arXiv subjects

Yue Wang

Publications and source records attributed to Yue Wang.

At least 181 records · Page 10Linked to original sources

arXiv2026

Rethinking LLMOps for Fraud and AML: Building a Compliance-Grade LLM Serving Stack

Fraud detection and anti-money-laundering (AML) compliance are high-value domains for large language models (LLMs), but their serving requirements differ sharply from generic chat workloads. Compliance prompts are often prefix-heavy, schema-constrained, and evidence-rich, combining reusable policy instructions, risk taxonomies, transaction or document context, and short structured outputs such as JSON labels or risk factors. These properties make prefix reuse, KV-cache efficiency, runtime tuning, model orchestration, and output validation first-order systems concerns. This paper introduces a workload-aware LLMOps stack for fraud and AML workloads using self-hosted open-weight models such as Meta Llama and Alibaba Qwen. The stack combines vLLM-style runtime tuning, PagedAttention, Automatic Prefix Caching, multi-adapter serving, adapter and prompt-length-aware batching, sleep/wake lifecycle management, speculative decoding, and optional prefill/decode disaggregation. To avoid exposing institution-specific data, the reproducibility track converts public synthetic AML datasets, including IBM AML and SAML-D, into prefix-heavy compliance prompts with reusable policy text, transaction evidence, typology definitions, and schema-constrained outputs. We also incorporate an LLM-as-judge quality gate using deterministic compliance checks, reference metrics, expert-adjudicated calibration data where available, and multi-judge rubric scoring. Across public-synthetic AML workloads and controlled serving benchmarks, workload-aware tuning improved throughput from 612-650 to 3,600 requests/hour, reduced P99 latency from 31-38 seconds to 6.4-8.7 seconds, and increased GPU utilization from 12% to 78%. These results show that regulated LLM performance is a workload-design, serving-optimization, and quality-gating problem, not only a model-selection problem.

cs.AI↗

arXiv2026

Pretraining Induces a Reusable Spectral Basis for Downstream Task Adaptation

Finetuning pretrained models occurs in a low-dimensional subspace of the full parameter space. Prior work has focused on characterizing this optimization subspace, but largely ignored the complementary question: why do certain directions remain unexplored during finetuning? Are these stable directions irrelevant to downstream tasks, or do they already encode task-relevant structure that requires no further adjustment? Answering this question is central to understanding how pretrained knowledge transfers. Through systematic spectral analysis across vision and language models, we show that the leading singular vectors of pretrained weight matrices remain highly stable under finetuning and are shared across unrelated downstream tasks, revealing that pretraining establishes a reusable spectral coordinate system. Models pretrained on larger datasets exhibit greater spectral stability under distribution shift or task change, directly linking pretraining scale to geometric transferability. Motivated by these findings, we propose a parameter-efficient method that freezes pretrained singular vectors and optimizes only leading spectral coefficients, achieving competitive performance on GLUE with 0.2% trainable parameters. Our results reveal that the stable directions encode transferable structure rather than irrelevant noise: successful pretraining discovers spectral bases that downstream tasks inherit and operate within.

cs.LG↗

arXiv2026

Implicit Compression Regularization: Concise Reasoning via Internal Shorter Distributions in RL Post-Training

Reinforcement learning with verifiable rewards improves LLM reasoning but often induces overthinking, where models generate unnecessarily long reasoning traces. Existing methods mainly rely on length penalties or early-exit strategies; however, the former may degrade accuracy and induce underthinking, whereas the latter assumes that substantial portions of reasoning traces can be safely truncated. To obtain a compression signal without these limitations, we revisit the training dynamics of existing compression methods. We observe that the length--accuracy correlation is initially negative but continually increases during compression, indicating that shorter responses are initially more likely to be correct but gradually lose this property as the policy moves toward underthinking. Based on this observation, we formalize overthinking: a negative correlation indicates an overthinking regime, while a positive one indicates underthinking. When overthinking, the shortest correct responses are shorter than the group-average response length in expectation, making them natural compression targets already present in on-policy rollouts. We therefore propose \emph{Implicit Compression Regularization} (ICR), an on-policy regularization method whose compression signal comes from a virtual shorter distribution induced by the shortest correct responses in rollout groups, guiding the policy toward concise yet correct trajectories. Training dynamics show that ICR maintains a better length--accuracy correlation during compression, indicating that short responses remain better aligned with correctness instead of drifting toward underthinking. Experiments on three reasoning backbones and multiple mathematical and knowledge-intensive benchmarks show that ICR consistently shortens responses while preserving or improving accuracy, achieving a stronger accuracy--length Pareto frontier.

