arXiv subjects

Liang Zhang

Publications and source records attributed to Liang Zhang.

At least 91 records · Page 5Linked to original sources

arXiv2026

A NICER view of the corona through time-dependent Comptonization of the quasi-periodic oscillations in nine black-hole X-ray binaries

We present a systematic study of the evolution of the corona geometry in nine black hole X-ray binaries (BHXRBs) using archival data from NICER. We identify 171 observations exhibiting quasi-periodic oscillations (QPOs) across various spectral states and model the time-averaged energy spectra of the source, as well as the energy-dependent rms and phase-lag spectra of the QPO, with the time-dependent Comptonization model vKompthdk. This allows us to simultaneously constrain the corona size and feedback fraction during outbursts. By using the power color hue diagnostics, we identify different spectral states, and observe that the QPO frequency increases from $\sim$0.1 Hz to $\sim$10 Hz in the low-hard and hard-intermediate states (LHS and HIMS), and remains approximately constant at 4--5~Hz in the soft-intermediate state (SIMS). The corona size shows significant evolution: the corona is large ($\sim10^4$--$10^5$ km) in the LHS, contracts rapidly to $\sim10^3$ km in the HIMS, and exhibits a flare-like expansion near the HIMS-to-SIMS transition. In the SIMS and high-soft state (HSS), the corona becomes compact and stable (4000--8000~km). The feedback fraction of the corona photons increases during the periods in which the corona contracts and decreases during the periods in which the corona expands, indicating a change of the disk-corona coupling. Our results are consistent with previous QPO-based studies using vKompthdk on some individual sources. This work, however, provides the first view of the coronal evolution across outbursts for a diverse BHXRB sample, offering critical insights into coronal behavior as a function of the spectral state of the source.

astro-ph.HE↗

arXiv2026

Long-term timing evolution of four Anomalous X-Ray Pulsars

Anomalous X-ray pulsars (AXPs) and soft gamma-ray repeaters (SGRs) are believed to be manifestations of magnetars. Typically, AXPs exhibit higher X-ray luminosities, whereas SGRs are generally fainter and display significantly high signal-to-noise ratios only during their outburst phases. In this work, we report the long-term timing evolution of four AXPs: 1E 2259+586, 4U 0142+61, 1RXS J170849.0-400910 and 1E 1841-045, which were regularly monitored with NICER from 2017 to 2024. Over this period, we identify a total of 10 timing events. In addition to one glitch and one anti-glitch in 1E 2259+586 reported in literature, we detect another 8 new timing events: 5 glitches, 2 anti-glitches, and 1 unusual state transition event. Notably, both anti-glitches were observed in 4U 0142+61, making it the most frequent source of such events, and there is a hint of regular evolution in its pulse profile. In the case of 1RXS J170849.0-400910, it continues to exhibit pronounced high-frequency timing anomalies and undergoes a state transition event. Finally, we study the evolution of the pulse profiles and find that the profiles of 1E 2259+586 and 4U 0142+61 both evolve. This is consistent with the earlier finding that pulse profile evolution is a generic feature of magnetars.

astro-ph.HE↗

arXiv2026

A State-Transition Framework for Efficient LLM Reasoning

While Long Chain-of-Thought (CoT) reasoning significantly improves Large Language Models (LLMs) performance on complex reasoning tasks, the substantial computational and memory costs of generating long CoT sequences limit their efficiency and practicality. Existing studies usually enhance the reasoning efficiency of LLMs by compressing CoT sequences. However, this approach conflicts with test-time scaling, limiting the reasoning capacity of LLMs. In this paper, we propose an efficient reasoning framework that models the reasoning process of LLMs as a state-transition process. Specifically, we first apply a linear attention mechanism to estimate the LLM's reasoning state, which records the historical reasoning information from previous reasoning steps. Then, based on the query prompt and the reasoning state, the LLM can efficiently perform the current reasoning step and update the state. With the linear attention, each token in the current reasoning step can directly retrieve relevant historical reasoning information from the reasoning state, without explicitly attending to tokens in previous reasoning steps. In this way, the computational complexity of attention is reduced from quadratic to linear, significantly improving the reasoning efficiency of LLMs. In addition, we propose a state-based reasoning strategy to mitigate the over-thinking issue caused by noisy reasoning steps. Extensive experiments across multiple datasets and model sizes demonstrate that our framework not only improves the reasoning efficiency of LLMs but also enhances their reasoning performance.

