Searcharxiv⌕ Search

arXiv subjects

Hao Wang

Publications and source records attributed to Hao Wang.

At least 145 records · Page 8Linked to original sources
arXiv2026

Design and Simulation Study of the Hadronic Calorimeter for the EicC Zero Degree Calorimeter

The Zero-Degree Calorimeter (ZDC) at the proposed Electron-Ion Collider in China (EicC) is essential for detecting forward-going neutral particles and supporting the core nucleon spin and 3D imaging physics programs. In this work, a highly compact Spaghetti Calorimeter (SPACAL) architecture is proposed and optimized as the baseline design for the ZDC hadronic section. To systematically and quantitatively evaluate its physics capabilities, a comprehensive end-to-end Geant4 simulation framework was developed, integrating energy deposition, optical photon transport, photo-detection, and modeled front-end signal digitization. The optimized detector demonstrates excellent performance for neutron detection, achieving an energy resolution of $33.42\%/\sqrt{E/\mathrm{GeV}} + 1.47\%$ and a timing resolution of approximately 500 ps, outperforming the targeted specification. Furthermore, full-system simulations incorporating the upstream electromagnetic calorimeter yield a sub-centimeter transverse position resolution. Moreover, topological shower shape analysis across the full detector system provides robust particle identification (PID), with photon-neutron separation accuracy reaching over $99\%$ for high-energy incident particles. These quantitative results confirm that the SPACAL design fully satisfies the stringent operational and physics requirements of the EicC forward kinematic region.

physics.ins-det↗
arXiv2026

Strong-coupling quantum optics in free space with holes in a Fermi sea

Coherent and efficient light-matter interfaces between an atom and a single mode of the electromagnetic field are essential for quantum technologies. Traditionally, these systems employ optical cavities or waveguides to isolate a specific mode of light, but a more recent approach is to engineer antennas from ordered configurations of trapped ultracold atoms that exhibit controllable and directed scattering. In this work, we propose a method to engineer an antenna from the center-of-mass wavefunction of a single atom, which can produce directed emission and thereby exhibit strong coupling to a target mode of light in free space. We predict that the resulting single-atom cooperativity can be comparable to the current state of the art in optical cavity and waveguide QED experiments. Building on this approach, we show that a wavepacket antenna becomes a single-atom mirror in the linear response regime. We study the modified dipole-dipole interactions and band structure of chains of such emitters, which can exhibit sub- and superradiance at spacings much larger than the wavelength of the light. Extending our approach to multi-level atoms, we propose a method to achieve arbitrary spatial scattering from a single atom, including uni-directional spontaneous emission. Finally, we elucidate how our approach can enhance cooperative scattering in conventional arrays of tightly trapped atoms.

quant-ph↗
arXiv2026

Volumetric Optical Scattering Neural Networks

Optical neural networks rely on planar photonic circuits or spatially separated diffractive planes, limiting volumetric integration and requiring precise alignment. Here we propose a volumetric optical scattering neural network (OSNN) comprising densely packed refractive-index neurons with weak contrast $Δn = 0.006$. Direction-dependent first-order Born scattering provides local coupling, while ordered propagation accumulates these weak updates into global task-specific transformations. Fabrication-aware training and two-photon nanolithography produced a sub-$4 \times 10^{-4}$ mm$^3$ device containing 376,320 neurons at $1.0 \times 10^9$/mm$^3$ in a continuously written, intrinsically registered volume. Experimentally, a classifier performed class-selective energy routing with 95.8% blind-test accuracy on MNIST. The calculated passive optical transit time is below 1 ps. A separately optimized imager provided fourfold full-field spatial-sampling compression, resolved 1-$μ$m features and enabled cross-domain transfer. An OSNN co-designed for multiple tasks encoded single-cell fluorescence-derived images into measured output maps, which were pooled into a 16-dimensional representation supporting organelle classification, morphology reconstruction, doublet detection and morphology regression, with all reported metrics exceeding 0.92. Volumetric weak scattering thus establishes an ultracompact, intrinsically registered optical computing architecture, opening a new route to scalable machine vision and biomedical analysis.

physics.optics↗
arXiv2026

NavPatch: Evidence-Guided Object-Level Costmap Correction with Vision-Language Models

