arXiv subjects

Zhi Li

Publications and source records attributed to Zhi Li.

At least 73 records · Page 4Linked to original sources

arXiv2026

Connection Between the Shadow Radius and Quasinormal Frequencies for Black Holes in STVG with Perfect Fluid Dark Matter

We investigate the connection between black hole shadow and quasinormal mode (QNM) spectra in the context of scalar-tensor-vector gravity (STVG) coupled to perfect fluid dark matter (PFDM), characterized by the MOG parameter $α$ and the dark matter intensity $λ$. Employing complementary methods -- namely the sixth-order WKB approximation, Padé resummation, and time-domain numerical integration -- we compute QNM frequencies for scalar ($s=0$), electromagnetic ($s=1$), and axial gravitational ($s=2$) perturbations. Both the real part of the QNM frequencies and the peak height of the effective potential display a consistent parametric dependence: they increase with $λ$ yet decrease with growing $α$. In the eikonal limit ($l \gg 1$), we derive an exact analytical link between the shadow radius $R_{\mathrm{sh}}$ and the QNM frequency $ω_R$. Noting that $R_{\mathrm{sh}}$ is determined by the critical impact parameter $b_c = r_{\mathrm{ph}}/\sqrt{f(r_{\mathrm{ph}})}$, while $ω_R = Ωl$ with photon angular velocity $Ω= \sqrt{f(r_{\mathrm{ph}})}/r_{\mathrm{ph}}$, we obtain the precise relation $ω_R = l / b_c$, identifying $R_{\mathrm{sh}} \equiv b_c$ for an asymptotically flat observer. This prediction is robustly validated by numerical results across all three computational approaches at large multipole numbers. Our findings reveal that the black hole shadow and gravitational ringdown are not independent phenomena, but dual observational signatures of the same underlying structure -- the unstable photon orbit -- thereby offering a unified multi-messenger framework to simultaneously constrain modified gravity and dark matter in the strong-field regime.

gr-qc↗

arXiv2026

High Uniformity GaN Micro-pyramids and Platelets by Selective Area Growth

The development of uniform GaN micro-pyramids and platelets via selective area growth is a critical step toward advancing III-nitride device technologies, particularly for micro-light-emitting diode applications. This work investigates the origins of morphological non-uniformity in micro-pyramids and micro-platelets grown by metal-organic chemical vapor deposition (MOCVD). We observe that a direct one-step growth approach leads to significant growth rate inhomogeneity across arrays. To shed light on this issue, we examine the mechanisms driving non-uniformity and explore process modifications aimed at mitigating these effects. Building on these insights, we propose a controlled multi-step growth strategy that combines sequen-tial growth and thermal treatment phases. This approach is demonstrated to enhance surface morphology and structural regularity. The work contributes to the broader objective of enabling scalable, high-precision GaN microstructure fabrication for next-generation optoelectronic applications.

physics.optics↗

arXiv2026

Explicit States with Two-sided Long-Range Magic

Nonstabilizerness, or magic, is a necessary resource for quantum advantage beyond the classically simulatable Clifford framework. Recent works have begun to chart the structure of magic in many-body states, introducing the concepts of long-range magic -- nonstabilizerness that cannot be removed by finite-depth local unitary (FDU) circuits -- and the magic hierarchy, which classifies quantum circuits by alternating layers of Clifford and FDUs. In this work, we construct explicit states that provably possess two-sided long-range magic, a stronger form of magic meaning that they cannot be prepared by a Clifford circuit and a FDU in either order, thus placing them provably outside the first level of the magic hierarchy. Our examples include the ``magical cat" state, $|ψ\rangle \propto |0^n\rangle + |+^n\rangle$, and ground states of certain nonabelian topological orders. These results provide new examples and proof techniques for circuit complexity, and in doing so, reveal the connection between long-range magic, the structure of many-body phases, and the principles of quantum error correction.

quant-ph↗

arXiv2026

RCPU: Rotation-Constrained Error Compensation for Structured Pruning of Large Language Models