cs.AI↗

arXiv2026

FSD-CAP: Fractional Subgraph Diffusion with Class-Aware Propagation for Graph Feature Imputation

Imputing missing node features in graphs is challenging, particularly under high missing rates. Existing methods based on latent representations or global diffusion often fail to produce reliable estimates, and may propagate errors across the graph. We propose FSD-CAP, a two-stage framework designed to improve imputation quality under extreme sparsity. In the first stage, a graph-distance-guided subgraph expansion localizes the diffusion process. A fractional diffusion operator adjusts propagation sharpness based on local structure. In the second stage, imputed features are refined using class-aware propagation, which incorporates pseudo-labels and neighborhood entropy to promote consistency. We evaluated FSD-CAP on multiple datasets. With $99.5\%$ of features missing across five benchmark datasets, FSD-CAP achieves average accuracies of $80.06\%$ (structural) and $81.01\%$ (uniform) in node classification, close to the $81.31\%$ achieved by a standard GCN with full features. For link prediction under the same setting, it reaches AUC scores of $91.65\%$ (structural) and $92.41\%$ (uniform), compared to $95.06\%$ for the fully observed case. Furthermore, FSD-CAP demonstrates superior performance on both large-scale and heterophily datasets when compared to other models.

cs.LG↗

arXiv2026

FAIR-ESI: Feature Adaptive Importance Refinement for Electrophysiological Source Imaging

An essential technique for diagnosing brain disorders is electrophysiological source imaging (ESI). While model-based optimization and deep learning methods have achieved promising results in this field, the accurate selection and refinement of features remains a central challenge for precise ESI. This paper proposes FAIR-ESI, a novel framework that adaptively refines feature importance across different views, including FFT-based spectral feature refinement, weighted temporal feature refinement, and self-attention-based patch-wise feature refinement. Extensive experiments on two simulation datasets with diverse configurations and two real-world clinical datasets validate our framework's efficacy, highlighting its potential to advance brain disorder diagnosis and offer new insights into brain function.

cs.CV↗

arXiv2026

Rapid jet production and suppression during fast state transitions in the black hole X-ray binary MAXI J1348$-$630

Black hole X-ray binaries (BH XRBs) launch powerful relativistic jets during bright outburst phases. The properties of these outflows change dramatically between different spectral/accretion states. Compact jets are observed during the hard state and are quenched during the soft state, while discrete ejecta are mainly launched during the hard-to-soft state transition. Currently, we do not understand what triggers the formation/destruction of compact jets or the launch of discrete ejecta. In this context, finding a unique link between the jet evolution and the properties of the X-ray emission, such as its fast variability, would imply major progress in our understanding of the fundamental mechanisms that drive relativistic outflows in BH XRBs. Here we show that a brief, strong radio re-brightening during a predominantly soft state of the BH XRB MAXI J1348$-$630 was contemporaneous with a significant increase in the X-ray rms variability observed with NICER in 2019. During this phase, the variability displayed significant changes and, at the same time, MAXI J1348$-$630 launched two relativistic discrete ejecta that we detected with the MeerKAT and ATCA radio-interferometers. We propose that short-lived compact jets were reactivated during this excursion to the hard-intermediate state and were switched off before the ejecta launch, a behavior that has been very rarely observed in these systems. Interestingly, with the caveat of gaps in our radio and X-ray coverage, we suggest a tentative correspondence between the launch of ejecta and the drop in X-ray rms variability in this source, while other typical X-ray signatures associated with discrete ejections are not detected. We discuss how these results provide us with insights into the complex dynamic coupling between the jets and hot corona in BH XRBs.

astro-ph.HE↗

arXiv2026

X-ray and radio observations of the AMXP MAXI J1957+032 covering the 2022-2025 outbursts

