Searcharxiv⌕ Search

arXiv subjects

Gang Li

Publications and source records attributed to Gang Li.

At least 73 records · Page 4Linked to original sources
arXiv2026

Final-state rescattering mechanism of the $Δ(1232)^{++}$ production in $Λ^+_c \to K^- π^+ p$ decay

We investigate the production of the $Δ(1232)^{++}$ resonance in the charmed baryon weak decay $Λ^+_c \to K^- π^+ p$, focusing on the $π^+ p$ final-state rescattering mechanism. The direct $W^+$ exchange diagram is expected to be suppressed, hence we adopt the $W^+$ internal emission process $Λ^+_c \to p \bar K^{*0}(892)$ followed by the subsequent decay $\bar{K}^{*0} \to K^- π^+$ as the dominant source of the final state particles. The $Δ(1232)^{++}$ resonance is then generated via $π^+ p$ rescattering within a triangle loop mechanism. Our calculations incorporate both the tree-level $\bar K^{*0}(892)$ and the dynamically generated $\bar{K}^*_0(700)$ state arising from the $S$-wave $K π$ final state interaction. We find that our theoretical results can reproduce the bump and peak structures in the $K^- π^+$ invariant mass distributions for the $\bar{K}^*_0(700)$ and $\bar{K}^{*0}(892)$, respectively. Meanwhile, the peak for the $Δ(1232)^{++}$ in the $π^+ p$ invariant mass distributions is also well described. The $Δ(1232)^{++}$ signal naturally emerges from rescattering effects, and adopting the pole parameters of $Δ(1232)$ resonance yields an improved description of the experimental data. In addition, we obtain a branching fraction ratio $\mathcal{B}[Λ_c^+ \to Δ(1232)^{++} K^-] / \mathcal{B}[Λ_c^+ \to p \bar{K}^{*0}(892)] \approx 0.5$, which is lower than the experimentally measured value. This discrepancy suggests that interference effects are likely significant in this decay process. Future high-precision measurements will further verify the proposed rescattering mechanism.

hep-ph↗
arXiv2026

Repurposing Adversarial Perturbations for Continual Learning: From Defense to Active Alignment

In dynamic environments, large language models need to keep adapting to new tasks, but continual learning often suffers from forgetting, limited transfer, and vulnerability to adversarial perturbations. To address this, we present AdvCL, which repurposes adversarial perturbations as a geometric control signal for stable continual adaptation. AdvCL combines three plug-in modules: Intra-Smooth promotes local smoothness via small adversarial perturbations; Proto-Clip uses similarity clipping to prevent excessive alignment to current task prototype; and Inter-Align applies directional alignment toward previous task prototype to reduce representational gaps. Experiments show consistent gains in both standard performance and robustness, with lower forgetting and stronger transfer. We further analyze key mechanisms by quantifying the sensitivity of Intra-Smooth to perturbation settings and the effect of Inter-Align on task similarity and geometric distance. In summary, the modules provide complementary gains when combined, and each can also be integrated individually into diverse CL paradigms, including replay, regularization, and dynamic architectures, thereby offering a geometric control mechanism for continual learning.

cs.LG↗
arXiv2026

Where to Refine, When to Stop: Rethinking Redundancy via Latent Discrepancy for Efficient Visual Autoregressive Generation

Visual Autoregressive (VAR) models deliver high-quality image generation but suffer from significant inference latency at high resolutions. Recent acceleration approaches most rely on heuristic measures with layer features to prune tokens. Such heuristics are sensitive to complex contextual semantics, leading to inaccurate identification of redundant computation and poor adaptability across prompts. We rethink redundancy in VAR from the perspective of its impact on pixel-space generation and introduce Latent Discrepancy. This unified metric quantifies a token's contribution by measuring the change in model states during generation. Our analysis shows that redundancy is more accurately identified when guided by image latent or pixel-space signals. We further observed that in classifier-free guidance (CFG), the convergence trend of the discrepancy between conditional and unconditional branches exhibits high dynamics with different prompts. Based on these findings, we propose LD-Pruning (Latent Discrepancy Pruning), a training-free framework that removes redundancy via latent discrepancy by integrating decoding-free region selection and adaptive unconditional-branch skipping. Extensive experiments show that LD-Pruning substantially reduces inference latency while maintaining high generation quality, achieving up to 2.35x speedup on Infinity-8B.

