Searcharxiv⌕ Search

arXiv subjects

Zheng Gong

Publications and source records attributed to Zheng Gong.

At least 19 recordsLinked to original sources
arXiv2026

Magnetic island structures in relativistic laser-driven plasma channels

We develop a theoretical model for self-generated magnetic islands in relativistic laser-driven channels in near-critical-density plasmas. The islands arise from the nonlinear superposition of the quasi-static magnetic fields generated by the longitudinal channel current $j_x$ and the laser-front driven transverse current $j_y$. By deriving the critical conditions among laser depletion, transversely symmetric channel formation, and magnetic-island formation, we identify the laser-plasma parameter window in which the magnetic island structures can exist. Within this window, the balance between the laser ponderomotive force and the charge-separation force, expressed through an effec tive electron density $n_\mathrm{eff}$, determines the transverse island width $H$, whereas the mismatch between the laser group and phase velocities determines the longitudinal period $L$. Large-scale particle-in-cell simulations over a broad range of laser intensities and plasma densities validate the resulting scaling laws. The model turns the island geometry from a qualitative feature of the channel field into a predictable quantity, providing a basis for tailoring electron transport, particle acceleration, high-energy radiation, and novel fusion ignition schemes in relativistic laser-plasma interactions.

physics.plasm-ph↗
arXiv2026

mRNA Design and Optimization with Deep Knowledge-Infused Approach

The mRNA optimization is essential for mRNA vaccines, therapies, and industrial protein production. Based on current explorations, an ideal optimization approach should simultaneously (i) prevent unintended amino-acid changes, (ii) optimize multiple, biologically relevant objectives, and (iii) retain computational efficiency. However, existing methods are forced to trade off between these perspectives, forming an "impossible triangle." We present RNop, a knowledge-infused Transformer that integrates mechanism-aligned losses to address this problem. By encoding biological prior knowledge in losses, RNop makes knowledge infusion explicit and controllable across optimization focus. Trained on over 6 million sequences, in silico analyses show RNop resolves the "impossible triangle" of mRNA optimization with absolute sequence fidelity, significantly improved biological metrics, and high throughput. In in vitro validation, it can deliver up to 2.28-fold expression gain. Ablation studies reveal how each prior contributes to targeted improvements, yielding mechanism-level interpretability. RNop represents a shift in mRNA optimization methodology: by infusing explicit and interpretable knowledge, the "black-box" mRNA design can be transformed into a predictable, explainable engineering problem. RNop is designed as an extensible platform: additional biological priors can be incorporated as modular, mechanism-aligned loss functions, enabling future development and adaptation to related sequence design problems.

q-bio.QM↗
arXiv2026

Wavefront-Guided Electron Injection for Direct Laser Acceleration in Relativistic Laser-Driven Plasma Channel

We investigate electron injection into direct laser acceleration (DLA) in relativistic laser-driven plasma channels using particle-in-cell simulations. We identify and characterize a wavefront-guided injection mechanism, in which electrons are continuously fed into the plasma channel through the density pile-up layer at the laser-pulse front. Phase-space analysis reveals a localized injectable region within the pile-up layer, indicating that only a selected subset of electrons satisfies the conditions required for the subsequent direct laser acceleration. This mechanism provides a physical interpretation for the high-charge capability of DLA by explaining how electrons are continuously supplied to the accelerating channel. Beyond this continuous supply process, the injection dynamics are further modulated by the periodic variation of the carrier phase at the laser-pulse front. The spatial locations of injected electrons are found to be closely associated with magnetic-island structures formed under laser-phase modulation, suggesting that the laser wavefront not only supplies electrons but also organizes their entry into the accelerating channel. These findings advance the physical understanding of energetic-electron generation in relativistic laser-driven subcritical-density plasma channels and are relevant to the development of compact DLA-based particle and radiation sources.

physics.plasm-ph↗
arXiv2026

Brewster-anomaly delocalization for free-electron radiation

Localization effects are central to disordered electronics and photonics. In electronics, Anderson localization governs electron confinement in randomly perturbed lattices. Similarly, its photonic counterpart inhibits light transport via disorder -- but with a unique exception: Brewster-anomaly delocalization, where the Brewster effect prevents multiple-scattering interference and counteracts the localization. Despite extensive research in electronics and photonics separately, the intricate role of localization effects in free-electron--light interactions -- vital for lasers, accelerators, microscopy and spectroscopy, and quantum information -- remains largely unexplored. At the same time, localization effects are widely regarded as a key factor limiting the efficient coupling between free electrons and light in random media. Here we overcome this key limitation via the unconventional interplay between Brewster-anomaly delocalization and free-electron radiation. In this way, free-electron radiation can be localization-free, intense and directional even in strongly disordered, unengineered multilayers. Essentially, this delocalization-mediated free-electron radiation is remarkably invariant not only to the random-medium configuration, but also to the light frequency and the electron velocity. Our findings unlock new opportunities for particle detectors and achromatic light sources operating in easy-to-fabricate complex media at previously inaccessible frequencies.

