribution in the Ji sum rule is $\sim 6\%$, while the gluonic component of the trace anomaly in the Hatta-Rajan-Tanaka sum rule is $\sim 35\%$. The method is general and can be straightforwardly adopted for lattice-QCD calculations of mass and spin decompositions as well as gravitational form factors of other hadrons and nuclei.

hep-lat↗

arXiv2026

Quantifying Depth Sufficiency in Residual Neural Networks: A First-Order Criterion

How can we determine whether a trained neural network is already deep enough? We study this under a fixed function-preserving residual-growth protocol specifying insertion locations, residual families, zero-output initializations, and zero-state first-order updates. We define first-order residual depth saturation as the absence of a strict local decrease from every admissible insertion. We prove residual non-degeneracy is necessary and sufficient: additional depth has first-order value exactly when conditional activation gradients have a nonzero projection onto at least one admissible residual tangent space. This boundary is shared by descent-compatible zero-state updates and invariant under regular local reparameterizations preserving that tangent space. Under residual-signal realizability, raw activation-gradient vanishing exactly certifies saturation. Across ResNets, GPT-2-style models, and continued-pretrained Pythia checkpoints, the maximum activation-gradient norm decreases toward a low-signal regime with depth. Function-preserving growth also achieves converged performance competitive with training from scratch. These results support activation-gradient magnitude as a conservative diagnostic of the remaining empirical first-order value of residual depth.

cs.LG↗

arXiv2026

Universal mapping of drop impact spreading from wetting to Leidenfrost regimes

We establish a universal mapping between the maximum spreading of a drop under wetting and Leidenfrost impact conditions from an energy-dissipation perspective, with the latter featuring a stable vapor film between the drop and the substrate. Experiments and direct numerical simulations demonstrate that the mapping remains valid over two to three orders of magnitude in the Ohnesorge and Weber numbers. This framework provides a simple route for converting established predictions for wetting impacts into their Leidenfrost counterparts, with implications for predicting and controlling droplet impact.

physics.flu-dyn↗

arXiv2026

TENT: A Declarative Slice Spraying Engine for Performant and Resilient Data Movement in Disaggregated LLM Serving

Modern GPU clusters rely on complex, heterogeneous interconnects. As large language model (LLM) serving shifts toward agentic reasoning, KVCache becomes a first-class mobile asset, driving frequent migrations and massive elephant flows that dominate the execution critical path. Operating Mooncake Transfer Engine (TE) on thousands of GPUs exposed a fundamental flaw in existing frameworks: imperative, early-binding path selection. This rigidity results in state-blind striping that ignores congestion and grey failures, leading to bandwidth stranding. This also produces operational fragility where routine faults require manual intervention. We present TENT, a declarative orchestration engine that decouples transfer intent from physical execution. By abstracting interconnects into a unified resource pool, TENT shifts path resolution from initialization to slice-time late binding. Applications simply declare transfer intents, while TENT dynamically "sprays" fine-grained slices across rails based on real-time telemetry and predictive cost modeling. This orchestration eliminates head-of-line (HoL) blocking and enables transparent, sub-50 ms self-healing by rerouting slices around failures or degradations without application-level intervention. TENT serves as the production data plane for LLM inference and reinforcement learning (RL) pipelines at multiple industrial clusters. Our evaluation shows that TE+ outperforms state-of-the-art baselines, including Mooncake TE, NIXL, and UCCL. In LLM inference with SGLang HiCache, TE+ achieves up to 1.36x higher throughput and 26% lower P90 time-to-first-token (TTFT) than Mooncake TE. In RL pipelines, TENT accelerates parameter updates in Moonshot Checkpoint Engine by 20-26%.

cs.DC↗

arXiv2026

Token-Operations-Oriented Inference Optimization Techniques for Large Models

Large model inference optimization serves as a key foundation for supporting the scalable, low-cost, and highly stable operation of large model services. Centered on token-oriented inference optimization technology, this paper proposes for the first time a four-layer technical architecture consisting of Multi-model Fusion, Model Optimization, Compute-Model Fusion, and Compute-Network-Model Fusion. It systematically reviews the key technologies and current industry status across these four levels and analyzes the application value of related technologies in real-world business scenarios. This paper provides a practical technical path for reducing token production costs, improving token service efficiency, ensuring the stability of token supply, and driving the transition of large model services from being merely callable to being operable.

cs.SE↗

arXiv2026

Plasma screening and configuration interaction effects induced large enhancement on L-shell photoionization cross sections and opacity