cs.CV↗
arXiv2026

Dasheng AudioGen: A Unified Model for Generating Coherent Audio Scenes from Text

Audio generation has long been fragmented, with speech, music, and sound effects produced by domain-specific models that fail to jointly generate coherent audio scenes from a single description. The key obstacles are insufficient fine-grained supervision for real-world mixed audio and limited acoustic representations for modeling concurrent audio components. We present Dasheng AudioGen, a unified framework for generating general mixed-audio scenes from text. Dasheng AudioGen introduces structured multi-view captions, which explicitly decouple complex acoustic scenes into complementary description views, thereby enabling fine-grained control over audio layers. Furthermore, we employ a high-dimensional unified semantic-acoustic representation as the shared latent space. It injects semantic priors that facilitate cross-modal training convergence, while its high-dimensional feature space provides sufficient capacity to disentangle and fuse concurrent audio components effectively. With these designs, a simple flow-matching DiT achieves high-quality end-to-end audio scene generation. We also establish a comprehensive evaluation pipeline for audio scene generation. Experiments demonstrate that Dasheng AudioGen achieves performance approaching real-world recordings in mixed-audio categories, while remaining competitive with specialized models in single-type generation tasks. Demos are available at https://nieeim.github.io/Dasheng-AudioGen-Web/.

cs.SD↗
arXiv2026

Probing Solar Wind Structures with Solar Energetic Particle Observations from Solar Orbiter

The propagation of solar energetic particles (SEPs) through the heliosphere is primarily guided by the interplanetary magnetic field (IMF) which is embedded in the solar wind plasma. Large-scale IMF structures can drive transient variations in SEP intensities. Using Solar Orbiter observations, we identify a distinct class of SEP variations: SEP flux deflections (SFDs), which are commonly detected in SEP events and frequently recur multiple times within a single event. SFDs are characterized by a sudden change in SEP flux directions where the intensities drop in one direction and increase in another direction, without a significant net change in total flux magnitude. These deflections occur dispersionlessly across a broad energy range-from tens of keV to over 100 MeV-and exhibit steep intensity gradients. SFDs are typically associated with magnetic flux tubes with boundary features consistent with tangential discontinuities. We further show that the solar wind inside these structures exhibits distinct plasma properties, and that the SEP streaming direction within SFDs aligns closely to the flux-tube axis. These observations suggest that magnetic flux tubes are a prevalent structural element of the solar wind, and demonstrate that SEPs can serve as an effective diagnostic tool for probing the topology and dynamics of solar wind structures.

astro-ph.SR↗
arXiv2026

ReSS: Learning Reasoning Models for Tabular Data Prediction via Symbolic Scaffold

Tabular data remains prevalent in high-stakes domains such as healthcare and finance, where predictive models are expected to provide both high accuracy and faithful, human-understandable reasoning. While symbolic models offer verifiable logic, they lack semantic expressiveness. Meanwhile, general-purpose LLMs often require specialized fine-tuning to master domain-specific tabular reasoning. To address the dual challenges of scalable data curation and reasoning consistency, we propose ReSS, a systematic framework that bridges symbolic and neural reasoning models. ReSS leverages a decision-tree model to extract instance-level decision paths as symbolic scaffolds. These scaffolds, alongside input features and labels, guide an LLM to generate grounded natural-language reasoning that strictly adheres to the underlying decision logic. The resulting high-quality dataset is used to fine-tune a pretrained LLM into a specialized tabular reasoning model, further enhanced by a scaffold-invariant data augmentation strategy to improve generalization and explainability. To rigorously assess faithfulness, we introduce quantitative metrics including hallucination rate, explanation necessity, and explanation sufficiency. Experimental results on medical and financial benchmarks demonstrate that ReSS-trained models improve traditional decision trees and standard fine-tuning approaches up to $10\%$ while producing faithful and consistent reasoning

