arXiv subjects

Tao Xu

Publications and source records attributed to Tao Xu.

At least 37 records · Page 2Linked to original sources

arXiv2026

Direct Collapse Black Hole Candidates from Decaying Dark Matter

Injecting 1-13.6 eV photons into the early universe can suppress the molecular hydrogen abundance and alter the star formation history dramatically enough to produce direct collapse black holes. These, in turn, could explain the recently observed population of puzzling high-redshift supermassive black holes that appear to require super-Eddington accretion. We show that axion dark matter decay in the intergalactic medium can account for this energy injection. We use a single zone model of the gas core and semi-analytically evolve its chemo-thermal properties to track the conditions for which the system becomes an atomic cooling halo-a necessary precursor for the production of heavy black hole seeds to explain the high-redshift black hole population. Windows of axions masses between 24.5-26.5 eV with photon couplings as low as $4\times 10^{-12}$/GeV may realize this atomic cooling halo condition. We highlight the significance of the band structure of molecular hydrogen on the effectiveness of this process and discuss estimates of the heavy seed population and prospects for testing this model.

hep-ph↗

arXiv2026

Inverse scattering theory for the discrete PT-symmetric nonlocal nonlinear Schröinger equation under arbitrarily large nonzero boundary conditions

In this paper, the theory of inverse scattering transform (IST) is developed for the discrete PT-symmetric nonlocal nonlinear Schröinger equation under large nonzero boundary conditions (NZBCs). By considering that the data at infinity have constant amplitudes, two cases are studied where the previous IST theory fails for large NZBCs. Based on a suitable uniformization variable, the rigorous proofs for the analyticity, symmetries and asymptotic behaviors of the eigenfunctions and scattering coefficients are provided for the direc problem, and the potential reconstruction formula is derived by solving the Riemann-Hilbert problem. Particularly, the focusing equation is found to admit two types of novel solitons under large NZBCs: oscillating soliton and breather, where the former has not been previously reported, while the latter does not occur under small NZBCs. In addition, the multi-soliton solutions are shown to exhibit the collisions among oscillating dark/anti-dark solitons, and the superposition of oscillating soliton and breather.

nlin.SI↗

arXiv2026

Huawei Cloud Model-as-a-Service on the CloudMatrix384 SuperPod

Scaled-out MoE LLMs and scaled-up SuperPods create new systems challenges for production Model-as-a-Service (MaaS), requiring disaggregation, low-latency communication, and decentralized serving. This report presents xDeepServe, the production serving system behind Huawei Cloud's MaaS offering on CloudMatrix384, a 48-server SuperPod with 384 Ascend 910C chips connected by a high-bandwidth UB fabric and global shared memory. It serves models including DeepSeek, Kimi, GLM, Qwen, and MiniMax, among others. xDeepServe is built around Transformerless, a disaggregated execution architecture that decomposes transformer inference into modular units -- attention, feedforward, and MoE -- and supports disaggregated Prefill-Decode and MoE-Attention deployments. To enable disaggregation, we develop XCCL, a memory-semantic communication layer providing microsecond-level point-to-point and scalable all-to-all primitives, and we extend FlowServe with decentralized DP groups and techniques to mitigate stragglers and synchronization variance. In a peak decoding configuration, xDeepServe reaches 2400 tokens/s per Ascend 910C chip at ~50ms time-per-output-token (TPOT).

cs.DC↗

arXiv2026

Index Light, Reason Deep: Deferred Visual Ingestion for Visual-Dense Document Question Answering

Existing multimodal document question answering methods predominantly adopt a Pre-Ingestion (PI) strategy: during the indexing phase, a Vision Language Model (VLM) is called on every page to generate page descriptions that are then encoded into vectors, and questions are answered via embedding similarity retrieval. However, this approach faces a dual dilemma on visual-dense engineering documents: VLM blind descriptions inevitably lose critical visual details, and embedding retrieval systematically fails on highly similar documents. This paper proposes the Deferred Visual Ingestion (DVI) framework: zero VLM calls during preprocessing, leveraging only document structural information (table of contents, drawing numbers) to automatically build a hierarchical index through the HDNC (Hierarchical Drawing Number Clustering) algorithm; during inference, candidate pages are located via BM25 retrieval, and the original images along with the specific question are sent to a VLM for targeted analysis. Large-scale experiments on three datasets validate the effectiveness of DVI: on Bridge engineering drawings (1,323 questions), end-to-end QA accuracy reaches 65.6\% vs. PI's 24.3\% (+41.3pp); on Steel catalog (186 questions), 30.6\% vs. 16.1\% (+14.5pp); on CircuitVQA, a public benchmark (9,315 questions), retrieval ImgR@3 achieves 31.2\% vs. 0.7\%. On the Bridge dataset, we evaluated ColPali (ICLR 2025 visual retrieval SOTA), which achieved only 20.1\% PageR@3, demonstrating that the failure of embedding retrieval on homogeneous engineering documents is structural rather than due to insufficient model capability. Ablation studies show that HDNC zero-cost automatic indexing yields a +27.5pp retrieval improvement, and VLM conversion rate analysis confirms that the bottleneck lies on the retrieval side rather than the comprehension side.