An opacity model that incorporates improved treatments of both plasma screening and configuration interaction (CI) effects is proposed, and a 25-30% enhancement on the iron L-shell opacity is predicted at solar interior temperatures. It is originated from the plasma screening induced 14-17% enhancement on the photoionization cross sections and the CI induced 10-20% enhancement on photoexcitation and photoionization cross sections for open L-shell ions. These explain the long-standing discrepancy between theoretical and experimental iron opacity [Nature 517, 56], and the relatively weaker enhancements on chromium and nickel opacity [Phys. Rev. Lett. 122, 235001] due to the sensitivity of these effects to the different L-shell electron population and plasma temperature/density. This letter provides the systematic interpretation of L-shell opacity measurements at solar interior temperatures, and advances the accurate simulation of opacity and radiative transport in high-energy-density plasma.

physics.atom-ph↗

arXiv2026

How Does Urban Context Relate to Residential Building Health? A Vision-POI Fusion Framework for Building-Level Housing Inspection

Housing-level urban physical examination is essential for identifying residential building problems and supporting targeted urban renewal. Existing automated inspection studies primarily rely on individual images and rarely examine whether surrounding urban functional context can provide supplementary information for building-level assessment. This study proposes a vision-POI fusion framework that combines multi-view visual inspection with POI-derived neighborhood context for residential building health assessment. The empirical dataset covers 92 old residential communities, 3,237 residential buildings, and 25,608 field-acquired inspection images in Qingdao, China, encompassing seven categories of housing-related issues. First, multiple object detection models are evaluated to extract issue locations, categories, and confidence scores from individual images. The image-level outputs are subsequently aggregated across multiple views to construct interpretable building-level representations. Second, POI features are extracted within 500m, 1,000m, and 1,500m neighborhood buffers to characterize surrounding functional environments. Pearson and Spearman correlation analyses, combined with false discovery rate correction, are used to identify candidate contextual features. Finally, visual and POI features are integrated using a cost-sensitive Random Forest classifier under community-isolated spatial cross-validation. The results show that multi-view aggregation provides the main performance improvement, increasing the building-level Macro-F1 from 60.84% under Direct Detection to 74.95%. Incorporating POI context further increases Macro-F1 to 76.79%, although the additional gain is modest and category-dependent. POI information therefore functions as a supplementary contextual prior rather than a substitute for direct visual evidence or a causal determinant of building condition.

cs.CV↗

arXiv2026

Workflow-GYM: Towards Long-Horizon Evaluation of Computer-use Agentic tasks in Real-World Professional Fields

Recent years have witnessed the rapid evolution of AI agents toward handling increasingly complex, real-world tasks. However, existing benchmarks rarely evaluate whether agents can operate graphical user interfaces to complete long-horizon, high-value professional workflows across diverse domains. Current GUI benchmarks still predominantly focus on general-purpose software, relatively simple applications, and short-horizon tasks, leaving it largely unknown whether modern agents can follow user instructions to autonomously operate domain-specific professional software and accomplish economically valuable work in an end-to-end manner. To bridge this gap, we introduce Workflow-GYM, a benchmark for long-horizon GUI tasks centered on professional domains and specialized software environments. Through extensive experiments on state-of-the-art models, we find that even the strongest models achieve only slightly above 30% success rates, highlighting that professional long-horizon GUI workflows remain highly challenging for current GUI agents. Further analysis reveals that current agents struggle to maintain long-horizon workflow consistency, frequently exhibiting workflow stage omission, error propagation, objective drift, and insufficient understanding of professional software environments. Our findings provide important insights into the limitations of current agent systems and suggest key directions for the next generation of GUI-agent research.

cs.AI↗

arXiv2026

An Empirical Study for Android-to-OpenHarmony GUI Test Migration

To reduce the substantial engineering effort required to test the corresponding applications from Android to OpenHarmony, migrating existing GUI test cases has become a critical problem. However, current research neither proposes solutions tailored for OpenHarmony nor provides a systematic evaluation of migration approaches on this system, leaving developers with limited empirical guidance in practice. In this paper, we present the first systematic empirical study of test migration from Android to OpenHarmony. Specifically, we first construct a dataset referred to as the ATH Benchmark, comprising 36 commercial applications with an average of over 9 billion downloads, along with 108 manually designed test cases. Second, we select two state-of-the-art test migration approaches (i.e., ReSPlay and ITeM) and adapt these two approaches to enable their execution on OpenHarmony. Third, we use the preceding infrastructure to evaluate these two approaches from three perspectives, including testing performance, root causes of failures, and the impact of OpenHarmony characteristics. Our results reveal that existing test migration approaches are less effective (15% success-rate on ReSPlay and 26% success-rate on ITeM) in Android-to-OpenHarmony scenarios. Through an in-depth analysis of failed cases, we identify that test performance is primarily hindered by OpenHarmony-specific characteristics, including technical architecture differences and unique ecosystem traits. Utilizing these findings, we propose an enhanced approach based on ITeM, referred as ITeM-HM, which incorporates specific OpenHarmony system features. As a result, ITeM-HM successfully achieves a 214% success-rate relative improvement over the original ITeM (from 26% to 81%).