physics.optics↗
arXiv2026

Distance to Class Prototypes: Active Learning for Object Detection

Deploying a deep object detector in a new setting is limited less by architecture than by the cost of annotating data from that setting. Active learning lowers the cost by choosing which images to label, and the choice is only as good as the signal used to score an unlabeled image. That signal is usually the class posterior, which is cheap but poorly calibrated, or the disagreement across several models or several stochastic passes, which is better but multiplies inference over a pool far larger than the labeled set. We propose a signal richer than the posterior yet still read from one forward pass of one network. A supervised contrastive term added to the training objective shapes a per-object embedding space in which distance encodes class membership, and an unlabeled detection is scored by how far it lies from the region occupied by its predicted category, weighted by its confidence. The criterion needs no ensemble, no auxiliary predictor and no repeated inference, and its entire cost is 2.89M parameters, an increase of 8.3% over a bare detector. On PASCAL VOC and MS-COCO it beats the posterior of the same detector in every round in which a selection is made, by up to 1.08% mAP50 against run to run deviations of 0.02% to 0.18%, and it stays competitive with ensemble and Monte Carlo dropout criteria costing three to fifty forward passes per unlabeled image. Experiments use the single-stage detector under which the compared criteria report their results, so that the selection decision is isolated from the strength of the detector.

cs.CV↗
arXiv2026

Generative Verification: Rethinking the Uncertainty Signal for Active Learning of Object Detection

Nearly every acquisition function for active object detection shares one arrangement, in that the model being improved is also the model being interrogated. We depart from it. In generative verification an independent generative model re-derives the label of a detection from the pixels inside its predicted box, and the disagreement between the two becomes the acquisition signal. Two properties follow from the arrangement itself rather than from any tuning. A displaced box, a box on background and a correct box carrying the wrong label all yield a crop that fails verification, so the failure modes arrive already combined in one scalar and the hand-weighted classification and localization terms of existing criteria are no longer needed. And because the verifier never observes the detector confidence, confidently wrong detections score highest, although a self-derived signal reads them as uninteresting and they are the costliest to leave unlabeled. We build the verifier as a conditional diffusion model whose diffusion target is a label representation rather than an image. Its reverse process is stochastic, so repeated generations return a distribution whose concentration reports how firmly the evidence determines the label, where a classifier returns a single point estimate. On PASCAL VOC and MS-COCO the signal outperforms output-uncertainty, feature-geometry, perturbation and ensemble criteria, gaining about one mAP50 point per round on MS-COCO, with its largest margins in the early rounds where confident detector errors are most common.

cs.CV↗
arXiv2026

EyeMVP: OCT-Informed Fundus Representation Learning via Paired CFP--OCT Pretraining

Color fundus photography (CFP) is the mainstay of large-scale retinal screening, but its diagnostic capacity is limited by the lack of depth-resolved structure, which optical coherence tomography (OCT) provides yet is less accessible at population scale. We present EyeMVP, a cross-modal retinal foundation model that uses paired CFP--OCT pretraining to learn OCT-informed CFP representations while requiring only CFP at inference. Pretrained on 674,893 same-eye same-day CFP--OCT triples from 112,642 patients across eight hospitals, EyeMVP uses cross-modal masked reconstruction to enrich CFP features with OCT-associated supervision, and combines source-constrained cross-attention with CFP-derived structural masks to accommodate the non-aligned geometry of en-face CFP and cross-sectional OCT. Across 15 dataset-level settings spanning classification and segmentation, under both full-data and few-shot regimes, EyeMVP performs on par with or better than representative retinal foundation models, with consistent gains on macular and optic-nerve tasks; it attains AUROCs of 0.923 for macular edema and 0.867 for myopic macular schisis, two conditions poorly resolved in CFP. In an exploratory reader study, EyeMVP surpasses junior and intermediate ophthalmologists but not seniors on macular edema, while exceeding all groups on myopic macular schisis. These results indicate that cross-modal reconstruction can enrich CFP representations with OCT-associated supervision, offering a practical route to stronger CFP-based screening.

