Searcharxiv⌕ Search

arXiv subjects

Yu Xia

Publications and source records attributed to Yu Xia.

At least 19 recordsLinked to original sources
arXiv2026

Measuring Mean Spanwise Flow with a Rotating Single Hot-Wire Probe over a Wide Yawed Riblet-Textured Surface

This study investigates the spatial development and recovery of near-wall flow steering induced by yawed widely-spaced riblets. To accurately measure the mean spanwise velocity in the near-wall region, we developed a unique, customised rotating probe equipped with a single hot-wire. This novel technique demonstrates high sensitivity, reliably detecting small spanwise velocity components. We applied this probe to map the flow over straight-to-yawed and yawed-to-straight riblet configurations. For the straight-to-yawed transition, the downstream distribution of the mean spanwise velocity closely aligns with the step-forced spatial Stokes layer (SSL) solution, achieving an equivalent active wall motion of $V^+_\mathrm{equ} \approx 2.0$. This offers a simplified analytical pathway for future passive flow manipulation studies. Conversely, flow recovery over the yawed-to-straight configuration diverges from the SSL step-down phase prediction. Instead, a spatial lag was observed where crest-adjacent flow recovered quickly, but significant residual spanwise flow ($V^+ > 0.5$) persisted far downstream ($\widehat{x}{}^{+} > 1000$).

physics.flu-dyn↗
arXiv2026

Caliber: Cross-Architecture Extraction-Cost Control for Score-Returning APIs

We present Caliber, an output-perturbation defense against model extraction that formulates noise selection as a calibration problem: how much the defense degrades the supervision signal used to train a surrogate, and the provable per-input query cost of recovering the clean logits. To defend against an attacker that uses returned scores for knowledge distillation, Caliber adds independent and identically distributed Gaussian noise to the internal logits. We establish two properties of the resulting perturbed predictions. Monotone agreement degradation: When the clean logits have a unique maximizer, agreement with the clean prediction decreases strictly with the noise scale, so every target in $(1/K,1)$ corresponds to a unique positive scale; task accuracy is bounded by computable lower and upper envelopes. Per-input recovery cost: We derive a closed-form minimax lower bound on the repeated queries needed to recover the clean logits for a fixed input. Caliber normalizes noise variance by the squared median top-two logit margin and fits the resulting noise-utility relationship with a logistic curve, either per model or shared within a task. Across more than thirty model-dataset combinations, per-model calibration achieves mean absolute relative errors of 0.6-1.4%. End-to-end experiments show that surrogate performance generally tracks the configured degradation, while fixed-input averaging follows the expected variance reduction.

cs.LG↗
arXiv2026

HetGPS: Scalable Graph Multi-Agent Reinforcement Learning with Physics-Anchored Adaptive Safety for EV Charging

Safety interventions for large populations of network-coupled agents must protect shared constraints without unnecessarily overriding task-oriented policy decisions. We present HetGPS, a hybrid graph-control framework synergizing learned graph risk with physics-anchored correction by separating intervention magnitude from corrective direction. An action-conditioned graph residual model schedules state-dependent intervention authority, while a physics model determines its direction. For electric vehicle (EV) charging, we couple this filter with a parameter-shared heterogeneous graph soft actor-critic policy, enabling topology-aware coordination with a learned model size independent of fleet size. Across five nested distribution networks with 200--3,218 EVs and 100 evaluation days, Adaptive Authority reduces bus--step voltage violations from 3.93--7.74\% without filtering to 0.52--3.44\%, while maintaining 99.06--100\% departure success. Relative to the same physics-directed projection with fixed authority, it improves mean reward on all five networks and lowers the mean safety score on four. The deployed policy-and-risk model contains 383,702 learned parameters at every scale; at 3,218 EVs, a matched centralized SAC actor is about $170\times$ larger. A policy trained on the eight-transformer system transfers zero-shot to the 16- and 32-transformer systems, attaining 0.57--0.75\% violation rates and at least 99.99\% departure success. These results show that learned graph risk can allocate intervention authority at scale while feeder physics anchors corrective action.