cs.CL↗

arXiv2026

AI-driven Large-scale Electron Microscopy enables Whole-tissue Subcellular Digitization

The distribution and interactions of cellular organelles play a critical role in mediating cellular physiology and pathology. Large-scale electron microscopy enables visualization of organelle distribution and interactions at the tissue level with nanometer resolution, but robust and efficient computational analysis tools are lacking. Here, we present a deep learning tool for universal large-scale 2D/3D electron microscopy analysis, DeepOrganelle. This new tool enables high-throughput, cell-resolved spatiotemporal mapping and digitization of organelle distribution and interactions. When applied to spermatogenesis across 12 stages and 22 differentiation status of the germ cells, DeepOrganelle uncovered previously unrecognized, stage-dependent dynamics of mitochondria-endoplasmic reticulum contact sites within one subphase of prophase I during meiosis. It also revealed coordinated organelle redistribution in Sertoli cells towards the blood-testis barrier, digitizing the remodeling dynamics of the tissue. This study demonstrates that DeepOrganelle provides a powerful framework that captures subcellular dynamics at the whole-tissue level.

physics.bio-ph↗

arXiv2026

Fast and Fewrious: Stochastic binary perturbations from fast compact objects

Massive compact objects soften binaries. This process has been used for decades to constrain the population of such objects, particularly as a component of dark matter (DM). The effects of light compact objects, such as those in the unconstrained asteroid-mass range, have generally been neglected. In principle, low-energy perturbers can harden binaries instead of softening them, but the standard lore is that this effect vanishes when the perturber velocities are large compared to the binary's orbital velocity, as is typical for DM constituents. Here, we revisit the computation of the hardening rate induced by light perturbers. We show that although the perturbations average to zero over many encounters, many scenarios of interest for DM constraints are in the regime where the variance cannot be neglected. We show that a few fast-moving perturbers can leave stochastic perturbations in systems that are measured with great precision, and we use this framework to revisit the constraint potential of systems such as binary pulsars and the Solar System. This opens a new class of dynamical probes with potential applications to asteroid-mass DM candidates.

hep-ph↗

arXiv2026

Physical Transformer

Digital AI systems spanning large language models, vision models, and generative architectures that operate primarily in symbolic, linguistic, or pixel domains. They have achieved striking progress, but almost all of this progress lives in virtual spaces. These systems transform embeddings and tokens, yet do not themselves touch the world and rarely admit a physical interpretation. In this work we propose a physical transformer that couples modern transformer style computation with geometric representation and physical dynamics. At the micro level, attention heads, and feed-forward blocks are modeled as interacting spins governed by effective Hamiltonians plus non-Hamiltonian bath terms. At the meso level, their aggregated state evolves on a learned Neural Differential Manifold (NDM) under Hamiltonian flows and Hamilton, Jacobi, Bellman (HJB) optimal control, discretized by symplectic layers that approximately preserve geometric and energetic invariants. At the macro level, the model maintains a generative semantic workspace and a two-dimensional information-phase portrait that tracks uncertainty and information gain over a reasoning trajectory. Within this hierarchy, reasoning tasks are formulated as controlled information flows on the manifold, with solutions corresponding to low cost trajectories that satisfy geometric, energetic, and workspace-consistency constraints. On simple toy problems involving numerical integration and dynamical systems, the physical transformer outperforms naive baselines in stability and long-horizon accuracy, highlighting the benefits of respecting underlying geometric and Hamiltonian structure. More broadly, the framework suggests a path toward physical AI that unify digital reasoning with physically grounded manifolds, opening a route to more interpretable and potentially unified models of reasoning, control, and interaction with the real world.

cs.LG↗

arXiv2025

Hawking Radiation of Nonrelativistic Scalars: Applications to Pion and Axion Production