cs.AI↗

arXiv2026

LANTERN: LLM-Augmented Neurosymbolic Transfer with Experience-Gated Reasoning Networks

Transfer learning in reinforcement learning (RL) seeks to accelerate learning in new tasks by leveraging knowledge from related sources. Existing neurosymbolic transfer methods, however, typically rely on manually specified task automata, assume a single source task, and use fixed knowledge-integration mechanisms that cannot adapt to varying source relevance. We propose LANTERN, a unified framework for multi-source neurosymbolic transfer that addresses these limitations through three components: (i) deterministic finite automata generated from natural language task descriptions using large language models, (ii) semantic embedding-based aggregation of multiple source policies weighted by cross-task similarity, and (iii) adaptive teacher-student gating based on temporal-difference error and semantic uncertainty. Across domains spanning resource management, navigation, and control, LANTERN achieves 40-60% improvements in sample efficiency over existing baselines while remaining robust to poorly aligned sources. These results demonstrate that multi-source, adaptively weighted neurosymbolic transfer can improve scalability and robustness in symbolic RL settings.

cs.AI↗

arXiv2026

MINT: Multi-Vector Search Index Tuning

Vector search plays a crucial role in many real-world applications. In addition to single-vector search, multi-vector search becomes important for multi-modal and multi-feature scenarios today. In a multi-vector database, each row is an item, each column represents a feature of items, and each cell is a high-dimensional vector. In multi-vector databases, the choice of indexes can have a significant impact on performance. Although index tuning for relational databases has been extensively studied, index tuning for multi-vector search remains unclear and challenging. In this paper, we define multi-vector search index tuning and propose a framework to solve it. Specifically, given a multi-vector search workload, we develop algorithms to find indexes that minimize latency and meet storage and recall constraints. Compared to the baseline, our latency achieves 2.1X to 8.3X speedup.

cs.DB↗

arXiv2026

Representation Fréchet Loss for Visual Generation

We show that Fréchet Distance (FD), long considered impractical as a training objective, can in fact be effectively optimized in the representation space. Our idea is simple: decouple the population size for FD estimation (e.g., 50k) from the batch size for gradient computation (e.g., 1024). We term this approach FD-loss. Optimizing FD-loss reveals several surprising findings. First, post-training a base generator with FD-loss in different representation spaces consistently improves visual quality. Under the Inception feature space, a one-step generator achieves0.72 FID on ImageNet 256x256. Second, the same FD-loss repurposes multi-step generators into strong one-step generators without teacher distillation, adversarial training or per-sample targets. Third, FID can misrank visual quality: modern representations can yield better samples despite worse Inception FID. This motivates FDr$^k$, a multi-representation metric. We hope this work will encourage further exploration of distributional distances in diverse representation spaces as both training objectives and evaluation metrics for generative models.