cs.AI↗
arXiv2026

LaRec: Unleashing LLM-based Latent Reasoning for Generative Recommendation

Large Language Models (LLMs) have shown great promise in recommendation due to superior reasoning abilities. However, existing methods mainly rely on explicit Chain-of-Thought (CoT), resulting in verbose reasoning texts and inefficient response times. latent reasoning aims to balance efficiency by thinking within a continuous latent space, yet it faces two major challenges: (1) Lack of Fine-grained Supervision: Latent reasoning relies solely on feedback from the final labels, providing sparse supervisory signals that struggle to effectively guide the optimization of multiple hidden reasoning steps. (2) Single Reasoning Path: The deterministic nature of latent reasoning impedes the exploration of users' diverse interests and preferences, thereby limiting the recommendation capabilities of LLMs. To address these issues, we propose \textbf{$LaRec$}, an efficient generative recommendation framework designed to unleash the potential of latent reasoning in LLMs. $LaRec$ consists of two core stages: First, we design Latent Pre-training that empowers LLMs with latent reasoning capabilities by providing rich supervisory signals to the latent space reasoning via step-level alignment and process direction alignment. Second, we introduce Personalized RL-tuning. Specifically, we construct a personalized Gaussian Mixture Distribution for each user based on their historical interests. By randomly sampling distinct reasoning starting points from this distribution during training, we guide the LLMs to traverse diverse reasoning paths within the latent space, enabling efficient exploration of user's multi-faceted interests. Experiments on multiple datasets show that $LaRec$ significantly outperforms existing baselines with comparable efficiency.

cs.IR↗
arXiv2026

OvisOCR2 Technical Report

We introduce OvisOCR2, a 0.8B document parsing model. OvisOCR2 is designed as an end-to-end parser: given a document page image, it generates a Markdown representation in natural reading order, covering text, formulas, tables, and visual regions. We build a data engine that combines filtered real-document annotations with synthetic pages whose rendered images and Markdown targets are derived from the same HTML source. The training recipe includes supervised fine-tuning, reinforcement learning on a 4B branch with a multi-component reward design, on-policy distillation into the 0.8B model, and model fusion. On OmniDocBench v1.6, OvisOCR2 achieves a state-of-the-art overall score of 96.58, placing an end-to-end model at the top of this leaderboard previously dominated by pipeline methods and highlighting the potential of end-to-end document parsing. On PureDocBench, OvisOCR2 also achieves the highest Avg3 score of 75.06. Beyond these two public benchmarks, we evaluate OvisOCR2 on an in-house benchmark designed to cover a broader set of long-tail and challenging scenarios. OvisOCR2 obtains the best overall performance among the compared methods, providing further evidence of its generalization and robustness. OvisOCR2 is available at https://huggingface.co/ATH-MaaS/OvisOCR2.

cs.CV↗
arXiv2026

Fourier Phase Retrieval for Finite Unions of Intervals

This paper investigates the one-dimensional Fourier phase retrieval problem for indicator functions of finite unions of intervals. Specifically, we study the recovery of a set $\Omega = \bigcup_{j=1}^m I_j \subset\mathbb{R}$ from the magnitude of its Fourier transform $|\widehat{\mathbf{1}_\Omega}|$, where each $I_j \subset \mathbb{R}$ is a bounded interval. For $m\le 2$, we prove that $\Omega$ is uniquely determined by $ |\widehat{\mathbf{1}_\Omega}|$ up to the natural ambiguities of translation and reflection, and we further establish a stability result for this reconstruction. In contrast, for $m\ge 3$, uniqueness fails in general. More precisely, for every $m\ge 3$, we explicitly construct functions $f_m,g_m\in\mathcal{I}_m$ such that $|\widehat{f_m}|=|\widehat{g_m}|,$ while $f_m$ cannot be obtained from $g_m$ by any translation or reflection, where $\mathcal{I}_m$ denotes the class of indicator functions of unions of exactly $m$ intervals. Furthermore, building on the theory of the turnpike problem, in which a finite integer set is uniquely determined by its multiset of pairwise differences under a collision-free condition, we establish an analogous result for finite subsets of $\mathbb{R}$. This, in turn, yields a sufficient condition for recovering indicator functions of finite unions of intervals. These results provide a complete characterization of the Fourier phase retrieval problem for indicator functions of finite unions of intervals and offer new insights into Fourier phase retrieval for indicator functions of more general domains in higher dimensions.