In studying secondary gamma-ray emissions from Primordial Black Holes (PBHs), the production of scalar particles like pions and axion-like particles (ALPs) via Hawking radiation is crucial. While previous analyses assumed relativistic production, asteroid-mass PBHs, relevant to upcoming experiments like AMEGO-X, likely produce pions and ALPs non-relativistically when their masses exceed 10 MeV. To account for mass dependence in Hawking radiation, we revisit the greybody factors for massive scalars from Schwarzschild black holes, revealing significant mass corrections to particle production rates compared to the projected AMEGO-X sensitivity. We highlight the importance of considering non-relativistic $π^0$ production in interpreting PBH gamma-ray signals, essential for determining PBH properties. Additionally, we comment on the potential suppression of pion production due to form factor effects when producing extended objects via Hawking radiation. We also provide an example code for calculating the Hawking radiation spectrum of massive scalar particles.

hep-ph↗

arXiv2025

Dynamical Heating from Dark Compact Objects and Axion Minihalos: Implications for the 21-cm Signal

The temperature of baryons at the end of the cosmic dark ages can be inferred from observations of the 21-cm hyperfine transition in neutral hydrogen. Any energy injection from the dark sector can therefore be detected through these measurements. Dark compact objects and dark-matter substructures can modify the baryon temperature by transferring heat via dynamical friction. In this work, we evaluate the prospects for detecting dynamical friction-induced heating from dark compact objects with a mass in the range $10^2 M_{\odot}$ to $10^5 M_{\odot}$, as well as from axion minihalos, using upcoming 21-cm experiments. We find that both the 21-cm global signal and power-spectrum measurements will be sensitive to dark compact objects that constitute about 10% of the dark matter, and will substantially improve our sensitivity to axion-like particles with masses in the range $10^{-18}$ eV to $10^{-9}$ eV.

astro-ph.CO↗

arXiv2025

L1 Sample Flow for Efficient Visuomotor Learning

Denoising-based models, such as diffusion and flow matching, have been a critical component of robotic manipulation for their strong distribution-fitting and scaling capacity. Concurrently, several works have demonstrated that simple learning objectives, such as L1 regression, can achieve performance comparable to denoising-based methods on certain tasks, while offering faster convergence and inference. In this paper, we focus on how to combine the advantages of these two paradigms: retaining the ability of denoising models to capture multi-modal distributions and avoid mode collapse while achieving the efficiency of the L1 regression objective. To achieve this vision, we reformulate the original v-prediction flow matching and transform it into sample-prediction with the L1 training objective. We empirically show that the multi-modality can be expressed via a single ODE step. Thus, we propose \textbf{L1 Flow}, a two-step sampling schedule that generates a suboptimal action sequence via a single integration step and then reconstructs the precise action sequence through a single prediction. The proposed method largely retains the advantages of flow matching while reducing the iterative neural function evaluations to merely two and mitigating the potential performance degradation associated with direct sample regression. We evaluate our method with varying baselines and benchmarks, including 8 tasks in MimicGen, 5 tasks in RoboMimic \& PushT Bench, and one task in the real-world scenario. The results show the advantages of the proposed method with regard to training efficiency, inference speed, and overall performance. \href{https://song-wx.github.io/l1flow.github.io/}{Project Website.}

cs.RO↗

arXiv2025

ChemBOMAS: Accelerated BO in Chemistry with LLM-Enhanced Multi-Agent System

Bayesian optimization (BO) is a powerful tool for scientific discovery in chemistry, yet its efficiency is often hampered by the sparse experimental data and vast search space. Here, we introduce ChemBOMAS: a large language model (LLM)-enhanced multi-agent system that accelerates BO through synergistic data- and knowledge-driven strategies. Firstly, the data-driven strategy involves an 8B-scale LLM regressor fine-tuned on a mere 1% labeled samples for pseudo-data generation, robustly initializing the optimization process. Secondly, the knowledge-driven strategy employs a hybrid Retrieval-Augmented Generation approach to guide LLM in dividing the search space while mitigating LLM hallucinations. An Upper Confidence Bound algorithm then identifies high-potential subspaces within this established partition. Across the LLM-refined subspaces and supported by LLM-generated data, BO achieves the improvement of effectiveness and efficiency. Comprehensive evaluations across multiple scientific benchmarks demonstrate that ChemBOMAS set a new state-of-the-art, accelerating optimization efficiency by up to 5-fold compared to baseline methods.

cs.LG↗

arXiv2025

Topological transition and emergent elasticity of dislocation in skyrmion lattice: Beyond Kittel's magnetic-polar analogy