Mobile robots typically rely on geometric maps for obstacle avoidance and path planning, but the resulting obstacle representation does not always match how an object should affect navigation. A low lying cable may be missed, a flexible curtain may create spurious blockage, and a traffic cone may require an exclusion region larger than its observed footprint. We present NavPatch, an object level correction layer that assigns ADD, REMOVE, or EXTEND to navigation relevant object categories through periodic scene understanding with a vision-language model. Open vocabulary grounding localizes object instances, and LiDAR and RGB-D observations provide 3D support. Observation quality filtering and cross frame maintenance determine when each correction patch is committed, replaced, or revoked. In 50 real robot trials across five layouts, NavPatch achieves an overall success rate of 86.0%. An ablation study of four configurations with 200 runs in total shows that NavPatch improves the success rate from 70.0% to 86.0% and reduces the false commit rate from 68.4% to 40.7% compared with updates based only on the current observation.

cs.RO↗
arXiv2026

YOLO12-MambaScan: An Efficient Object Detector with High-Frequency Enhancement and State-Space Modeling

The rapid development of unmanned aerial vehicle (UAV) technology has made aerial-image object detection increasingly important for natural-resource monitoring, traffic management, and disaster response. Detecting small objects in aerial images remains difficult because objects occupy very few pixels, high-frequency cues are easily lost, and global context is hard to model in cluttered scenes. Existing detectors often retain insufficient edge, corner, and texture information. We propose \ours, an aerial-image detector built on the YOLO12 architecture. The model combines a triple-path high-frequency enhancement convolution module (TriPathHFConv), receptive-field coordinate-attention convolution (RFCAConv), and a Mamba-based global-context module. On VisDrone, at an input resolution of 960*960, ours achieves 60.0% mAP@50 and 38.6%mAP@50:95, demonstrating a favorable accuracy--efficiency trade-off for small-object detection. The benchmark and dataset protocol follow the VisDrone challenge setup.

cs.CV↗
arXiv2026

UniCAR-RL: Seeing Better before Thinking Deeper in Visual Mathematics

Multimodal Large Language Models (MLLMs) often struggle with complex mathematical visual reasoning primarily due to a lack of fine-grained perception, causing initial visual hallucinations to directly trigger cascading reasoning failures. In traditional end-to-end reinforcement learning (RL), sparse rewards fail to decouple perceptual hallucinations from logical missteps, hindering targeted perception optimization. Alternatively, fine-tuning with perception-enhanced CoT data incurs high costs and hallucinations. In this paper, we address these challenges by proposing UniCAR-RL, an annotation-free RL framework. By explicitly decoupling the optimization of perception and reasoning during the training process, it achieves isolation and optimization of both capabilities. Specifically, UniCAR-RL consists of three synergistic branches: 1) a Caption-RL branch that optimizes perception capabilities through verifier-guided reasoning validation; 2) a Reasoning-RL branch that performs logical reasoning based on a gold image description to halt cascading errors; 3) a QA-RL branch that retains native end-to-end alignment to ensure robust question-answering performance. Experiments show that UniCAR-RL substantially improves MLLMs' mathematical and visual reasoning using only raw short-answer data. Furthermore, it demonstrates strong generalization across diverse architectures and scales.

cs.AI↗
arXiv2026

A Multi-Resolution Multi-Domain Pre-Training Framework for Universal Traffic Forecasting

Spatio-temporal traffic data are central to intelligent transportation systems, yet their heterogeneity poses significant challenges for large-scale modeling. Existing pre-trained models often rely on a homogeneous modeling paradigm to handle highly heterogeneous traffic data. This fundamental mismatch not only limits model generalization but also leads to computationally expensive and parameter-inefficient designs. To this end, we propose FlexST, a novel pre-training framework that introduces modularity and adaptivity for traffic modeling. Specifically, we first propose a multi-resolution spatio-temporal diffusion module that captures both short-term fluctuations and long-range trends, effectively reconciling inputs with divergent temporal and spatial resolutions. After that, we construct a domain-adaptive mixture-of-experts that dynamically routes data to specialized sub-networks, enabling selective knowledge transfer while preventing negative interference across diverse domains. Moreover, we devise a unified periodic encoding strategy that injects resolution- and domain-aware inductive biases to harmonize periodic inconsistencies across datasets. Extensive experiments on 23 real-world traffic datasets demonstrate that FlexST significantly outperforms state-of-the-art baselines in zero- and few-shot settings, showcasing superior generalization, adaptability and efficiency. This work offers a new direction for building general-purpose pre-trained models capable of handling the complexity and variability of urban traffic systems.