math.FA↗
arXiv2026

LLM-Based User Personas for Recommendations at Scale

Large Language Models (LLMs) offer unprecedented potential for enhancing recommendation systems through their world knowledge and reasoning capabilities. However, existing approaches often rely on structured IDs or offline processing, limiting semantic richness, real-time adaptability, and user-facing interpretability. In this paper, we introduce a novel framework that enables real-time generation of LLM-based user interest personas for a large-scale commercial video recommendation platform. Our method generates natural-language user interest personas that address the exploitation-exploration trade-off by combining the summarization of existing interests with novel topics, directly during serving. To overcome the computational challenges of online LLM inference at a billion-user scale, we design a cost-efficient architecture leveraging knowledge distillation, asynchronous inference, and input optimization via semantically clustered video representations. Extensive offline evaluations, user studies, and live A/B tests demonstrate significant improvements in viewer value. This work bridges the gap between high-level semantic understanding and industrial-scale recommendation, paving the way for more dynamic, explainable, and satisfying personalized experiences.

cs.IR↗
arXiv2026

Kwai Keye-VL-2.0 Technical Report

We introduce Kwai Keye-VL-2.0-30B-A3B, an open-source Mixture-of-Experts (MoE) multimodal foundation model designed to advance long-video understanding and agentic intelligence. To address the challenges of ultra-long contexts, information redundancy, and prohibitive computational costs inherent in hour-level videos, Keye-VL-2.0 is the first to adapt DeepSeek Sparse Attention (DSA) to GQA-based multimodal architectures, enabling lossless 256K context processing while capturing critical frames and long-range temporal dependencies. This architecture is underpinned by a highly optimized training and inference infrastructure, including scalable video I/O, heterogeneous ViT-LM parallelism, and custom DSA kernels that significantly maximize throughput and minimize computational overhead. Furthermore, to overcome the algorithmic dilemma of catastrophic forgetting during multi-task alignment, we introduce Cross-Modal Multi-Teacher On-Policy Distillation (MOPD) paired with Context-RL and Video-RL. By distilling dense token-level teacher feedback from on-policy rollouts back into the MoE backbone, which activates only 3B parameters, Keye-VL-2.0 natively empowers advanced agent collaboration across Code, Tool, and Search scenarios with multimodal self-correction. Extensive evaluations across video understanding, temporal grounding, reasoning, STEM, and agent benchmarks demonstrate that Keye-VL-2.0-30B-A3B achieves state-of-the-art performance among models of similar scale, particularly excelling in fine-grained temporal localization on TimeLens and long-video comprehension on Video-MME-v2 and LongVideoBench. We release our model checkpoints to accelerate community progress toward scalable and robust multimodal agentic applications.

cs.CV↗
arXiv2026

Agentic Chain-of-Thought Steering for Efficient and Controllable LLM Reasoning

Large language models improve final-answer accuracy through extended chain-of-thought reasoning, but often spend tokens inefficiently and offer little inference-time control. Existing efficient reasoning methods control thinking length by shortening, early-stopping, or compressing traces, leaving how the model thinks implicit. In this paper, we propose Agentic Chain-of-Thought Steering (ACTS), which formulates reasoning steering as a Markov decision process where a controller agent adaptively steers a frozen reasoner during inference. At each step, the controller observes the reasoning trace and remaining thinking budget, then issues a steering action consisting of a reasoning strategy and a steering phrase that initiates the next reasoner step. This enables budget-aware strategy control for efficient reasoning while preserving the reasoner's generation continuity. We initialize the controller agent from our constructed synthetic steering trajectories with multi-budget augmentation, and further optimize it via reinforcement learning with budget-conditioned reward shaping. Experiments across multiple benchmarks show that ACTS achieves competitive accuracy with substantial token savings, and enables controllable accuracy-token trade-offs across different reasoners and tasks. The code is available at https://github.com/Andree-9/ACTS.