We presented a comprehensive multi-epoch timing and multiwavelength analysis of the accreting millisecond X-ray pulsar MAXI J1957+032, covering two major outbursts in 2022 and 2025. By reanalyzing the 2022 outburst data from the Neutron Star Interior Composition Explorer (NICER), we found the spin frequency and orbital parameters from the observations in 0.3-5 keV. For the 2025 outburst, we reported the detection of pulsations with the Einstein Probe (EP). Based on the $\sim$3-year baseline between these two outbursts, we measured a significant long-term spin-down rate of $\dotν= (-5.73 \pm 0.28) \times 10^{-14}~{\rm Hz~s^{-1}}$. Assuming that the quiescent spin-down is driven by magnetic dipole radiation, we inferred a spin-down luminosity of $L \approx 1.1 \times 10^{36}~{\rm erg~s^{-1}}$ and a surface dipolar magnetic field of $B \approx (7.3 - 10.4) \times 10^8$ G. Furthermore, we conducted a deep radio pulsation search with the Five-hundred-meter Aperture Spherical radio Telescope (FAST) during the X-ray quiescent state in 2024, resulting in a non-detection with a 7$σ$ flux density upper limit of 12.3 $μ$Jy. This corresponds to a radio efficiency upper limit of $ξ< 2.8 \times 10^{-10}$, which is significantly lower than that of typical millisecond pulsars with a similar spin-down power. This profound radio pulsation faintness can be explained by two primary scenarios: either a geometric effect, wherein the pulsar's radio beam is directed away from our line of sight, or a physical suppression of the emission mechanism, potentially caused by a persistent low-level accretion flow during the X-ray quiescent state.

astro-ph.HE↗

arXiv2026

POLYCHARTQA: Benchmarking Large Vision-Language Models with Multilingual Chart Question Answering

Charts are a universally adopted medium for data communication, yet existing chart understanding benchmarks are overwhelmingly English-centric, limiting their accessibility and relevance to global audiences. To address this limitation, we introduce PolyChartQA, the first large-scale multilingual benchmark for chart question answering, comprising 22,606 charts and 26,151 QA pairs across 10 diverse languages. PolyChartQA is constructed through a scalable pipeline that enables efficient multilingual chart generation via data translation and code reuse, supported by LLM-based translation and rigorous quality control. We systematically evaluate multilingual chart understanding with PolyChartQA on state-of-the-art LVLMs and reveal a significant performance gap between English and other languages, particularly low-resource ones. Additionally, we introduce a companion multilingual chart question answering training set, PolyChartQA-Train, on which fine-tuning LVLMs yields substantial gains in multilingual chart understanding across diverse model sizes and architectures. Together, our benchmark provides a foundation for developing globally inclusive vision-language models capable of understanding charts across diverse linguistic contexts.

cs.CL↗

arXiv2026

Improved LLM Agents for Financial Document Question Answering

Large language models (LLMs) have shown impressive capabilities on numerous natural language processing tasks. However, LLMs still struggle with numerical question answering for financial documents that include tabular and textual data. Recent works have showed the effectiveness of critic agents (i.e., self-correction) for this task given oracle labels. Building upon this framework, this paper examines the effectiveness of the traditional critic agent when oracle labels are not available, and show, through experiments, that this critic agent's performance deteriorates in this scenario. With this in mind, we present an improved critic agent, along with the calculator agent which outperforms the previous state-of-the-art approach (program-of-thought) and is safer. Furthermore, we investigate how our agents interact with each other, and how this interaction affects their performance.

cs.CL↗

arXiv2026

Characteristics of the High-frequency Humps in the Black hole X-ray Binary Swift J1727.8--1613

We present a detailed timing analysis of the two high-frequency humps observed in the power density spectrum of Swift J1727.8--1613 up to 100 keV, using data from the Hard X-ray Modulation Telescope (Insight-HXMT). Our analysis reveals that the characteristic frequencies of the humps increase with energy up to $\sim30$ keV, followed by a plateau at higher energies. The fractional rms amplitudes of the humps increase with energy, reaching approximately 15\% in the 50-100 keV band. The lag spectrum of the hump is characterized primarily by a soft lag that varies with energy. Our results suggest that the high-frequency humps originate from a corona close to the black hole. Additionally, by applying the relativistic precession model, we constrain the mass of Swift J1727.8--1613 to $2.84 < M / M_{\odot} < 120.01$ and the spin to $0.14 < a < 0.43$ from the full-energy band dataset, using triplets composed of a type-C quasi-periodic oscillation and two high-frequency humps. When considering only the high-energy bands with stable characteristic frequencies, we derive additional constraints of $2.84 < M/M_{\odot} < 13.98$ and $0.14 < a < 0.40$.

astro-ph.HE↗

arXiv2026

High-Q AlN microresonators for nonlinear near-infrared and near-visible photonics