cs.LG↗
arXiv2026

Non-uniform B-spline optimization method for generating swept surfaces

Swept surface construction is widely used in computer-aided design. We propose a novel optimization method using non-uniform B-splines to improve the approximate accuracy of swept surfaces. First, discrete points on the swept shape are computed, and geometric properties such as surface area, discrete curvature, first-order derivatives, and their rotation angles are used to derive a distribution function representing surface irregularity, with weights adjusted from samples. Then, feature points are selected based on the distribution function to determine control points for the approximate non-uniform B-spline surface via inverse calculation, producing an optimized approximation. Finally, the number of feature points is adjusted based on the estimated approximation error. Experiments on 969 randomly generated sweep samples and 1 pipe example show that the proposed algorithm achieves similar accuracy with fewer control points, reducing them by about 15.85% at a specified accuracy of 0.01. Moreover, with ample sampling points, it reduces the average error by approximately 51.35% when the feature point multiple is 10 times the path control points, outperforming comparable methods.

cs.GR↗
arXiv2026

Autonomous multifunctional image processing via programmable multimode lasing

Optical image processing offers a promising pathway to overcome the latency and energy limitations of conventional electronic image processors. However, existing approaches based on passive photonic devices are often constrained by signal attenuation, lack of nonlinearity, fixed functionality, and heavy training overhead. Here, we introduce a programmable image processor based on a highly multimode degenerate cavity laser (DCL), shifting the computational framework from passive extracavity transformation to active intracavity evolution. By manipulating the intracavity loss distribution and exploiting the nonlinear lasing dynamics, we map computational tasks to the spontaneous mode selection within the DCL, realizing image processing directly at the source. We experimentally demonstrate that by simply altering the encoding scheme of input images, the platform can be flexibly reconfigured for multifunctional tasks, including high-fidelity edge detection and robust image denoising. The computation proceeds autonomously without dataset training, featuring an intrinsically low latency ($\sim$90 $μ\mathrm{s}$) with $\mathcal{O}(1)$ complexity. Furthermore, the high intensity sustained within the resonator enables intracavity nonlinear upconversion of the processed image. This work extends the frontiers of laser applications, providing a compelling candidate for next-generation optical processors.

physics.optics↗
arXiv2026

Quantifying the Value of Privileged Information Using a PAC-Bayesian Approach

In practice, various learning scenarios provide access to auxiliary features exclusively during training. Incorporating such data to enhance model performance gave rise to a paradigm known as Learning Using Privileged Information (LUPI). While this extra information is intended to improve the resulting model, establishing a generalized, cohesive understanding of how privileged information (PI) transfers useful knowledge remains a challenge. Vapnik's original theory and subsequent works offer performance guarantees in certain cases, but these results are inherently per-algorithm and rely on setting-specific proof approaches. Consequently, a more general framework explaining how and when PI transfers useful knowledge is still missing. To bridge this gap, we introduce an algorithm-agnostic, information-theoretic approach based on the PAC-Bayes framework. Rather than asking whether a particular algorithm exploits PI, we ask how much value it could offer: comparing the tightest achievable risk bound with and without PI yields its potential - an upper limit on the extractable gain. We introduce a metric that quantifies this potential directly from empirical training risk, bypassing the need for test-time data access, and validate our findings in both supervised and unsupervised settings. The results demonstrate a robust correspondence between our training-time metric and true test-time performance gains. Ultimately, this work takes a necessary step toward an information-theoretic understanding of LUPI, and quantifying the potential of privileged features before committing to a model.

cs.LG↗
arXiv2026

BenchShield: Formal Model-Backed Instrumentation for Reward Integrity in LLM-Agent Evaluation Infrastructure