cs.CV↗

arXiv2026

Retrieval-Augmented Multimodal Model for Fake News Detection

In recent years, multimodal multidomain fake news detection has garnered increasing attention. Nevertheless, this direction presents two significant challenges: (1) Failure to Capture Cross-Instance Narrative Consistency: existing models usually evaluate each news in isolation, fail to capture cross-instance narrative consistency, and thus struggle to address the spread of cluster based fake news driven by social media; (2) Lack of Domain Specific Knowledge for Reasoning: conventional models, which rely solely on knowledge encoded in their parameters during training, struggle to generalize to new or data-scarce domains (e.g., emerging events or niche topics). To tackle these challenges, we introduce Retrieval-Augmented Multimodal Model for Fake News Detection (RAMM). First, RAMM employs a Multimodal Large Language Model (MLLM) as its backbone to capture cross-modal semantic information from news samples. Second, RAMM incorporates an Abstract Narrative Alignment Module. This component adaptively extracts abstract narrative consistency from diverse instances across distinct domains, aggregates relevant knowledge, and thereby enables the modeling of high-level narrative information. Finally, RAMM introduces a Semantic Representation Alignment Module, which aligns the model's decision-making paradigm with that of humans - specifically, it shifts the model's reasoning process from direct inference on multimodal features to an instance-based analogical reasoning process. Extensive experimental results on three public datasets validate the efficacy of our proposed approach. Our code is available at the following link: https://github.com/li-yiheng/RAMM

cs.CL↗

arXiv2026

EvolveSignal: A Large Language Model Powered Coding Agent for Discovering Traffic Signal Control Strategies

In traffic engineering, fixed-time traffic signal control remains widely used for its low cost, stability, and interpretability. However, its design relies on hand-crafted formulas (e.g., Webster) and manual re-timing by engineers to adapt to demand changes, which is labor-intensive and often yields suboptimal results under heterogeneous or congested conditions. This paper introduces EvolveSignal, an LLM-powered coding agent for automatically discovering interpretable heuristic strategies for fixed-time traffic signal control. Rather than deriving entirely new analytical formulations, the proposed framework focuses on exploring code-level variations of existing control logic and identifying effective combinations of heuristic modifications. We formulate the problem as program synthesis, where candidate strategies are represented as Python functions with fixed input-output structures and iteratively optimized through external evaluations (e.g., a traffic simulator) and evolutionary search. Experiments on a signalized intersection demonstrate that the discovered strategies outperform a classical baseline (Webster's method), reducing average delay by 20.1\% and average stops by 47.1\%. Beyond performance, ablation and incremental analyses reveal that EvolveSignal can identify meaningful modifications, such as adjusting cycle length bounds, incorporating right-turn demand, and rescaling green allocations, that provide useful insights for traffic engineers. This work highlights the potential of LLM-driven program synthesis for supporting interpretable and automated heuristic design in traffic signal control.

cs.LG↗

arXiv2026

WebGen-R1: Incentivizing Large Language Models to Generate Functional and Aesthetic Websites with Reinforcement Learning

While Large Language Models (LLMs) excel at function-level code generation, project-level tasks such as generating functional and visually aesthetic multi-page websites remain highly challenging. Existing works are often limited to single-page static websites, while agentic frameworks typically rely on multi-turn execution with proprietary models, leading to substantial token costs, high latency, and brittle integration. Training a small LLM end-to-end with reinforcement learning (RL) is a promising alternative, yet it faces a critical bottleneck in designing reliable and computationally feasible rewards for website generation. Unlike single-file coding tasks that can be verified by unit tests, website generation requires evaluating inherently subjective aesthetics, cross-page interactions, and functional correctness. To this end, we propose WebGen-R1, an end-to-end RL framework tailored for project-level website generation. We first introduce a scaffold-driven structured generation paradigm that constrains the large open-ended action space and preserves architectural integrity. We then design a novel cascaded multimodal reward that seamlessly couples structural guarantees with execution-grounded functional feedback and vision-based aesthetic supervision. Extensive experiments demonstrate that our WebGen-R1 substantially transforms a 7B base model from generating nearly nonfunctional websites into producing deployable, aesthetically aligned multi-page websites. Remarkably, our WebGen-R1 not only consistently outperforms heavily scaled open-source models (up to 72B), but also rivals the state-of-the-art DeepSeek-R1 (671B) in functional success, while substantially exceeding it in valid rendering and aesthetic alignment. These results position WebGen-R1 as a viable path for scaling small open models from function-level code generation to project-level web application generation.