cs.AI↗
arXiv2026

WARC-Bench: Web Archive Based Benchmark for GUI Subtask Executions

Training web agents to navigate complex, real-world websites requires them to master $\textit{subtasks}$ - short-horizon interactions on multiple UI components (e.g., choosing the correct date in a date picker, or scrolling in a container to extract information). We introduce WARC-Bench (Web Archive Benchmark), a novel web navigation benchmark featuring 438 tasks designed to evaluate multimodal AI agents on subtasks. WARC-Bench enables sandboxed interactions with dynamic and realistic webpages using Web ARChive files. We show that WARC-Bench is challenging for leading computer-use models, with the highest observed success rate being 64.8%. To improve open source models on subtask, we explore two common training techniques: supervised fine-tuning (SFT) and reinforcement learning with verifiable rewards (RLVR). Experiments show that SFT models obtain a 48.8% success rate on the benchmark. Training with RLVR over SFT checkpoints, even in data-scarce settings, improves the score to 52.8% on WARC-Bench, outperforming many frontier models. Our analysis concludes that mastering these subtasks is essential for robust web planning and navigation, and is a capability not extensively evaluated by existing benchmarks.

cs.LG↗
arXiv2026

Knowledge-to-Verification: Exploring RLVR for LLMs in Knowledge-Intensive Domains

Reinforcement learning with verifiable rewards (RLVR) has demonstrated promising potential to enhance the reasoning capabilities of large language models (LLMs) in domains such as mathematics and coding. However, its applications on knowledge-intensive domains have not been effectively explored due to the scarcity of high-quality verifiable data. Furthermore, current RLVR focuses solely on the correctness of final answers, leading to the limitations of flawed reasoning and sparse reward signals. In this work, we propose Knowledge-to-Verification (K2V), a framework that extends RLVR to knowledge-intensive domains through automated verifiable data synthesis, while enabling verification of the LLM's reasoning process. Extensive experiments demonstrate that K2V enhances the reasoning of LLM in knowledge-intensive domains without significantly compromising the model's general capabilities. This study also suggests that integrating automated data synthesis with reasoning verification is a promising direction to enhance model capabilities in these broader domains. Code is available at https://github.com/SeedScientist/K2V.

cs.CL↗
arXiv2026

A Loewner-Nirenberg phenomena for Ricci flow on compact manifolds with boundary.II

This is a continuation of the research in [16]. Let $(\overline{M},g_{-1})$ be a closed geodesic $r_0$-ball in the hyperbolic space $(\mathbb{H}^n,g_{-1})$. Let $m\neq1$ be a positive constant. In this paper, we show that for $n\geq3$, starting from the metric $m g_{-1}$ on $\overline{M}$, with certain prescribed non-decreasing rotationally symmetric mean curvature and the fixed conformal class $[g_{\mathbb{S}^{n-1}}]$ on the boundary $\partial M$, the solution $g(t)$ to the normalized Ricci flow $(1.2)$ which is continuous up to the boundary, exists for all $t>0$, and converges locally uniformly in the interior $M$ of $\overline{M}$ to a complete hyperbolic metric as $t\to\infty$(see Theorem 1.1 for details). Under some additional conditions, we show the same conclusion holds for $n=2$.

math.DG↗
arXiv2026

PD-4DGS:Progressive Decomposition of 4D Gaussian Splatting for Bandwidth-Adaptive Dynamic Scene Streaming