High Q-factors of microresonators are crucial for nonlinear integrated photonics, as many nonlinear dynamics have quadratic or even cubic dependence on Q-factors. The unique material properties make AlN microresonators invaluable for microcomb generation, Raman lasing and visible integrated photonics. However, the loss level of AlN falls behind other integrated platforms. By optimizing the fabrication, we demonstrate record Q-factors of 5.4$\times$10$^6$ and 2.2$\times$10$^6$ for AlN microresonators in the near-infrared and near-visible, respectively. Polarized-mode-interaction was used to create anomalous dispersion to support bright AlN Dirac solitons. Measurement of polarization-dependent spectra reveals the polarization hybridization of the Dirac soliton. In a microresonator with normal dispersion, Raman assisted four-wave-mixing (RFWM) was observed to initiate platicon formation, adding an approach to generate normal dispersion microcombs. A design of width-varying waveguides was used to ensure both efficient coupling and high Q-factor for racetrack microresonators at 780 nm. The microresonator was pumped to generate near-visble Raman laser at 820 nm with a fundamental linewidth narrower than 220 Hz. Our work unlocks new opportunities for integrated AlN photonics by improving Q-factors and uncovering nonlinear dynamics in AlN microresonators.

physics.optics↗

arXiv2026

Generalized $2$-split for higher-derivative YM and GR amplitudes at tree-level

We study the generalized $2$-split of higher-derivative amplitudes, including Yang-Mills (YM) and Gravity (GR) amplitudes with special insertions of higher-derivative vertices, by expanding them into ${\rm YM}\oplus{\rm BAS}$, ${\rm GR}\oplus{\rm YM}$, and ${\rm GR}\oplus{\rm YM}\oplus{\rm BAS}$ amplitude, respectively. By leveraging the established $2$-split properties of these constituent theories, we show that these higher-derivative amplitudes -- which also exhibit another newly discovered phenomenon called hidden zero -- do not factorize into a single product of two currents. Instead, their factorization universally appears as a sum of multiple $2$-split contributions.

hep-th↗

arXiv2025

MatDecompSDF: High-Fidelity 3D Shape and PBR Material Decomposition from Multi-View Images

We present MatDecompSDF, a novel framework for recovering high-fidelity 3D shapes and decomposing their physically-based material properties from multi-view images. The core challenge of inverse rendering lies in the ill-posed disentanglement of geometry, materials, and illumination from 2D observations. Our method addresses this by jointly optimizing three neural components: a neural Signed Distance Function (SDF) to represent complex geometry, a spatially-varying neural field for predicting PBR material parameters (albedo, roughness, metallic), and an MLP-based model for capturing unknown environmental lighting. The key to our approach is a physically-based differentiable rendering layer that connects these 3D properties to the input images, allowing for end-to-end optimization. We introduce a set of carefully designed physical priors and geometric regularizations, including a material smoothness loss and an Eikonal loss, to effectively constrain the problem and achieve robust decomposition. Extensive experiments on both synthetic and real-world datasets (e.g., DTU) demonstrate that MatDecompSDF surpasses state-of-the-art methods in geometric accuracy, material fidelity, and novel view synthesis. Crucially, our method produces editable and relightable assets that can be seamlessly integrated into standard graphics pipelines, validating its practical utility for digital content creation.

cs.CV↗

arXiv2025

Ask, Answer, and Detect: Role-Playing LLMs for Personality Detection with Question-Conditioned Mixture-of-Experts

Understanding human personality is crucial for web applications such as personalized recommendation and mental health assessment. Existing studies on personality detection predominantly adopt a "posts -> user vector -> labels" modeling paradigm, which encodes social media posts into user representations for predicting personality labels (e.g., MBTI labels). While recent advances in large language models (LLMs) have improved text encoding capacities, these approaches remain constrained by limited supervision signals due to label scarcity, and under-specified semantic mappings between user language and abstract psychological constructs. We address these challenges by proposing ROME, a novel framework that explicitly injects psychological knowledge into personality detection. Inspired by standardized self-assessment tests, ROME leverages LLMs' role-play capability to simulate user responses to validated psychometric questionnaires. These generated question-level answers transform free-form user posts into interpretable, questionnaire-grounded evidence linking linguistic cues to personality labels, thereby providing rich intermediate supervision to mitigate label scarcity while offering a semantic reasoning chain that guides and simplifies the text-to-personality mapping learning. A question-conditioned Mixture-of-Experts module then jointly routes over post and question representations, learning to answer questionnaire items under explicit supervision. The predicted answers are summarized into an interpretable answer vector and fused with the user representation for final prediction within a multi-task learning framework, where question answering serves as a powerful auxiliary task for personality detection. Extensive experiments on two real-world datasets demonstrate that ROME consistently outperforms state-of-the-art baselines, achieving improvements (15.41% on Kaggle dataset).