Magnetic and polar skyrmions exhibit topologically protected quasiparticle behavior, including emergent fields, deformation, and the formation of a densely packed skyrmion lattice, beyond conventional domain configurations described by Kittel's law. Analogous to atomic crystals, lattice defects, especially dislocations and their associated strain fields, are crucial for understanding the lattice behavior of skyrmions; however, their features and roles remain insufficiently understood. Here, we show that magnetic skyrmion dislocations develop a core-split structure due to a significant skyrmion elongation up to 180% of their original length, reaching a topological transition from a single skyrmion to two half-skyrmions. Despite such a distinct structure, the long-range strain fields around the dislocation perfectly obey conventional Volterra's elasticity theory, in contrast to polar skyrmion lattices, where skyrmion deformations cause a breakdown of the elasticity theory. Furthermore, an energetic analysis shows that Dzyaloshinskii-Moriya interaction drives the large skyrmion deformation of the dislocation core. Our findings not only clarify the coexistence of topological core-reconstruction and a robust long-range elastic field of dislocations in magnetic skyrmion lattices, but also reveal that magnetic and electric domains, long regarded as dual and analogous, exhibit fundamental differences when extended into the regime of collective topological quasiparticles.

cond-mat.mtrl-sci↗

arXiv2025

Robust Source-Free Domain Adaptation for Medical Image Segmentation based on Curriculum Learning

Recent studies have uncovered a new research line, namely source-free domain adaptation, which adapts a model to target domains without using the source data. Such a setting can address the concerns on data privacy and security issues of medical images. However, current source-free domain adaptation frameworks mainly focus on the pseudo label refinement for target data without the consideration of learning procedure. Indeed, a progressive learning process from source to target domain will benefit the knowledge transfer during model adaptation. To this end, we propose a curriculum-based framework, namely learning from curriculum (LFC), for source-free domain adaptation, which consists of easy-to-hard and source-to-target curricula. Concretely, the former curriculum enables the framework to start learning with `easy' samples and gradually tune the optimization direction of model adaption by increasing the sample difficulty. While, the latter can stablize the adaptation process, which ensures smooth transfer of the model from the source domain to the target. We evaluate the proposed source-free domain adaptation approach on the public cross-domain datasets for fundus segmentation and polyp segmentation. The extensive experimental results show that our framework surpasses the existing approaches and achieves a new state-of-the-art.

cs.CV↗

arXiv2025

Thermal Gravitons from Warm Inflation

In warm inflation (WI), the persistent thermal bath that is sustained by dissipative interactions with the inflaton field produces a stochastic background of gravitational waves (GWs). In this paper we study the production and evolution of these GWs. Specifically, we investigate the emission of thermal gravitons (gravitons emitted by a thermal bath) from particle scattering in the bath and the evolution of the corresponding GWs. We find that the bulk of thermal graviton production in WI occurs during the transition to radiation domination after inflation. Further, the energy density of thermal gravitons is enhanced by roughly one to two orders of magnitude compared to that in a radiation-dominated scenario with the same reheating temperature. We also calculate the spectrum of the resulting stochastic GW background and find that it has a distinctive shape, consisting of a peak at high frequencies ~100 GHz and an almost flat spectrum extending to low frequencies. The peak arises from emission of sub-horizon modes that follow the temperature of the bath. The flat part of the spectrum corresponds to the modes that exit the horizon during WI and re-enter during radiation domination. We show that the detection prospects for the high-frequency peak of the GW spectrum, while improved slightly compared to the radiation-dominated case, still remain challenging. The thermal spectrum's low-frequency plateau is typically subdominant to the amplitude of the standard vacuum tensor modes from inflation, although WI models can exist where the thermal graviton plateau surpasses the vacuum contribution without exceeding current observational limits on the tensor-to-scalar ratio. Furthermore, we calculate the thermal graviton contribution from WI to dark radiation and show that WI models are generally expected to satisfy current observational bounds, including those from the cosmic microwave background.

hep-ph↗

arXiv2025

Generating Transferrable Adversarial Examples via Local Mixing and Logits Optimization for Remote Sensing Object Recognition

Deep Neural Networks (DNNs) are vulnerable to adversarial attacks, posing significant security threats to their deployment in remote sensing applications. Research on adversarial attacks not only reveals model vulnerabilities but also provides critical insights for enhancing robustness. Although current mixing-based strategies have been proposed to increase the transferability of adversarial examples, they either perform global blending or directly exchange a region in the images, which may destroy global semantic features and mislead the optimization of adversarial examples. Furthermore, their reliance on cross-entropy loss for perturbation optimization leads to gradient diminishing during iterative updates, compromising adversarial example quality. To address these limitations, we focus on non-targeted attacks and propose a novel framework via local mixing and logits optimization. First, we present a local mixing strategy to generate diverse yet semantically consistent inputs. Different from MixUp, which globally blends two images, and MixCut, which stitches images together, our method merely blends local regions to preserve global semantic information. Second, we adapt the logit loss from targeted attacks to non-targeted scenarios, mitigating the gradient vanishing problem of cross-entropy loss. Third, a perturbation smoothing loss is applied to suppress high-frequency noise and enhance transferability. Extensive experiments on FGSCR-42 and MTARSI datasets demonstrate superior performance over 12 state-of-the-art methods across 6 surrogate models. Notably, with ResNet as the surrogate on MTARSI, our method achieves a 17.28% average improvement in black-box attack success rate.