4D Gaussian Splatting (4DGS) enables high-quality dynamic novel view synthesis, yet current models remain monolithic bitstreams that clients must download in full before any frame can be rendered, causing black-screen waits of tens to hundreds of seconds on mobile bandwidth and leaving 4DGS incompatible with modern adaptive-bitrate delivery. Progressive 3DGS compression alleviates this for static scenes, but it acts only on spatial anchors and cannot partition the temporal deformation networks that dominate dynamic-scene size. We present PD-4DGS, the first framework for progressive compression and on-demand transmission of 4DGS. Hierarchical Deformation Decomposition (HDD) externalises the coarse-to-fine motion hierarchy already latent in 4DGS into three independently transmittable layers -- a static scaffold, a global deformation, and a local refinement -- so that any prefix of the bitstream is already renderable, turning a single training run into a scalable, DASH/HLS-compatible bitstream. A Gaussian-entropy attribute rate-distortion loss together with a temporal mask consistency regulariser shrink the base layer while suppressing low-bitrate flicker; a capacity-weighted rollout schedule, gated online by a learnt activation rate rho, then prevents deformation-network under-training without any per-scene hyperparameter. On the Dycheck iPhone benchmark, PD-4DGS cuts the streamed bitstream by >60% at matched rendering fidelity and reduces first-frame latency from 73--930 s to ~1.7 s on a 2 Mbps link, uniquely enabling true on-demand progressive streaming for 4DGS.

cs.CV↗
arXiv2026

MECAT: A Multi-Experts Constructed Benchmark for Fine-Grained Audio Understanding Tasks

While large audio-language models have advanced open-ended audio understanding, they still fall short of nuanced human-level comprehension. This gap persists largely because current benchmarks, limited by data annotations and evaluation metrics, fail to reliably distinguish between generic and highly detailed model outputs. To this end, this work introduces MECAT, a Multi-Expert Constructed Benchmark for Fine-Grained Audio Understanding Tasks. Generated via a pipeline that integrates analysis from specialized expert models with Chain-of-Thought large language model reasoning, MECAT provides multi-perspective, fine-grained captions and open-set question-answering pairs. The benchmark is complemented by a novel metric: DATE (Discriminative-Enhanced Audio Text Evaluation). This metric penalizes generic terms and rewards detailed descriptions by combining single-sample semantic similarity with cross-sample discriminability. A comprehensive evaluation of state-of-the-art audio models is also presented, providing new insights into their current capabilities and limitations. The data and code are available at https://github.com/xiaomi-research/mecat

eess.AS↗
arXiv2026

Data-Constrained Modeling of Electron Transport and Asymmetric Precipitation in the 2011 August 4 Solar Flare

Energetic electrons accelerated at coronal reconnection sites during solar flares precipitate into the lower solar atmosphere, generating nonthermal emissions and regulating energy deposition. However, how their transport and precipitation are jointly governed by the three-dimensional (3D) magnetic topology, turbulent scattering, and Coulomb collisions remains unclear. Here, we aim to disentangle these physical processes by using a data-constrained 3D particle transport model for the 2011 August 4 flare. The simulated distribution of precipitated electrons aligns closely with photospheric quasi-separatrix layers and reproduces the observed two-ribbon morphology in 1700~Å. We reveal a strong polarity asymmetry, with the 10~s precipitation fraction about six times higher in the weak positive polarity. This arises primarily from distinct mirror ratios of different polarities under the 3D magnetic configuration and can be understood via a modified escape probability for an asymmetric magnetic bottle. Varying strengths of turbulent scattering lead to a rise-then-fall trend and a pronounced energy dependence in the precipitation fraction. Coulomb collisions globally suppress precipitation, especially at low energies, and further amplify the polarity asymmetry. This integrated modeling framework bridges detailed transport physics to observable flare emissions and advances the development of quantitative models for realistic solar flare events.

astro-ph.SR↗
arXiv2026

Memory-Efficient Continual Learning with CLIP Models

Contrastive Language-Image Pretraining (CLIP) models excel at understanding image-text relationships but struggle with adapting to new data without forgetting prior knowledge. To address this, models are typically fine-tuned using both new task data and a memory buffer of past tasks. However, CLIP's contrastive loss suffers when the memory buffer is small, leading to performance degradation on previous tasks. We propose a memory-efficient, distributionally robust method that dynamically reweights losses per class during training. Our approach, tested on class incremental settings (CIFAR-100, ImageNet1K) and a domain incremental setting (DomainNet) adapts CLIP models quickly while minimizing catastrophic forgetting, even with minimal memory usage.