In this paper, we propose a rotation-constrained compensation method to address the errors introduced by structured pruning of large language models (LLMs). LLMs are trained on massive datasets and accumulate rich semantic knowledge in their representation space. In contrast, pruning is typically carried out with only a small amount of calibration data, which makes output mismatches unavoidable. Although direct least-squares fitting can reduce such errors, it tends to overfit to the limited calibration set, destructively modifying pretrained weights. To overcome this difficulty, we update the pruned parameters under a rotation constraint. This constrained update preserves the geometry of output representations (i.e., norms and inner products) and simultaneously re-aligns the pruned subspace with the original outputs. Furthermore, in rotation-constrained compensation, removing components that strongly contribute to the principal directions of the output makes error recovery difficult. Since input dimensions with large variance strongly affect these principal directions, we design a variance-aware importance score that ensures such dimensions are preferentially kept in the pruned model. By combining this scoring rule with rotation-constrained updates, the proposed method effectively compensates errors while retaining the components likely to be more important in a geometry-preserving manner. In the experiments, we apply the proposed method to Llama-7B and Llama-2-13B, and evaluate it on WikiText2 and multiple language understanding benchmarks. The results demonstrate consistently better perplexity and task accuracy compared with existing baselines.

cs.CL↗

arXiv2026

StreamDiffusionV2: A Streaming System for Dynamic and Interactive Video Generation

Generative models are reshaping the live-streaming industry by redefining how content is created, styled, and delivered. Previous image-based streaming diffusion models have powered efficient and creative live streaming products but have hit limits on temporal consistency due to the foundation of image-based designs. Recent advances in video diffusion have markedly improved temporal consistency and sampling efficiency for offline generation. However, offline generation systems primarily optimize throughput by batching large workloads. In contrast, live online streaming operates under strict service-level objectives (SLOs): time-to-first-frame must be minimal, and every frame must meet a per-frame deadline with low jitter. Besides, scalable multi-GPU serving for real-time streams remains largely unresolved so far. To address this, we present StreamDiffusionV2, a training-free pipeline for interactive live streaming with video diffusion models. StreamDiffusionV2 integrates an SLO-aware batching scheduler and a block scheduler, together with a sink-token--guided rolling KV cache, a motion-aware noise controller, and other system-level optimizations. Moreover, we introduce a scalable pipeline orchestration that parallelizes the diffusion process across denoising steps and network layers, achieving near-linear FPS scaling without violating latency guarantees. The system scales seamlessly across heterogeneous GPU environments and supports flexible denoising steps (e.g., 1--4), enabling both ultra-low-latency and higher-quality modes. Without TensorRT or quantization, StreamDiffusionV2 renders the first frame within 0.5s and attains 58.28 FPS with a 14B-parameter model and 64.52 FPS with a 1.3B-parameter model on four H100 GPUs, making state-of-the-art generative live streaming practical and accessible--from individual creators to enterprise-scale platforms.

cs.CV↗

arXiv2026

WebWorld: A Large-Scale World Model for Web Agent Training

Web agents require massive trajectories to generalize, yet real-world training is constrained by network latency, rate limits, and safety risks. We introduce \textbf{WebWorld} series, the first open-web simulator trained at scale. While existing simulators are restricted to closed environments with thousands of trajectories, WebWorld leverages a scalable data pipeline to train on 1M+ open-web interactions, supporting reasoning, multi-format data, and long-horizon simulations of 30+ steps. For intrinsic evaluation, we introduce WebWorld-Bench with dual metrics spanning nine dimensions, where WebWorld achieves simulation performance comparable to Gemini-3-Pro. For extrinsic evaluation, Qwen3-14B trained on WebWorld-synthesized trajectories improves by +9.2\% on WebArena, reaching performance comparable to GPT-4o. WebWorld enables effective inference-time search, outperforming GPT-5 as a world model. Beyond web simulation, WebWorld exhibits cross-domain generalization to code, GUI, and game environments, providing a replicable recipe for world model construction.

cs.AI↗

arXiv2026

Not All Negative Samples Are Equal: LLMs Learn Better from Plausible Reasoning

Learning from negative samples holds great promise for improving Large Language Model (LLM) reasoning capability, yet existing methods treat all incorrect responses as equally informative, overlooking the crucial role of sample quality. To address this, we propose Plausible Negative Samples (PNS), a method that synthesizes high-quality negative samples exhibiting expected format and structural coherence while ultimately yielding incorrect answers. PNS trains a dedicated model via reverse reinforcement learning (RL) guided by a composite reward combining format compliance, accuracy inversion, reward model assessment, and chain-of-thought evaluation, generating responses nearly indistinguishable from correct solutions. We further validate PNS as a plug-and-play data source for preference optimization across three backbone models on seven mathematical reasoning benchmarks. Results demonstrate that PNS consistently outperforms other negative sample synthesis methods, achieving an average improvement of 2.03% over RL-trained models.