cs.CL↗
arXiv2026

UAV-OVO: Out-of-Viewpoint Generalization in UAV Action Recognition

UAV action recognition faces a deployment shift that standard benchmarks often obscure: a model trained on UAV footage captured from low-depression viewpoints may be required to recognize the same action classes from high-depression viewpoints. While the action labels remain unchanged, this shift alters body visibility, motion projection, and scene context, encouraging models to rely on viewpoint-specific shortcuts. We introduce UAV-OVO, an Out-of-Viewpoint generalization benchmark for UAV action recognition. UAV-OVO derives view scores from uncalibrated videos, uses a view-isolation band to assign low-depression videos to the training and in-distribution test splits while reserving high-depression videos for out-of-distribution testing, and constructs ID/OOD test sets matched by class distribution so that performance differences reflect viewpoint shift rather than label imbalance. Across representative video recognizers, UAV-OVO reveals a substantial ID/OOD gap: models that fit the low-depression training distribution well often fail to transfer to held-out high-depression views, exposing viewpoint shortcuts hidden by aggregate accuracy. We further propose LATER, LoRA-Anchored Test-time Re-centering, which first adapts the recognizer with Low-Rank Adaptation (LoRA) and then uses the learned LoRA subspace as a semantic anchor for online feature re-centering. Specifically, LATER projects target-domain displacement onto the orthogonal complement of the LoRA subspace before re-centering features, reducing viewpoint-induced drift while preserving task-relevant semantics. Together, UAV-OVO and LATER provide a controlled testbed and a practical adaptation method for viewpoint-robust UAV video understanding.

cs.CV↗
arXiv2026

Overcoming noise-agility trade-off in integrated lasers for precision sensing

Lasers that combine narrow linewidths with rapid tunability are critical for applications such as coherent optical ranging, distributed fiber-optic sensing, and precision spectroscopy. Despite significant progress in integrated laser technologies, the concurrent realization of low phase noise and frequency agility on a single integrated platform remains challenging owing to a fundamental architectural trade-off: conventional integrated laser designs typically suppress phase noise via high-$Q$ resonators, yet the extended photon lifetimes inherent to such resonators intrinsically constrain tuning speed. Here, we address this noise-agility trade-off by introducing a laser architecture that achieves ultralow phase noise and ultrafast tunability simultaneously. Rather than relying on ultrahigh-$Q$ resonators for self-injection locking, our design employs strong synthetic feedback within a Pockels-tunable, resonator-enhanced distributed Bragg reflector to suppress phase noise. As a proof of concept, we demonstrate a hybrid integrated laser with a short-term linewidth of 29 Hz, realized using a lithium niobate external cavity with a loaded $Q$ of only 0.62 million. The adoption of a moderate resonator $Q$ relaxes the photon-lifetime constraint on tuning speed, enabling sub-exahertz-per-second tuning rates and a chirp nonlinearity as low as 0.14%. Leveraging this laser, we implement a frequency-modulated continuous-wave LiDAR system that achieves a relative ranging precision of $1.7 \times 10^{-4}$ at a measurement rate of $1\,\text{MSa s}^{-1}$, without requiring complex chirp linearization techniques. We further demonstrate fiber-optic acoustic sensing capable of detecting sub-$\mu\epsilon$ dynamic strain, underscoring the platform's versatility for high-speed precision optical measurements. Our work provides a route toward cost-effective yet high-performance sensing and metrology systems.

physics.optics↗
arXiv2026

THEMol dataset: Torsion, Hessian, and Energy of Molecules

We present THEMol (Torsion, Hessian, Energy of Molecules), a massive open-source collection of quantum mechanical properties tailored for closed-shell organic molecules, with up to 50 heavy atoms. THEMol includes a Hessian subset with more than 3 million relaxed geometries with Hessian matrices, a TorsionScan subset with nearly 100 million constrained relaxed geometries with energies and forces, and relaxation-trajectory subsets (HessianRelax and TorsionScanRelax) that together comprise about 3 billion DFT calculations. The chemical space sampling is comprehensive, spanning twelve essential elements and diverse molecular architectures relevant to drug discovery, electrolytes, ionic liquids, and beyond. The dataset also features exhaustive conformational sampling through the TorsionScan and TorsionScanRelax subsets, including comprehensive in-ring and non-ring torsional scans. Furthermore, it contains an extensive library of Hessian matrices, computed at relaxed geometries, to capture critical second-derivative information of the potential energy landscape. Additionally, we supply electron density-derived atomic multipoles computed via the Minimal Basis Iterative Stockholder partition scheme. Organized into five distinct subsets (Hessian, TorsionScan, HessianRelax, TorsionScanRelax, and MBIS), the data encompasses optimized geometries, relaxation trajectories, and derived molecular properties. We anticipate that this massive and diverse dataset will significantly empower the development of highly accurate and transferable molecular potentials.