cs.LG↗
arXiv2026

Multi-fidelity surrogates for mechanics of composites: from co-kriging to multi-fidelity neural networks

Composite materials exhibit strongly hierarchical and anisotropic properties governed by coupled mechanisms spanning constituents, plies, laminates, structures, and manufacturing history. This intrinsic complexity makes predictive modeling of composites expensive, because repeated experiments and high-fidelity simulations are needed to cover large design spaces of material, structure, and manufacturing. Multi-fidelity surrogate modeling addresses this challenge by combining abundant, less expensive data with limited high-accuracy data to recover reliable high-fidelity predictions. This review presents a structured overview of multi-fidelity modeling for composite mechanics, covering Gaussian-process or Kriging-based methods, including co-Kriging, coregionalization models, autoregressive formulations, nonlinear autoregressive Gaussian processes, multi-fidelity deep Gaussian processes, and multi-fidelity neural networks. Their distinctions are examined in terms of cross-fidelity correlation, discrepancy representation, uncertainty quantification, and scalability. Selected examples of their applications to composites are introduced according to the roles that multi-fidelity surrogates play in engineering problems, including forward prediction for rapid exploration of material design spaces, inverse optimization for composite parameter identification and design search under limited high-fidelity access, and workflow integration, where heterogeneous data sources, constraints, and validation requirements determine model utility. Open question discussions highlight recurring challenges specific to composites, such as regime-dependent fidelity gaps associated with nonlinear damage and manufacturing history, mismatches between simulations and experiments, and uncertainty propagation across multi-fidelity models.

physics.comp-ph↗
arXiv2026

Two distinct superconducting regimes in Ti4Co2O under pressures

We report on the pressure dependence of superconducting transition temperature Tc and upper critical field Bc2(0) through electrical transport of the Ti4Co2O superconductor (eg.,the superconducting transition temperature Tc = 2.5 K and the Bc2(0)=7.2T=2.9Tc). We find that the Tc exhibits non-monotonic pressure dependence:it rises monotonically at first with a pressure coefficient of dTc/dP=0.034 K/GPa, but rapidly decreases around 10-20 GPa, and then increases with the dTc/dP = 0.023 K/GPa, up to= 4.31 K at 69.7 GPa. Concurrently, the Bc2(0)exhibits a dome shaped pressure dependence, with its maximum at 5 GPa of almost twice the value at ambient pressure, exceeding the weak-coupling Pauli paramagnetic limit Bp throughout the whole pressure range. By comparing the normal-state and superconducting properties, we identify two distinct superconducting regimes, with a low-pressure superconducting phase characterized by an enhanced Bc2(0)values and Fermi-liquid normal-state electrical transport (the exponent n = 2), and a high-pressure superconducting phase with a monotonically increased Tc and an enhancement in phonon scatterings (the exponent n = 4). Room-temperature synchrotron X-ray diffraction indicates that there is no structural transition up to 55.8 GPa, which gives a relatively large bulk modulus of 192 GPa in comparison with other alloy superconductors. First-principles calculations suggest that the nonmonotonic Tc maybe closely related to the evolution of the density of states of Ti4Co2O upon compression, which is different from those of isostructural superconductors Ti4Ir2O and Nb4Rh2C. Our results show that even in the Ti4Co2O with weak spin-orbit coupling, superconductivity remains highly sensitive to the external stimuli such as pressure.

cond-mat.supr-con↗
arXiv2026

Hunting for $B\bar B$ molecular state $X_{b0}$ via radiative transition of $Υ(10753)$