LM-agent benchmarks increasingly function as interactive evaluation infrastructure. Agents observe state, call tools, modify workspaces, submit artifacts, and receive rewards from outcome procedures. This interactivity makes evaluations vulnerable to reward hacking: an agent improves its measured score by exploiting the reward-relevant trajectory instead of solving the intended task. Existing defenses rely largely on task-specific patches, prompt instructions, or post-hoc detectors. They do not provide reusable evidence that a concrete run remained within its intended evaluation boundary. This paper presents BenchShield, a model-backed instrumentation layer for reward integrity in LLM-agent evaluation. BenchShield grounds detection in a finite lifecycle model of an evaluation's reward-relevant events. Within the benchmark infrastructure, two complementary analyses operate over this model. A static, phase-aware taint analysis exposes reward-hacking paths before a run. Its runtime counterpart uses infrastructure-side evidence to attribute concrete agent use and emit evidence-backed claims. We construct BenchShield Trajectories, a human-labeled corpus of 456 adjudicated trajectories from more than 31,000 public agent runs across three benchmarks. Compared with an agentic hackability scanner baseline on the same tasks and model, BenchShield improves full-chain recall from 23-94% to 77-100%, same-vector coverage from 16-56% to 43-78%, and reduces per-task cost by up to 65%. Its runtime analysis achieves 96% accuracy in detecting reward hacking from infrastructure-side evidence.

cs.CR↗
arXiv2026

X2-NativeCursor: Native-Token Text Progress Tracking for Incremental-Text Streaming Codec TTS

Incremental-text streaming text-to-speech (TTS) needs online text progress tracking for synchronized highlighting, interruption handling, and dialogue-history updates. Input text arrives before it is spoken, so text arrival alone cannot indicate speech progress. Existing waveform-based alignment requires complete audio or adds acoustic processing during streaming. We propose X2-NativeCursor, a lightweight observer that tracks progress from native speech tokens before waveform decoding without changing the TTS generator. Its normalization plan links spoken labels to their original-text spans. Text and native-token encoders feed a local matcher that estimates the current label position. A separate output rule converts revisable position estimates into a cursor that never moves backward. Mean absolute error against an automatic reference is 0.151 Chinese characters with 80-ms lookahead, versus 1.253 characters with 320-ms lookahead for an online waveform baseline. Alignment real-time factor also decreases from 0.3598 to 0.0180 relative to this baseline. Lower tracking error is retained under a second automatic alignment reference. We evaluate X2-NativeCursor on Qwen3-TTS and validate its adaptation to CosyVoice2 by training a separate observer for each backbone. Code is publicly available at https://github.com/X-Square-Robot/X2Streaming-TTS.

cs.CL↗
arXiv2026

Search for neutrinoless quadruple beta decay of $^{136}$Xe in PandaX-4T detector

The observation of neutrinoless quadruple beta decay (0$ν$4$β$) in the absence of neutrinoless double beta decay (0$ν$2$β$) has been argued to provide a strong indication that neutrinos are Dirac particles. We report a search for 0$ν$4$β$ decay of $^{136}\text{Xe}$ using a total $^{136}\text{Xe}$ exposure of 148.4 kg$\cdot$yr, collected during the commissioning and the first science runs of the PandaX-4T experiment. No significant excess of events over the background is observed. A lower limit on the 0$ν$4$β$ decay half-life of $^{136}\text{Xe}$ is set at 6.01 x $10^{24}$ yr at the 90% confidence level. This result establishes the most stringent constraint on this process in xenon, demonstrating the unique capability of the PandaX-4T detector in probing lepton number violation and shedding light on the fundamental nature of neutrinos.

nucl-ex↗
arXiv2026

Compressing Sequences in the Latent Embedding Space: $K$-Token Merging for Large Language Models

Large Language Models (LLMs) incur significant computational and memory costs when processing long prompts, as full self-attention scales quadratically with input length. Token compression aims to address this challenge by reducing the number of tokens representing inputs. However, existing prompt-compression approaches primarily operate in token space and overlook inefficiencies in the latent embedding space. In this paper, we propose K-Token Merging, a latent-space compression framework that merges each contiguous block of K token embeddings into a single embedding via a lightweight encoder. The compressed sequence is processed by a LoRA-adapted LLM, while generation remains in the original vocabulary. Experiments on structural reasoning (Textualized Tree), sentiment classification (Amazon Reviews), and code editing (CommitPackFT) show that K-Token Merging lies on the Pareto frontier of performance vs. compression, achieving up to 75% input length reduction with minimal performance degradation. Code is available at https://github.com/shsjxzh/K-Token-Merging.