cs.CV↗
arXiv2026

Electron penetration heating in turbulent magnetic loops driven by nonrelativistic laser-plasma interaction

Using particle-in-cell simulations to study nonrelativistic laser pulse propagation in a under-critical plasma, we identify a novel mechanism that occurs during the growth of turbulent magnetic loops: electron penetration heating. The loops have an electromagnetic left-hand chirality distinct from that of well-known quasistatic magnetic islands. The fast electrons penetrate through the loops and thus are accelerated to unexpected relativistic energies due to the symmetry breaking induced by the coupling between the loop field and the non-relativistic electromagnetic wave. The identified features of penetration heating and magnetic loops might provide an alternative perspective for understanding superponderomotive electron heating in under-critical plasmas irradiated by nonrelativistic laser pulses. This is a potential explanation for anomalous hot electron generation in scenarios of laser-driven inertial confinement fusion.

physics.plasm-ph↗
arXiv2026

Directory-Aware Query and Maintenance in Vector Databases

Vector databases typically manage metadata as flat scalar attributes, which limits their ability to express hierarchical directory semantics commonly used to organize code repositories, enterprise documents, and agent memories. As a result, directory-scoped retrieval and structural updates are often implemented as application-layer workarounds, making recursive scope resolution expensive and directory maintenance difficult to keep consistent. This paper studies native directory semantics as a first-class capability for vector databases. We formalize two core operators: Directory-Semantic Query (DSQ) for hierarchically scoped retrieval, and Directory-Semantic Maintenance (DSM) for structural updates. We then evaluate three implementation strategies: query-time path expansion (PE-Online), ingestion-time path expansion (PE-Offline), and a Trie-based Hierarchical Index (TrieHI). Our analysis exposes the fundamental limitations of expansion-based designs: flattening the hierarchy incurs high recursive-query latency in PE-Online and unscalable write amplification during structural changes in both expansion strategies. In contrast, TrieHI keeps the directory topology as a native prefix tree, enabling efficient recursive retrieval through tree traversal and reducing maintenance cost through topological node manipulation. We benchmark these design points within ByteDance's Viking vector search engine and release two large-scale datasets, WIKI-Dir and ARXIV-Dir, to support future research on directory-semantic vector search. Finally, TrieHI has been integrated into OpenViking, an open-source context database for AI agents, where it supports filesystem-style context organization and directory-recursive retrieval.

cs.DB↗
arXiv2026

Arbor: Tree Search as a Cognition Layer for Autonomous Agents

Arbor is a multi-agent framework that introduces structured tree search as a cognition layer for autonomous agents operating in large, stateful action spaces. Prior autonomous optimization systems operate on isolated targets with stateless evaluation. Arbor instead maintains an explicit search tree of scored hypotheses that serves as the shared working memory across agents, evolving with every measurement, treating failures as diagnostic signal that reshapes subsequent exploration, and expanding as prior successes shift the bottleneck distribution. We validate Arbor on full-stack LLM inference optimization, a domain where achieving peak performance has historically required coordinated effort from engineering teams across the application, framework, compiler, kernel, and hardware stack. Arbor pairs an Orchestrator agent, which drives optimization by delegating to Domain Specialists across the inference stack, with a Critic agent that safeguards stability through root-cause analysis, introspection, and measurement validation -- a checks-and-balances architecture where neither agent can unilaterally drive the system. Agent capabilities are decomposed into hard skills (domain expertise) and soft skills (coordination protocols that determine how contributions compose), enabling fully autonomous multi-day campaigns. Arbor achieves up to 193% inference throughput-latency Pareto improvement over vendor-optimized baselines, while a single agent without the harness plateaus at +33% throughput improvement and crashes irrecoverably within hours. Arbor generalizes to multiple generations of hardware platform, and run-to-run variance is within 2 percentage points demonstrating that the method is hardware-agnostic and reproducible.

cs.AI↗
arXiv2026

Radiation-induced electron spin polarization in ultrarelativistic kinetic turbulence

Electron spin polarization in radiative plasmas with ultrarelativistic kinetic turbulence under highly magnetized conditions is investigated using particle-in-cell simulations. We observe that a significant spin polarization can be sustained when the leptons undergo energetic photon emission accompanied by spin flips during the nonequilibrium turbulent evolution. By analyzing the time evolution of spatially dependent spin polarization, we identify an electromagnetic (EM) regime of kinetic turbulence, distinct from the well-known density-dominated regime characterized by vortex currents and magnetic islands. While in the latter regime the spin polarization exists only transiently, in the EM regime significant anisotropic net polarization emerges and persists in non-dissipative scenarios. The correlation between spin signals and turbulence features is leveraged to introduce the characteristic parameter delimiting the EM regime via the ratio of electric and magnetic energy densities and to gain insight into complex plasma turbulence. This study demonstrates the versatility of a spin-resolved study of the plasma turbulence in extreme environments, such as black holes and magnetar magnetospheres.