cs.CL↗

arXiv2026

Targeted Exploration via Unified Entropy Control for Reinforcement Learning

Recent advances in reinforcement learning (RL) have improved the reasoning capabilities of large language models (LLMs) and vision-language models (VLMs). However, the widely used Group Relative Policy Optimization (GRPO) consistently suffers from entropy collapse, causing the policy to converge prematurely and lose diversity. Existing exploration methods introduce additional bias or variance during exploration, making it difficult to maintain optimization stability. We propose Unified Entropy Control for Reinforcement Learning (UEC-RL), a framework that provides targeted mechanisms for exploration and stabilization. UEC-RL activates more exploration on difficult prompts to search for potential and valuable reasoning trajectories. In parallel, a stabilizer prevents entropy from growing uncontrollably, thereby keeping training stable as the model consolidates reliable behaviors. Together, these components expand the search space when needed while maintaining robust optimization throughout training. Experiments on both LLM and VLM reasoning tasks show consistent gains over RL baselines on both Pass@1 and Pass@$k$. On Geometry3K, UEC-RL achieves a 37.9\% relative improvement over GRPO, indicating that it sustains effective exploration without compromising convergence and underscoring UEC-RL as a key for scaling RL-based reasoning in large models. Our code is available at https://github.com/597358816/UEC-RL.

cs.AI↗

arXiv2026

FedGUI: Benchmarking Federated GUI Agents across Heterogeneous Platforms, Devices, and Operating Systems

Training GUI agents with traditional centralized methods faces significant cost and scalability challenges. Federated learning (FL) offers a promising solution, yet its potential is hindered by the lack of benchmarks that capture real-world, cross-platform heterogeneity. To bridge this gap, we introduce FedGUI, the first comprehensive benchmark for developing and evaluating federated GUI agents across mobile, web, and desktop platforms. FedGUI provides a suite of six curated datasets to systematically study four crucial types of heterogeneity: cross-platform, cross-device, cross-OS, and cross-source. Extensive experiments reveal several key insights: First, we show that cross-platform collaboration improves performance, extending prior mobile-only federated learning to diverse GUI environments; Second, we demonstrate the presence of distinct heterogeneity dimensions and identify platform and OS as the most influential factors. FedGUI provides a vital foundation for the community to build more scalable and privacy-preserving GUI agents for real-world deployment. Our code and data are publicly available at https://github.com/wwh0411/FedGUI..

cs.MA↗

arXiv2026

Energy threshold in Smith-Purcell radiation

Smith Purcell radiation has emerged as a crucial platform for investigating light-matter interactions and developing compact, tunable light sources that span from microwaves to X-rays. In classical theory, it is believed that Cherenkov radiation exhibits an energy threshold for electrons, while Smith Purcell radiation is considered free of such a threshold. Although quantum theory suggests there is an emission cutoff in Smith-Purcell radiation, the behavior of this radiation near the threshold remains understudied. In this article, we address this gap by examining the behavior of Smith-Purcell radiation near the threshold from quantum perspectives. Specifically, we derive a quantum energy threshold based on energy-momentum conservation, providing a rigorous limit for the onset of Smith Purcell radiation. Furthermore, we find that around the threshold the incident electron emits a photon and subsequently reverses its direction of motion. Additionally, we establish a classical energy threshold below which the classical theory breakdown by applying the Duane Hunt limit to Smith Purcell radiation. Accordingly, quantum theory is required when the electron energy falls between the classical and quantum thresholds. Our findings enrich the understanding of Smith Purcell radiation and provide valuable insights for developing low energy driven and heralded quantum light sources.

physics.optics↗

arXiv2026

RoboCOIN: An Open-Sourced Bimanual Robotic Data Collection for Integrated Manipulation

Despite the critical role of bimanual manipulation in endowing robots with human-like dexterity, large-scale and diverse datasets remain scarce due to the significant hardware heterogeneity across bimanual robotic platforms. To bridge this gap, we introduce RoboCOIN, a large-scale multi-embodiment bimanual manipulation dataset comprising over 180,000 demonstrations collected from 15 distinct robotic platforms. Spanning 16 diverse environments-including residential, commercial, and industrial settings-the dataset features 421 bimanual tasks systematically categorized by 39 bimanual collaboration actions and 432 objects. A key innovation of our work is the hierarchical capability pyramid, which provides granular annotations ranging from trajectory-level concepts to segment-level subtasks and frame-level kinematics. Furthermore, we present CoRobot, an efficient data processing pipeline powered by the Robot Trajectory Markup Language (RTML), designed to facilitate quality assessment, automated annotation, and unified multi-embodiment and data management. Extensive experiments demonstrate the effectiveness of RoboCOIN in enhancing the performance of various bimanual manipulation models across a wide spectrum of robotic embodiments. The entire dataset and codebase are fully open-sourced, providing a valuable resource for advancing research in bimanual and multi-embodiment manipulation.