cs.LG↗

arXiv2026

Human-in-the-Loop Failure Recovery with Adaptive Task Allocation

Since the recent Covid-19 pandemic, mobile manipulators and humanoid assistive robots with higher levels of autonomy have increasingly been adopted for patient care and living assistance. Despite advancements in autonomy, these robots often struggle to perform reliably in dynamic and unstructured environments and require human intervention to recover from failures. Effective human-robot collaboration is essential to enable robots to receive assistance from the most competent operator, in order to reduce their workload and minimize disruptions in task execution. In this paper, we propose an adaptive method for allocating robotic failures to human operators (ARFA). Our proposed approach models the capabilities of human operators, and continuously updates these beliefs based on their actual performance for failure recovery. For every failure to be resolved, a reward function calculates expected outcomes based on operator capabilities and historical data, task urgency, and current workload distribution. The failure is then assigned to the operator with the highest expected reward. Our simulations and user studies show that ARFA outperforms random allocation, significantly reducing robot idle time, improving overall system performance, and leading to a more distributed workload among operators.

cs.RO↗

arXiv2026

The Intrinsic Connection between Dynamical Phase Transitions and Magnetization in the 1D XY Model

In this manuscript, we study the quench dynamics of a transverse-field XY model starting from coherent Gibbs states. The results reveal that the initial strength of magnetization plays a crucial role in the emergence of dynamical quantum phase transitions. In concrete terms, when quenching within the same phase, through the properties of observables such as Fisher zeros and magnetization, we show that the stronger the initial magnetization, the more difficult the emergence of dynamical quantum phase transitions. The underlying mechanism is that the strong initial magnetization provides a directional effect, which inhibits the spin flipping in the process of quantum quench, making the dynamical quantum phase transition difficult to emerge. Since dynamical quantum phase transitions can be experimentally realized in various artificial systems, we hope that the physics predicted here can be experimentally verified in tabletop platforms.

quant-ph↗

arXiv2026

V2X-DSC: Multi-Agent Collaborative Perception with Distributed Source Coding Guided Communication

Collaborative perception improves 3D understanding by fusing multi-agent observations, yet intermediate-feature sharing faces strict bandwidth constraints as dense BEV features saturate V2X links. We observe that collaborators view the same physical world, making their features strongly correlated; thus receivers only need innovation beyond their local context. Revisiting this from a distributed source coding perspective, we propose V2X-DSC, a framework with a Conditional Codec (DCC) for bandwidth-constrained fusion. The sender compresses BEV features into compact codes, while the receiver performs conditional reconstruction using its local features as side information, allocating bits to complementary cues rather than redundant content. This conditional structure regularizes learning, encouraging incremental representation and yielding lower-noise features. Experiments on DAIR-V2X, OPV2V, and V2X-Real demonstrate state-of-the-art accuracy-bandwidth trade-offs under KB-level communication, and generalizes as a plug-and-play communication layer across multiple fusion backbones.

cs.CV↗

arXiv2026

Variance & Greediness: A comparative study of metric-learning losses

Metric learning is central to retrieval, yet its effects on embedding geometry and optimization dynamics are not well understood. We introduce a diagnostic framework, VARIANCE (intra-/inter-class variance) and GREEDINESS (active ratio and gradient norms), to compare seven representative losses, i.e., Contrastive, Triplet, N-pair, InfoNCE, ArcFace, SCL, and CCL, across five image-retrieval datasets. Our analysis reveals that Triplet and SCL preserve higher within-class variance and clearer inter-class margins, leading to stronger top-1 retrieval in fine-grained settings. In contrast, Contrastive and InfoNCE compact embeddings are achieved quickly through many small updates, accelerating convergence but potentially oversimplifying class structures. N-pair achieves a large mean separation but with uneven spacing. These insights reveal a form of efficiency-granularity trade-off and provide practical guidance: prefer Triplet/SCL when diversity preservation and hard-sample discrimination are critical, and Contrastive/InfoNCE when faster embedding compaction is desired.