We investigate the radiative decay $Υ(10753) \to γX_{b0}$ within the framework of nonrelativistic effective field theory (NREFT). The $Υ(10753)$ is treated as an $S$-$D$ mixed state of the $Υ(4S)$ and $\YD$, while the $X_{b0}$ is interpreted as a weakly bound $B\bar{B}$ molecule with $J^{PC}=0^{++}$. The decay process was assumed to occur via the intermediate meson loops involving the $S$-wave $B^{(*)}$ and $P$-wave $B_1^{(\prime)}$ mesons. Our calculated results indicate that the decay $Υ(10753) \to γX_{b0}$ is dominated by the $B_1^{(\prime)}$ meson loops. The partial decay width is predicted to be $0.2-1.5~\mathrm{keV}$ for a binding energy range of $ε_X=0-10~\mathrm{MeV}$, corresponding to a branching fraction of $10^{-6}-10^{-5}$. The decay width is found to be insensitive to the full width of the $B_1^{\prime}$ meson. Our study suggests that the radiative decay of the $Υ(10753)$ is a promising channel to search for the $X_{b0}$ state, which is crucial for testing heavy-quark symmetries and understanding the exotic hadron spectrum in the bottom sector.

hep-ph↗
arXiv2026

Probabilistic Predictions of Process-Induced Deformation in Carbon/Epoxy Composites Using a Deep Operator Network

Fiber reinforcement and polymer matrix respond differently to manufacturing conditions due to mismatch in coefficient of thermal expansion and matrix shrinkage during curing of thermosets. These heterogeneities generate residual stresses over multiple length scales, whose partial release leads to process-induced deformation (PID), requiring accurate prediction and mitigation via optimized non-isothermal cure cycles. This study considers a unidirectional AS4 carbon fiber/amine bi-functional epoxy prepreg and models PID using a two-mechanism framework that accounts for thermal expansion/shrinkage and cure shrinkage. The model is validated against manufacturing trials to identify initial and boundary conditions, then used to generate PID responses for a diverse set of non-isothermal cure cycles (time-temperature profiles). Building on this physics-based foundation, we develop a data-driven surrogate based on Deep Operator Networks (DeepONets). A DeepONet is trained on a dataset combining high-fidelity simulations with targeted experimental measurements of PID. We extend this to a Feature-wise Linear Modulation (FiLM) DeepONet, where branch-network features are modulated by external parameters, including the initial degree of cure, enabling prediction of time histories of degree of cure, viscosity, and deformation. Because experimental data are available only at limited time instances (for example, final deformation), we use transfer learning: simulation-trained trunk and branch networks are fixed and only the final layer is updated using measured final deformation. Finally, we augment the framework with Ensemble Kalman Inversion (EKI) to quantify uncertainty under experimental conditions and to support optimization of cure schedules for reduced PID in composites.

cs.CE↗
arXiv2026

Certain Head, Uncertain Tail: Expert-Sample for Test-Time Scaling in Fine-Grained MoE

Test-time scaling improves LLM performance by generating multiple candidate solutions, yet token-level sampling requires temperature tuning that trades off diversity against stability. Fine-grained MoE, featuring hundreds of well-trained experts per layer and multi-expert activation per token, offers an unexplored alternative through its rich routing space. We empirically characterize fine-grained MoE routing and uncover an informative pattern: router scores exhibit a certain head of high-confidence experts followed by an uncertain tail of low-confidence candidates. While single-run greedy accuracy remains stable when fewer experts are activated, multi-sample pass@n degrades significantly-suggesting that the certain head governs core reasoning capability while the uncertain tail correlates with reasoning diversity. Motivated by these findings, we propose Expert-Sample, a training-free method that preserves high-confidence selections while injecting controlled stochasticity into the uncertain tail, enabling diverse generation without destabilizing outputs. Evaluated on multiple fine-grained MoE models across math, knowledge reasoning, and code tasks, Expert-Sample consistently improves pass@n and verification-based accuracy. On Qwen3-30B-A3B-Instruct evaluated on GPQA-Diamond with 32 parallel samples, pass@32 rises from 85.4% to 91.9%, and accuracy improves from 59.1% to 62.6% with Best-of-N verification.

cs.LG↗