cs.RO↗

arXiv2026

Fast-SegSim: Real-Time Open-Vocabulary Segmentation for Robotics in Simulation

Open-vocabulary panoptic reconstruction is crucial for advanced robotics and simulation. However, existing 3D reconstruction methods, such as NeRF or Gaussian Splatting variants, often struggle to achieve the real-time inference frequency required by robotic control loops. Existing methods incur prohibitive latency when processing the high-dimensional features required for robust open-vocabulary segmentation. We propose Fast-SegSim, a novel, simple, and end-to-end framework built upon 2D Gaussian Splatting, designed to realize real-time, high-fidelity, and 3D-consistent open-vocabulary segmentation reconstruction. Our core contribution is a highly optimized rendering pipeline that specifically addresses the computational bottleneck of high-channel segmentation feature accumulation. We introduce two key optimizations: Precise Tile Intersection to reduce rasterization redundancy, and a novel Top-K Hard Selection strategy. This strategy leverages the geometric sparsity inherent in the 2D Gaussian representation to greatly simplify feature accumulation and alleviate bandwidth limitations, achieving render rates exceeding 40 FPS. Fast-SegSim provides critical value in robotic applications: it serves both as a high-frequency sensor input for simulation platforms like Gazebo, and its 3D-consistent outputs provide essential multi-view 'ground truth' labels for fine-tuning downstream perception tasks. We demonstrate this utility by using the generated labels to fine-tune the perception module in object goal navigation, successfully doubling the navigation success rate. Our superior rendering speed and practical utility underscore Fast-SegSim's potential to bridge the sim-to-real gap.

cs.RO↗

arXiv2026

Ψ-Map: Panoptic Surface Integrated Mapping Enables Real2Sim Transfer

Open-vocabulary panoptic reconstruction is essential for advanced robotics perception and simulation. However, existing methods based on 3D Gaussian Splatting (3DGS) often struggle to simultaneously achieve geometric accuracy, coherent panoptic understanding, and real-time inference frequency in large-scale scenes. In this paper, we propose a comprehensive framework that integrates geometric reinforcement, end-to-end panoptic learning, and efficient rendering. First, to ensure physical realism in large-scale environments, we leverage LiDAR data to construct plane-constrained multimodal Gaussian Mixture Models (GMMs) and employ 2D Gaussian surfels as the map representation, enabling high-precision surface alignment and continuous geometric supervision. Building upon this, to overcome the error accumulation and cumbersome cross-frame association inherent in traditional multi-stage panoptic segmentation pipelines, we design a query-guided end-to-end learning architecture. By utilizing a local cross-attention mechanism within the view frustum, the system lifts 2D mask features directly into 3D space, achieving globally consistent panoptic understanding. Finally, addressing the computational bottlenecks caused by high-dimensional semantic features, we introduce Precise Tile Intersection and a Top-K Hard Selection strategy to optimize the rendering pipeline. Experimental results demonstrate that our system achieves superior geometric and panoptic reconstruction quality in large-scale scenes while maintaining an inference rate exceeding 40 FPS, meeting the real-time requirements of robotic control loops.

cs.RO↗

arXiv2026

Reinforcement Learning-Guided NSGA-II Enhanced with Gray Relational Coefficient for Multi-Objective Optimization: Application to NASDAQ Portfolio Optimization