cs.CL↗
arXiv2026

X2-Turn: Frame-Synchronous Dual-Head Modeling for Joint Streaming ASR and Turn State Prediction

Accurate and responsive turn-taking is essential for spoken dialogue systems, which must distinguish in real time between user interruptions, backchannels that should be ignored, and the completion of an utterance. Prior modular approaches typically optimize turn state prediction at the utterance or fixed-chunk level, creating a mismatch with the continuous turn state estimate, and often depend on an auxiliary ASR model, which limits responsiveness and increases overall system complexity. Therefore, we present X2-Turn, a frame-synchronous turn state prediction method via delayed-stream modeling. Specifically, building on the pretrained Voxtral Realtime model, we introduce a frame-synchronous turn state head that operates in parallel with the ASR head on shared streaming representations, jointly predicting ASR tokens and fine-grained turn states at the frame level. Experiments on bilingual EasyTurn and Full-Duplex-Bench demonstrate that the proposed method achieves an effective trade-off between turn state accuracy and decision latency.

cs.CL↗
arXiv2026

TruthInsightBench: An Evidence-Grounded Benchmark for Automated Evaluation of Open-Ended Scientific Discovery Agents

Autonomous coding agents are increasingly proposed as AI-scientist systems that conduct analyses and write research reports, but executing a prescribed analysis is not the same as making a discovery. Existing benchmarks are configured for reproduction: tasks, data, and rubrics are built around a hidden target study, and recovery of its result is rewarded. We present TruthInsightBench, a benchmark configured for discovery. Its 40 blind tasks, drawn from 40 peer-reviewed studies across 10 scientific domains, expose only a neutral scientific objective and frozen data; source conclusions, expected values, and analysis paths are withheld, leaving the agent to determine what claim the data support. A fixed LLM-based judge scores the evidentiary maturity of an agent's own claims along six dimensions, operationalized as 29 artifact-grounded items, with automated, deterministic aggregation and no per-instance human grading, so evaluation can be repeated automatically as agents evolve. On one frozen base model, four coding agents form a narrow plateau (58.4-60.3 of 100) with no statistically reliable pairwise separation: they execute and document analyses competently, with comparatively strong evidence auditability and novelty, but largely lack the discriminating acts that establish a trustworthy claim (controls, robustness, falsifiability, and cross-dataset generalization). The bottleneck is scientific judgment rather than coding, and genuine discovery remains out of reach. TruthInsightBench makes this gap a measurable target; data and scoring code are at https://github.com/TruthInsight-stack/TruthInsightBench.

cs.AI↗
arXiv2026

NeuCME: Toward Dynamic Multimodal Continual Learning via Neural Combinatorics of Multiple Experts

Multimodal continual learning has recently shown great potential for developing agents with human-like intelligence by continuously learning new tasks across multiple modalities. However, existing methods typically assume that the set of modalities per task is predefined and fixed. In this paper, we investigate a more realistic learning setting, referred to as dynamic multimodal continual learning, in which the set of modalities may vary across tasks rather than remaining fixed. This setting involves two primary challenges: (i) spatio-temporal catastrophic forgetting and (ii) adaptive multimodal fusion. To address these challenges, we propose NeuCME (as shorthand for \textbf{Neu}ral \textbf{C}ombinatorics of \textbf{M}ultiple \textbf{E}xperts), a novel framework designed to effectively learn and integrate knowledge across tasks with varying modalities. The proposed NeuCME model comprises three key components, namely modality-combinational rehearsal, multi-gated mixture-of-experts, and task relevance-guided distillation. Furthermore, we formulate an evaluation metric to quantify the dynamism of task sequences and then set up a comprehensive benchmark with different degrees of dynamism. Extensive experiments using four real-world datasets demonstrate that the proposed NeuCME outperforms state-of-the-art methods markedly.

cs.LG↗
arXiv2026

A dual theory of Zhu's theory

We establish a dual theory of Zhu's $A(\mathcal {V})$-theory, i.e., for a graded vertex operator coalgebra $V$, the coassociative coalgebra $C(V)$ are investigated, any admissible $V$-comodule gives a $C(V)$-comodule, and vice versa. We also prove that $V$ is corational, which means every admissible $V$-comodule is completely reducible, if and only if its dual vertex operator algebra $V'$ is rational.

math.QA↗