cs.CV↗

arXiv2026

From Observations to Events: Event-Aware World Model for Reinforcement Learning

While model-based reinforcement learning (MBRL) improves sample efficiency by learning world models from raw observations, existing methods struggle to generalize across structurally similar scenes and remain vulnerable to spurious variations such as textures or color shifts. From a cognitive science perspective, humans segment continuous sensory streams into discrete events and rely on these key events for decision-making. Motivated by this principle, we propose the Event-Aware World Model (EAWM), a general framework that learns event-aware representations to streamline policy learning without requiring handcrafted labels. EAWM employs an automated event generator to derive events from raw observations and introduces a Generic Event Segmentor (GES) to identify event boundaries, which mark the start and end time of event segments. Through event prediction, the representation space is shaped to capture meaningful spatio-temporal transitions. Beyond this, we present a unified formulation of seemingly distinct world model architectures and show the broad applicability of our methods. Experiments on Atari 100K, Craftax 1M, and DeepMind Control 500K, DMC-GB2 500K demonstrate that EAWM consistently boosts the performance of strong MBRL baselines by 10%-45%, setting new state-of-the-art results across benchmarks. Our code is released at https://github.com/MarquisDarwin/EAWM.

cs.LG↗

arXiv2026

Data-Driven Exploration and Insights into Temperature-Dependent Phonons in Inorganic Materials

Phonons, quantized vibrations of the atomic lattice, are fundamental to understanding thermal transport, structural stability, and phase behavior in crystalline solids. Despite advances in computational materials science, most predictions of vibrational properties in large materials databases rely on the harmonic approximation and overlook crucial temperature-dependent anharmonic effects. Here, we present a scalable computational framework that combines machine learning interatomic potentials, anharmonic lattice dynamics, and high-throughput calculations to investigate temperature-dependent phonons across thousands of materials. By fine-tuning the universal M3GNet interatomic potential using high-quality phonon data, we improve phonon prediction accuracy by a factor of four while preserving computational efficiency. Integrating this refined model into a high-throughput implementation of the stochastic self-consistent harmonic approximation, we compute temperature-dependent phonons for 4,669 inorganic compounds. Our analysis identifies systematic elemental and structural trends governing anharmonic phonon renormalization, with particularly strong manifestations in alkali metals, perovskite-derived frameworks, and related systems. Machine learning models trained on this dataset identify key atomic-scale features driving strong anharmonicity, including weak bonding, large atomic radii, and specific coordination motifs. First-principles validation confirms that anharmonic effects can dramatically alter lattice thermal conductivity by factors of two to four in some materials. This work establishes a robust and efficient data-driven approach for predicting finite-temperature phonon behavior, offering new pathways for the design and discovery of materials with tailored thermal and vibrational properties.

cond-mat.mtrl-sci↗

arXiv2026

Lightweight Yet Secure: Secure Scripting Language Generation via Lightweight LLMs

The security of scripting languages such as PowerShell is critical given their powerful automation and administration capabilities, often exercised with elevated privileges. Today, securing these languages still demands substantial human effort to craft and enforce rules, imposing heavy burdens on typical administrators and creating critical production risks (e.g., misoperations that shut down servers).Large language models (LLMs) have demonstrated strong capabilities in code generation, vulnerability detection, and automated repair for languages like Python and JavaScript. However, their ability to assist with generating secure scripting-language code remains largely underexplored. In this paper, we present SecGenEval-PS, a benchmark designed to systematically evaluate LLMs on secure scripting generation, security analysis, and automated repair. Our results show that both proprietary and open-source models fall short in these areas. For instance, over 60% of PowerShell scripts produced by GPT-4o and o3-mini are insecure without structured guidance.To bridge this gap, we propose PSSec, a framework that combines data synthesis with fine-tuning to enhance model security capabilities. We develop a self-debugging agent that integrates static analyzers with the reasoning abilities of advanced LLMs to synthesize large-scale structured triplets of insecure scripts, violation analyses, and corresponding repairs. We then fine-tune lightweight LLMs (as small as 1.7B parameters) using supervised fine-tuning (SFT) and reinforcement learning (RL), enabling security-aware reasoning and the generation of secure PowerShell code.Across multiple LLM families, including GPT and Qwen, \textit{PSSec}-trained models match or surpass general-purpose large models on PowerShell security tasks while reducing inference cost by more than an order of magnitude.