In modern financial markets, decision-makers increasingly rely on quantitative methods to navigate complex trade-offs among multiple, often conflicting objectives. This paper addresses constrained multi-objective optimization (MOO) with an application to portfolio optimization for minimizing risk and maximizing return. To address existing gaps, we propose a novel reinforcement learning (RL)-guided non-dominated sorting genetic algorithm II (NSGA-II) enhanced with gray relational coefficients (GRC), termed RL-NSGA-II-GRC, which combines an RL agent controller and GRC-based selection to improve convergence and diversity of Pareto fronts. The agent adapts evolutionary parameters online using metrics of hypervolume, feasibility, and diversity, while the GRC tournament operator ranks parents via a unified score considering dominance rank, crowding distance, and proximity to ideal reference. We evaluate the framework on the Kursawe and CONSTR benchmarks and a NASDAQ portfolio application. On the benchmarks, RL-NSGA-II-GRC achieves convergence improvements of about 5.8% and 4.4% over NSGA-II, while preserving well-distributed non-dominated solutions. In the portfolio application, it produces a smooth, densely populated efficient frontier supporting identification of the maximum Sharpe ratio portfolio (annualized Sharpe =1.92) and utility-optimal portfolios for different risk-aversion levels. The main contributions are three-fold: 1) we propose an RL-NSGA-II-GRC method integrating an RL agent into the evolutionary framework to adaptively control parameters via generational feedback; 2) we design a GRC-enhanced binary tournament operator providing a comprehensive indicator to guide the search toward the Pareto front; 3) we demonstrate, on benchmark MOO and a NASDAQ case study, that the method delivers improved convergence and well-populated frontiers supporting actionable insights.

cs.LG↗

arXiv2026

Joint Observation of SGR J1935+2154 with \textit{Insight}-HXMT and KM40m during the active episode of October 2022

SGR J1935+2154 is the unique magnetar so far from which fast radio bursts have been detected. In October 2022, it resumed its burst activity, and we implemented a dedicated target-of-opportunity (ToO) observation on it from Oct. 13th to Nov. 1st, 2022 (about 940 ks in total) with \textit{Insight}-HXMT, while the KM40m radio telescope observed this source for about 1400 hours since Oct. 15th. We searched the LE, ME, and HE data of \textit{Insight}-HXMT in the overlapping observation time windows with the KM40m radio telescope and revealed 60 magnetar X-ray bursts (MXBs), while KM40m only detected 1 radio burst. In particular, we find that there is an X-ray burst on October 21 (denoted as MXB 221021) temporally associated with this radio burst. Interestingly, this association event shows very different morphology from those X-ray and radio association events from this source reported before (e.g., MXB/FRB 200428). Moreover, we systematically analyzed the temporal and spectral properties of the sample of MXBs during this observation and found that % the (radio-associated) MXB 221021 shows some different properties from other MXBs without associated radio bursts. These findings shed new light on the physical mechanisms of X-ray bursts and radio burst emission in magnetars.

astro-ph.HE↗

arXiv2026

GECAM discovery of a peculiar magnetar X-ray burst (MXB 221120) from SGR J1935+2154 associated with a fast radio burst

Fast radio bursts (FRBs) are enigmatic cosmic transients of millisecond duration observed in the radio band. The identification of FRB-associated magnetar X-ray bursts (MXBs) from galactic magnetar SGR J1935+2154 suggests that at least a fraction of FRBs can be produced from magnetar activity. However, the sample size of FRB-associated MXBs is still very small. Here we report a bright and peculiar FRB-associated MXB from SGR J1935+2154 detected by GECAM on November 20, 2022, dubbed MXB 221120. We find that both temporal and spectral properties of MXB 221120 exhibit distinctive features. Its light curve could be generally described by a single FRED function with superposition of several narrow pulses. Interestingly, we identify a possible QPO feature with center frequency of ~18 Hz in this MXB. The time-integrated spectrum is best fitted by a blackbody model with temperature (kT ) of 18.6 keV, rendering it the first thermal spectrum FRB-associated MXB from SGR J1935+2154. Compared to other MXBs with single emission episode, MXB 221120 has longer duration and higher blackbody temperature, making it an outlier in the burst sample. These results indicate that MXB 221120 may be produced by a special mechanism with extreme physical conditions.

astro-ph.HE↗