physics.plasm-ph↗
arXiv2026

Enabling Real-Time Colonoscopic Polyp Segmentation on Commodity CPUs via Ultra-Lightweight Architecture

Real-time polyp segmentation is essential for early colorectal cancer detection, yet clinical deployment remains blocked by GPU dependency. We introduce the UltraSeg family, a set of CPU-native segmentation models operating below 0.3M parameters. UltraSeg-108K (0.108M) establishes the extreme-compression frontier, while UltraSeg-130K (0.130M) integrates cross-layer lightweight fusion for enhanced multi-center generalization. The architecture replaces parameter-heavy components with grouped multi-rate dilated convolutions and attention-gated cross-layer fusion, achieving real-time throughput on a single CPU core (exceeding 50 FPS at 256*256 and 30 FPS at 352*352) without sacrificing clinical-grade accuracy. Evaluated on seven public datasets, UltraSeg-130K attains Dice scores exceeding 0.8 at both resolutions, substantially outperforming all existing sub-0.3M competitors. Notably, it approaches or exceeds UNet-Medium (7.76M parameters) on zero-shot external validations while using only 1.7% of its parameters, establishing the first strong baseline for CPU-native real-time polyp segmentation. When scaled to 4.38M parameters, UltraSeg achieves accuracy competitive with heavyweight state-of-the-art models while maintaining an order-of-magnitude parameter advantage, demonstrating that the proposed design principles yield intrinsic representational gains across the entire efficiency spectrum. By delivering the first clinically deployable, CPU-native real-time solution, this work provides an immediately usable tool for resource-limited settings and a reproducible blueprint for real-time medical AI beyond endoscopy. Source code is publicly available.

cs.CV↗
arXiv2026

Solving Reach- and Stabilize-Avoid Problems Using Discounted Reachability

In this article, we consider the infinite-horizon reach-avoid (RA) and stabilize-avoid (SA) zero-sum game problems for general nonlinear continuous-time systems, where the goal is to find the set of states that can be controlled to reach or stabilize to a target set, without violating constraints even under the worst-case disturbance. Based on the Hamilton-Jacobi reachability method, we address the RA problem by designing a new Lipschitz continuous RA value function, whose zero sublevel set exactly characterizes the RA set. We establish that the associated Bellman backup operator is contractive and that the RA value function is the unique viscosity solution of a Hamilton-Jacobi variational inequality. Finally, we develop a two-step framework for the SA problem by integrating our RA strategies with a recently proposed Robust Control Lyapunov-Value Function, thereby ensuring both target reachability and long-term stability. We numerically verify our RA and SA frameworks on a 3D Dubins car system to demonstrate the efficacy of the proposed approach.

math.OC↗
arXiv2026

Optimization of laser-driven proton acceleration in a near-critical-density plasma

Optimizing laser and plasma parameters is crucial for enhancing accelerated proton energy in laser-driven proton acceleration with finite laser energy for applications such as cancer therapy. Tight focusing plays a significant role in improving laser-driven proton acceleration, which is generally believed as a result of the enhancement of laser intensity. However, we find that even at a fixed laser intensity, reducing the focal spot size still enhances the proton energy. Through particle-in-cell simulations and theoretical modeling, we find that at a small spot size (0.8 μm), the maximum proton energy is enhanced by 56.3% compared to that obtained at a conventional spot size (3 μm). This improvement is attributed to the dominance of ponderomotive-force-driven electrons at reduced spot sizes, which generate stronger charge-separation fields that propagate at higher velocities. Furthermore, to optimize proton acceleration, we analytically derive an ideal plasma density profile that promotes phase-stable proton acceleration, yielding an additional energy increase of 61.3% over the case of a tightly focused laser interacting with a planar target of uniform density. These findings remain robust under parameter variations, indicating that advanced focusing techniques combined with optimized plasma profiles could relax the demand for high laser energies, thereby reducing the reliance on large-scale laser facilities in medical and scientific applications.