cs.CV↗

arXiv2025

From Fluctuation to Polarization: Imprints of $O(1-10)\, \mathrm{Mpc}^{-1}$ Curvature Perturbations in CMB B-modes from Scalar-Induced Gravitational Waves

Probing primordial curvature perturbations on small scales, beyond those accessible using cosmic microwave background (CMB) primary anisotropies and Lyman-$α$ forest data, remains a major open challenge. Current constraints on the scalar power spectrum at these scales are either weak or rely heavily on model-dependent assumptions about small-scale structure. In this work, we propose a novel method to probe the small-scale primordial power spectrum using scalar-induced tensor perturbations, which are inevitably sourced by curvature perturbations at second order in cosmological perturbation theory. While induced tensor modes have traditionally been studied in the context of the stochastic gravitational wave background, we highlight a complementary observable: the distinctive pattern of B-mode polarization they imprint on the CMB. We compute the angular spectrum of these B-modes arising from enhanced scalar perturbations and show that the resulting signal can be competitive with inflationary predictions for values of the tensor-to-scalar ratio targeted in upcoming CMB experiments, most notably CMB-Stage 4. We map the region of the scalar power spectrum to which these future B-mode experiments will be sensitive and compare with existing constraints, finding it to exceed current sensitivities at $k \sim O(1-10)\, \mathrm{Mpc}^{-1}$. In addition to providing a new CMB-based probe of the small-scale power spectrum, this work also motivates dedicated B-mode searches at higher multipoles ($\ell \gtrsim 100$).

hep-ph↗

arXiv2025

Adversarial Patch Attack for Ship Detection via Localized Augmentation

Current ship detection techniques based on remote sensing imagery primarily rely on the object detection capabilities of deep neural networks (DNNs). However, DNNs are vulnerable to adversarial patch attacks, which can lead to misclassification by the detection model or complete evasion of the targets. Numerous studies have demonstrated that data transformation-based methods can improve the transferability of adversarial examples. However, excessive augmentation of image backgrounds or irrelevant regions may introduce unnecessary interference, resulting in false detections of the object detection model. These errors are not caused by the adversarial patches themselves but rather by the over-augmentation of background and non-target areas. This paper proposes a localized augmentation method that applies augmentation only to the target regions, avoiding any influence on non-target areas. By reducing background interference, this approach enables the loss function to focus more directly on the impact of the adversarial patch on the detection model, thereby improving the attack success rate. Experiments conducted on the HRSC2016 dataset demonstrate that the proposed method effectively increases the success rate of adversarial patch attacks and enhances their transferability.

cs.CV↗

arXiv2025

Boosting Adversarial Transferability for Hyperspectral Image Classification Using 3D Structure-invariant Transformation and Weighted Intermediate Feature Divergence

Deep Neural Networks (DNNs) are vulnerable to adversarial attacks, which pose security challenges to hyperspectral image (HSI) classification based on DNNs. Numerous adversarial attack methods have been designed in the domain of natural images. However, different from natural images, HSIs contains high-dimensional rich spectral information, which presents new challenges for generating adversarial examples. Based on the specific characteristics of HSIs, this paper proposes a novel method to enhance the transferability of the adversarial examples for HSI classification using 3D structure-invariant transformation and weighted intermediate feature divergence. While keeping the HSIs structure invariant, the proposed method divides the image into blocks in both spatial and spectral dimensions. Then, various transformations are applied on each block to increase input diversity and mitigate the overfitting to substitute models. Moreover, a weighted intermediate feature divergence loss is also designed by leveraging the differences between the intermediate features of original and adversarial examples. It constrains the perturbation direction by enlarging the feature maps of the original examples, and assigns different weights to different feature channels to destroy the features that have a greater impact on HSI classification. Extensive experiments demonstrate that the adversarial examples generated by the proposed method achieve more effective adversarial transferability on three public HSI datasets. Furthermore, the method maintains robust attack performance even under defense strategies.

cs.CV↗