cs.CR↗

arXiv2026

Tunable chiral spiral phases in a non-Hermitian Ising-Gamma spin chain

We study the influence of dissipation on the Ising-Gamma model. Through observables such as ground-state energy, order parameters, entanglement entropy, etc., we identify each phase region and provide the global phase diagram of the system. The results show that the region of the spiral phase will continuously expand with the increase of dissipation, gradually squeezing the original paramagnetic and antiferromagnetic phase regions. Remarkably, unlike the conservative system, the introduction of dissipation will cause two spiral phases with opposite chiralities to emerge simultaneously in the system, which provides a possibility for the manipulation of spiral chirality in cold atomic experiments. Moreover, we reveal the mechanism of the dependence of the transformation between these two spiral phases with distinct chirality on the strength of the relative coefficient of off-diagonal Gamma interactions in the Ising-Gamma model. Since both the relevant order parameters and dissipation can be well controlled within a detectable range, these phenomena can be observed in ultracold atomic experiments.

cond-mat.quant-gas↗

arXiv2026

Physically-Grounded Manifold Projection Model for Generalizable Metal Artifact Reduction in Dental CBCT

Metal artifacts in Dental CBCT severely obscure anatomical structures, hindering diagnosis. Current deep learning for Metal Artifact Reduction (MAR) faces limitations: supervised methods suffer from spectral blurring due to "regression-to-the-mean", while unsupervised ones risk structural hallucinations. Denoising Diffusion Models (DDPMs) offer realism but rely on slow, stochastic iterative sampling, unsuitable for clinical use. To resolve this, we propose the Physically-Grounded Manifold Projection (PGMP) framework. First, our Anatomically-Adaptive Physics Simulation (AAPS) pipeline synthesizes high-fidelity training pairs via Monte Carlo spectral modeling and patient-specific digital twins, bridging the synthetic-to-real gap. Second, our DMP-Former adapts the Direct x-Prediction paradigm, reformulating restoration as a deterministic manifold projection to recover clean anatomy in a single forward pass, eliminating stochastic sampling. Finally, a Semantic-Structural Alignment (SSA) module anchors the solution using priors from medical foundation models (MedDINOv3), ensuring clinical plausibility. Experiments on synthetic and multi-center clinical datasets show PGMP outperforms state-of-the-art methods on unseen anatomy, setting new benchmarks in efficiency and diagnostic reliability. Code and data: https://github.com/ricoleehduu/PGMP.

cs.CV↗

arXiv2025

The transport of angular momentum for massive stars I. Formation of slowly rotating WNE stars

The evolutionary scenario of early-type nitrogen-sequence Wolf-Rayet (WNE) stars predicts a slowly rotating subclass that typically forms after the red supergiant (RSG) phase. Their slow rotation rates are attributed to stellar winds that remove angular momentum transferred outward during core contraction. We incorporate improved prescriptions for internal gravity waves and the magnetic Tayler instability into single massive star evolution models. Our simulations successfully produce slowly rotating WNE stars and determine optimal parameters for both mechanisms ($A \ge 10$ for internal gravity waves (IGWs), $α= 0.01$ for revised Tayler instability (TSF)). The results demonstrate that the efficiency of angular momentum transfer in massive stars is significantly enhanced compared to low-mass stars, both processes can self-consistently explain the slow rotation of WNE stars, confirming their efficiency in angular momentum redistribution and providing crucial theoretical support for the existence of this predicted stellar population.

astro-ph.SR↗

arXiv2025

Emergent topological re-entrant phase transition in a generalized quasiperiodic modulated Su-Schrieffer-Heeger model

We study the topological properties of the one-dimensional generalized quasiperiodic modulated Su-Schrieffer-Heeger model. The results reveal that topological re-entrant phase transition emerges. Through the analysis of a real-space winding number , we divide the emergent topological re-entrant phase transitions into two types. The first is the re-entrant phase transition from the traditional topological insulator phase into the topological Anderson insulator phase, and the second is the re-entrant phenomenon from one topological Anderson insulator phase into another topological Anderson insulator phase. These two types of re-entrant phase transition correspond to bounded and unbounded cases of quasiperiodic modulation, respectively. Furthermore, we verify the above topological re-entrant phase transitions by analyzing the Lyapunov exponent and bulk gap. Since Su-Schrieffer-Heeger models have been realized in various artificial systems (such as cold atoms, optical waveguide arrays, ion traps, Rydberg atom arrays, etc.), the two types of topological re-entrant phase transition predicted in this paper are expected to be realized in the near future.

cond-mat.mes-hall↗