physics.chem-ph↗
arXiv2026

A Survey on LLM-based Conversational User Simulation

User simulation has long played a vital role in computer science due to its potential to support a wide range of applications. Language, as the primary medium of human communication, forms the foundation of social interaction and behavior. Consequently, simulating conversational behavior has become a key area of study. Recent advancements in large language models (LLMs) have significantly catalyzed progress in this domain by enabling high-fidelity generation of synthetic user conversation. In this paper, we survey recent advancements in LLM-based conversational user simulation. We introduce a novel taxonomy covering user granularity and simulation objectives. Additionally, we systematically analyze core techniques and evaluation methodologies. We aim to keep the research community informed of the latest advancements in conversational user simulation and to further facilitate future research by identifying open challenges and organizing existing work under a unified framework.

cs.CL↗
arXiv2026

Meter-long broadband chirped Bragg gratings for on-chip dispersion control and pulse shaping

Precise on-chip dispersion control is essential for advanced integrated photonic technologies, enabling applications ranging from high-speed communications and sensing to signal processing and biomedical imaging. However, existing on-chip dispersion control methods still suffer from substantial loss and a limited dispersion-bandwidth product (DBP) far from application needs. As a result, on-chip systems continue to rely exclusively on off-chip dispersion control solutions provided by optical fiber or bulky free-space optics. To overcome these limitations, we design and fabricate meter-long chirped spiral Bragg gratings (CSBGs) on the ultra-low-loss silicon nitride (SiN) photonic platform for advanced dispersion control. Our device achieves a 10-nanosecond group delay with customizable bandwidths exceeding 10 nanometers within a compact footprint of only 30 $\text {mm} ^2$, surpassing the physical limits of fiber-based grating devices. More importantly, CSBGs can simultaneously possess the characteristics of high stability, low latency, and a large DBP, thanks to the ultra-low-loss SiN platform with a loss of only 0.3 dB/m. Leveraging the precise and stable dispersion profile, we demonstrate high-fidelity pulse shaping and compression of electro-optic frequency combs (EOCs) with a 1-GHz repetition rate centered across the entire reflection bandwidth. The compressed pulse has an on-chip peak (average) power of 21.6 watts (580 milliwatts). Furthermore, we showcase for the first time the application of on-chip pulse-compressed EOC in wavelength-swept coherent anti-Stokes Raman scattering (CARS) microscopy. Our work provides integrated photonics with a long-sought, scalable, and robust solution for high-performance on-chip dispersion control, empowering a new generation of on-chip functionalities.

physics.optics↗
arXiv2026

Scalable 3D silicon nitride photonic interposer for high-density optical interconnects

Modern computing workloads demand energy-efficient, high-bandwidth interconnects, motivating photonic interposers as an alternative to electrical links. Here we demonstrate a compact 3D silicon nitride (SiN) photonic interposer prototype comprising two routing layers, with the 3D routing scheme optimized by a global optimization algorithm. The 3D interposer realizes a fully connected 12-node optical network that reduces the total number of intralayer crossings from 495 for all-planar routing to merely 150 (69.7% reduction), below the theoretical lower bound of 153 for all-planar interconnects. Comparing the two schemes, our 3D design achieves a 45.8% reduction experimentally in the average loss per waveguide. The proposed 3D routing architecture also features inherent symmetry and is scalable to higher node counts, flexible node placements, additional routing layers, and other operating wavelengths, enabling denser, lower-loss photonic interposers for next-generation scale-up and high-performance computing (HPC) systems.

physics.optics↗
arXiv2026

Building reliable 3D photonic integrated circuits and cavities at the wafer scale