cs.CL↗

arXiv2025

$2$-Split of Form Factors via BCFW Recursion Relation

Recently, \cite{Cao:2025hio} demonstrated the $2$-split for form factor under specific kinematic constraints. This factorization is analogous to that observed in scattering amplitudes. A key consequence of this structure is the presence of hidden zeros, where the form factors vanish on specific kinematic loci. We first establish these zeros and a new zero for the form factors of the composite operators ${\cal O} =\frac{1}{2}\Tr((\partial ϕ)^2) + \Tr(ϕ^3)$ and ${\cal O} = \Tr(F^2)$, and then employ an inductive proof based on the BCFW recursion relation to prove the $2$-split factorization for any number of external particles.

hep-th↗

arXiv2025

Unilaw-R1: A Large Language Model for Legal Reasoning with Reinforcement Learning and Iterative Inference

Reasoning-focused large language models (LLMs) are rapidly evolving across various domains, yet their capabilities in handling complex legal problems remains underexplored. In this paper, we introduce Unilaw-R1, a large language model tailored for legal reasoning. With a lightweight 7-billion parameter scale, Unilaw-R1 significantly reduces deployment cost while effectively tackling three core challenges in the legal domain: insufficient legal knowledge, unreliable reasoning logic, and weak business generalization. To address these issues, we first construct Unilaw-R1-Data, a high-quality dataset containing 17K distilled and screened chain-of-thought (CoT) samples. Based on this, we adopt a two-stage training strategy combining Supervised Fine-Tuning (SFT) and Reinforcement Learning (RL), which significantly boosts the performance on complex legal reasoning tasks and supports interpretable decision-making in legal AI applications. To assess legal reasoning ability, we also introduce Unilaw-R1-Eval, a dedicated benchmark designed to evaluate models across single- and multi-choice legal tasks. Unilaw-R1 demonstrates strong results on authoritative benchmarks, outperforming all models of similar scale and achieving performance on par with the much larger DeepSeek-R1-Distill-Qwen-32B (54.9%). Following domain-specific training, it also showed significant gains on LawBench and LexEval, exceeding Qwen-2.5-7B-Instruct (46.6%) by an average margin of 6.6%.

cs.CL↗

arXiv2025

A Scene-aware Models Adaptation Scheme for Cross-scene Online Inference on Mobile Devices

Emerging Artificial Intelligence of Things (AIoT) applications desire online prediction using deep neural network (DNN) models on mobile devices. However, due to the movement of devices, unfamiliar test samples constantly appear, significantly affecting the prediction accuracy of a pre-trained DNN. In addition, unstable network connection calls for local model inference. In this paper, we propose a light-weight scheme, called Anole, to cope with the local DNN model inference on mobile devices. The core idea of Anole is to first establish an army of compact DNN models, and then adaptively select the model fitting the current test sample best for online inference. The key is to automatically identify model-friendly scenes for training scene-specific DNN models. To this end, we design a weakly-supervised scene representation learning algorithm by combining both human heuristics and feature similarity in separating scenes. Moreover, we further train a model classifier to predict the best-fit scene-specific DNN model for each test sample. We implement Anole on different types of mobile devices and conduct extensive trace-driven and real-world experiments based on unmanned aerial vehicles (UAVs). The results demonstrate that Anole outwits the method of using a versatile large DNN in terms of prediction accuracy (4.5% higher), response time (33.1% faster) and power consumption (45.1% lower).

cs.CV↗

arXiv2025

Beyond Last-Click: An Optimal Mechanism for Ad Attribution

Accurate attribution for multiple platforms is critical for evaluating performance-based advertising. However, existing attribution methods rely heavily on the heuristic methods, e.g., Last-Click Mechanism (LCM) which always allocates the attribution to the platform with the latest report, lacking theoretical guarantees for attribution accuracy. In this work, we propose a novel theoretical model for the advertising attribution problem, in which we aim to design the optimal dominant strategy incentive compatible (DSIC) mechanisms and evaluate their performance. We first show that LCM is not DSIC and performs poorly in terms of accuracy and fairness. To address this limitation, we introduce the Peer-Validated Mechanism (PVM), a DSIC mechanism in which a platform's attribution depends solely on the reports of other platforms. We then examine the accuracy of PVM across both homogeneous and heterogeneous settings, and provide provable accuracy bounds for each case. Notably, we show that PVM is the optimal DSIC mechanism in the homogeneous setting. Finally, numerical experiments are conducted to show that PVM consistently outperforms LCM in terms of attribution accuracy and fairness.

cs.GT↗