cs.SE↗

arXiv2026

Proton isovector helicity PDF at NNLO and the twist-3 moment $\tilde{d}_2$ from lattice QCD at physical quark masses

We present a lattice quantum chromodynamics calculation of the $x$-dependent isovector quark helicity parton distribution function (PDF) of the proton in the large momentum effective theory (LaMET) framework. Through operator product expansion (OPE) we also extract the $\tilde{d}_2$ moment of the twist-3 PDF $g_T(x)$ for the first time in the $\overline{\rm MS}$ scheme, which is proportional to the average color Lorentz force experienced by the quark in the proton. This calculation is performed on a lattice of spacing $a$ = 0.076 fm at physical quark masses. The quasi-PDF matrix elements are measured in proton states boosted to momenta $P_z=\{0, 0.25, 1.02, 1.53\}$ GeV. We first extract the lowest few helicity PDF moments from the renormalization-group (RG) invariant ratios of the matrix elements with OPE. Combined with the matrix elements relevant for $g_T(x)$, we obtain $\tilde{d}_2^{u-d}(2\ {\rm GeV})=0.0024(46)$ at next-to-leading order in $\overline{\rm MS}$. Then, the helicity quasi-PDF matrix elements are renormalized in the hybrid scheme with linear renormalon resummation and Fourier transformed to the $x$-space after an asymptotic extrapolation. The quasi-PDF is perturbatively matched to the $\overline{\rm MS}$ PDF with RG and threshold resummations at next-to-leading power and next-to-next-to-leading logarithmic accuracies. After resummations, we determine the PDF in the region $x\in[0.25,0.75]$ with controlled systematic uncertainties. The end-point regions are then parameterized, combined with the LaMET prediction at moderate $x$, and fitted to the short-distance matrix elements in coordinate space.

hep-lat↗

arXiv2026

RIMRULE: Improving Tool-Using Language Agents via MDL-Guided Rule Learning

Large language models (LLMs) often struggle to use tools reliably in domain-specific settings, where APIs may be idiosyncratic, under-documented, or tailored to private workflows. This highlights the need for effective adaptation to task-specific tools. We propose RIMRULE, a neuro-symbolic approach for LLM adaptation based on dynamic rule injection. Compact, interpretable rules are distilled from failure traces and injected into the prompt during inference to improve task performance. These rules are proposed by the LLM itself and consolidated using a Minimum Description Length (MDL) objective that favors generality and conciseness. Each rule is stored in both natural language and a structured symbolic form, supporting efficient retrieval at inference time. Experiments on tool-use benchmarks show that this approach improves accuracy on both seen and unseen tools without modifying LLM weights. It outperforms prompting-based adaptation methods and complements finetuning. Moreover, rules learned from one LLM can be reused to improve others, including long reasoning LLMs, highlighting the portability of symbolic knowledge across architectures.

cs.CL↗

arXiv2026

Detecting Vulnerability-Inducing Commits via Multi-Stage Reasoning with LLM-Based Agents

Detecting vulnerability-inducing commits (VICs) at submission time is critical for improving the security and reliability of software systems. However, this task is highly challenging because it requires reasoning about the semantic impact of code changes from heterogeneous information sources, including code diffs, commit messages, and the surrounding contextual code. Existing approaches often struggle to fully capture these complex interactions, resulting in limited detection performance. In this paper, we propose VIC-RAGENT, an LLM-based multi-agent framework for effective and explainable vulnerability detection. VIC-RAGENT leverages multiple specialized agents to provide complementary perspectives, including structural analysis, intent understanding, and vulnerability inspection. To further improve detection reliability, the framework employs a multi-stage reasoning process that progressively refines candidate vulnerabilities through preliminary inspection, reanalysis, and a final decision stage. Experimental results on a real-world dataset across multiple LLMs demonstrate that VIC-RAGENT consistently outperforms baselines, including Direct, CoT, and CodeAgent. Compared to the strongest baseline, VIC-RAGENT achieves 1.2-1.7x higher F1-scores across different models. Overall, VIC-RAGENT offers a robust, explainable, and practical solution for detecting VICs in modern software development workflows.

cs.SE↗