Three-dimensional (3D) photonic integrated circuits (PIC) are emerging as an indispensable scheme for high density and multifunctional photonic systems. However, the wafer-scale scaling of PICs towards a 3D configuration is constrained by two key factors: (i) the trade-off between inter-layer taper efficiency and footprint, and (ii) wafer-scale uniformity of inter-layer transition loss. In this work, we introduce etch-back assisted chemical mechanical polishing (E-CMP) to achieve high wafer-scale uniformity of the spacer layer. Moreover, we break the efficiency-footprint trade-off by demonstrating a novel $\kappa$-engineered taper, achieving a reliability metric that is 75\% higher than the traditional linearly tapered structure. Building on these design and fabrication developments, we enable reliable 3D PICs with typical loss of 0.077 and 0.068 dB/cm on two silicon nitride (SiN) waveguide layers and typical 3D transition loss as low as 6 mdB. Furthermore, the low 3D transition loss enables the first class of 3D high-Q optical cavities occupying two distinct device layers, providing new design space for high-Q optical cavities. The scalable fabrication process and design methodology provide routes for wafer-scale reliable 3D PICs that are promising in a series of applications ranging from photonic interconnects and computing networks to high-density photonic sensors and nonlinear photonics.

physics.optics↗
arXiv2026

Walk the Talk: Bridging the Reasoning-Action Gap for Thinking with Images via Multimodal Agentic Policy Optimization

Recent advancements in Multimodal Large Language Models (MLLMs) have incentivized models to ``think with images'' by actively invoking visual tools during multi-turn reasoning. The common Reinforcement Learning (RL) practice of relying on outcome-based rewards ignores the fact that textual plausibility often masks executive failure, meaning that models may exhibit intuitive textual reasoning while executing imprecise or irrelevant visual actions within their agentic reasoning trajectories. This reasoning-action discrepancy introduces noise that accumulates throughout the multi-turn reasoning process, severely degrading the model's multimodal reasoning capabilities and potentially leading to training collapse. In this paper, we introduce Multimodal Agentic Policy Optimization (MAPO), bridging the gap between textual reasoning and visual actions generated by models within their Multimodal Chain-of-Thought (MCoT). Specifically, MAPO mandates the model to generate explicit textual descriptions for the visual content obtained via tool usage. We then employ a novel advantage estimation that couples the semantic alignment between these descriptions and the actual observations with the task reward. Theoretical findings are provided to justify the rationale behind MAPO, which inherently reduces the variance of gradients, and extensive experiments demonstrate that our method achieves superior performance across multiple visual reasoning benchmarks.

cs.CV↗
arXiv2026

Learning to Hint for Reinforcement Learning

Group Relative Policy Optimization (GRPO) is widely used for reinforcement learning with verifiable rewards, but it often suffers from advantage collapse: when all rollouts in a group receive the same reward, the group yields zero relative advantage and thus no learning signal. For example, if a question is too hard for the reasoner, all sampled rollouts can be incorrect and receive zero reward. Recent work addresses this issue by adding hints or auxiliary scaffolds to such hard questions so that the reasoner produces mixed outcomes and recovers a non-zero update. However, existing hints are usually fixed rather than adapted to the current reasoner, and a hint that creates learning signal under the hinted input does not necessarily improve the no-hint policy used at test time. To this end, we propose Hint Learning for Reinforcement Learning (HiLL), a framework that jointly trains a hinter policy and a reasoner policy during RL. For each hard question, the hinter generates hints online conditioned on the current reasoner's incorrect rollout, allowing hint generation to adapt to the reasoner's evolving errors. We further introduce hint reliance, which measures how strongly correct hinted trajectories depend on the hint. We derive a transferability result showing that lower hint reliance implies stronger transfer from hinted success to no-hint success, and we use this result to define a transfer-weighted reward for training the hinter. Therefore, HiLL favors hints that not only recover informative GRPO groups, but also produce signals that are more likely to improve the original no-hint policy. Experiments across multiple benchmarks show that HiLL consistently outperforms GRPO and prior hint-based baselines, demonstrating the value of adaptive and transfer-aware hint learning for RL. The code is available at https://github.com/Andree-9/HiLL.

cs.LG↗