physics.plasm-ph↗
arXiv2026

Robust Control Lyapunov-Value Functions for Nonlinear Disturbed Systems

Control Lyapunov Functions (CLFs) have been extensively used in the control community. A well-known drawback is the absence of a systematic way to construct CLFs for general nonlinear systems, and the problem can become more complex with input or state constraints. Our preliminary work on constructing Control Lyapunov Value Functions (CLVFs) using Hamilton-Jacobi (HJ) reachability analysis provides a method for finding a non-smooth CLF. In this paper, we extend our work on CLVFs to systems with bounded disturbance and define the Robust CLVF (R-CLVF). The R-CLVF naturally inherits all properties of the CLVF; i.e., it first identifies the "smallest robust control invariant set (SRCIS)" and stabilizes the system to it with a user-specified exponential rate. The region from which the exponential rate can be met is called the "region of exponential stabilizability (ROES)." We provide clearer definitions of the SRCIS and more rigorous proofs of several important theorems. Since the computation of the R-CLVF suffers from the "curse of dimensionality," we also provide two techniques (warmstart and system decomposition) that solve it, along with necessary proofs. Three numerical examples are provided, validating our definition of SRCIS, illustrating the trade-off between a faster decay rate and a smaller ROES, and demonstrating the efficiency of computation using warmstart and decomposition.

math.OC↗
arXiv2026

Frequency-Enhanced Hilbert Scanning Mamba for Short-Term Arctic Sea Ice Concentration Prediction

While Mamba models offer efficient sequence modeling, vanilla versions struggle with temporal correlations and boundary details in Arctic sea ice concentration (SIC) prediction. To address these limitations, we propose Frequency-enhanced Hilbert scanning Mamba Framework (FH-Mamba) for short-term Arctic SIC prediction. Specifically, we introduce a 3D Hilbert scan mechanism that traverses the 3D spatiotemporal grid along a locality-preserving path, ensuring that adjacent indices in the flattened sequence correspond to neighboring voxels in both spatial and temporal dimensions. Additionally, we incorporate wavelet transform to amplify high-frequency details and we also design a Hybrid Shuffle Attention module to adaptively aggregate sequence and frequency features. Experiments conducted on the OSI-450a1 and AMSR2 datasets demonstrate that our FH-Mamba achieves superior prediction performance compared with state-of-the-art baselines. The results confirm the effectiveness of Hilbert scanning and frequency-aware attention in improving both temporal consistency and edge reconstruction for Arctic SIC forecasting. Our codes are publicly available at https://github.com/oucailab/FH-Mamba.

eess.IV↗
arXiv2026

StoxLSTM: A Stochastic Extended Long Short-Term Memory Network for Time Series Forecasting

The Extended Long Short-Term Memory (xLSTM) network has demonstrated strong capability in modeling complex long-term dependencies in time series data. Despite its success, the deterministic architecture of xLSTM limits its representational capacity and forecasting performance, especially on challenging real-world time series datasets characterized by inherent uncertainty, stochasticity, and complex hierarchical latent dynamics. In this work, we propose StoxLSTM, a stochastic xLSTM within a designed state space modeling framework, which integrates latent stochastic variables directly into the recurrent units to effectively model deep latent temporal dynamics and uncertainty. The designed state space model follows an efficient non-autoregressive generative approach, achieving strong predictive performance without complex modifications to the original xLSTM architecture. Extensive experiments on publicly available benchmark datasets demonstrate that StoxLSTM consistently outperforms state-of-the-art baselines, achieving superior performance and generalization.

cs.LG↗
arXiv2025

Step-DeepResearch Technical Report

As LLMs shift toward autonomous agents, Deep Research has emerged as a pivotal metric. However, existing academic benchmarks like BrowseComp often fail to meet real-world demands for open-ended research, which requires robust skills in intent recognition, long-horizon decision-making, and cross-source verification. To address this, we introduce Step-DeepResearch, a cost-effective, end-to-end agent. We propose a Data Synthesis Strategy Based on Atomic Capabilities to reinforce planning and report writing, combined with a progressive training path from agentic mid-training to SFT and RL. Enhanced by a Checklist-style Judger, this approach significantly improves robustness. Furthermore, to bridge the evaluation gap in the Chinese domain, we establish ADR-Bench for realistic deep research scenarios. Experimental results show that Step-DeepResearch (32B) scores 61.4% on Scale AI Research Rubrics. On ADR-Bench, it significantly outperforms comparable models and rivals SOTA closed-source models like OpenAI and Gemini DeepResearch. These findings prove that refined training enables medium-sized models to achieve expert-level capabilities at industry-leading cost-efficiency.

cs.CL↗