arXiv subjects

Yilong Wang

Publications and source records attributed to Yilong Wang.

At least 37 records · Page 2Linked to original sources

arXiv2025

Accretion geometry in neutron star low-mass X-ray binaries during the hard spectral state

We investigate the accretion geometry in neutron star low-mass X-ray binaries (LMXBs) in the hard spectral state. It is commonly accepted that, for low mass transfer rates, an advection-dominated accretion flow (ADAF) is present in the inner region. But the observed relativistically broadened emission lines in the reflection spectra clearly indicate the existence of discs near the innermost stable circular orbit $(R_{\rm{ISCO}})$. We investigate the interaction between the coronal flow and the disc in neutron star LMXBs, and find that gas condensation from the dominant, coronal accretion flow to an inner disc is enhanced as compared to that in black hole LMXBs as a consequence of irradiation of the corona by the neutron star surface. Computations show that for low mass transfer rates ($\sim 0.005-0.02$ Eddington rate) a persistent weak disc can coexist with a coronal flow in the innermost region, where a pure ADAF would have been expected. The inner disc extends outwards from $R_{\rm{ISCO}}$ to $\sim 10 R_{\rm{ISCO}}$ for Eddington ratios ($L/L_{\rm{Edd}}$) as low as $\sim 0.002$, covers a larger region for higher Eddington ratios, and eventually connects to the outer disc at $L/L_{\rm{Edd}} \sim 0.02$, thereby transiting to a soft state. We demonstrate that the observationally inferred region of the broad iron lines in the hard-state sources generally lies within the extension of the inner discs predicted by the condensation model. Disappearance of the broad iron lines is predicted at very low luminosities, either caused by very low accretion rates or disc truncation by strong magnetic fields.

astro-ph.HE↗

arXiv2025

Unlearning Inversion Attacks for Graph Neural Networks

Graph unlearning methods aim to efficiently remove the impact of sensitive data from trained GNNs without full retraining, assuming that deleted information cannot be recovered. In this work, we challenge this assumption by introducing the graph unlearning inversion attack: given only black-box access to an unlearned GNN and partial graph knowledge, can an adversary reconstruct the removed edges? We identify two key challenges: varying probability-similarity thresholds for unlearned versus retained edges, and the difficulty of locating unlearned edge endpoints, and address them with TrendAttack. First, we derive and exploit the confidence pitfall, a theoretical and empirical pattern showing that nodes adjacent to unlearned edges exhibit a large drop in model confidence. Second, we design an adaptive prediction mechanism that applies different similarity thresholds to unlearned and other membership edges. Our framework flexibly integrates existing membership inference techniques and extends them with trend features. Experiments on four real-world datasets demonstrate that TrendAttack significantly outperforms state-of-the-art GNN membership inference baselines, exposing a critical privacy vulnerability in current graph unlearning methods.

cs.LG↗

arXiv2025

Generalized Witt and Morita equivalences

In this work, we introduce a family of new equivalence relations among fusion categories that are less refined than the usual Morita equivalence. We obtain abelian groups by quotienting these new equivalence relations from the commutative monoids of the equivalence classes of all fusion categories. Moreover, we upgrade them to equivalence relations among nondegenerate braided fusion categories that are more refined than the usual Witt equivalence. As a consequence, we obtain new abelian groups that are more refined than the usual Witt group. These new groups allow us to access the internal structures within Witt classes. We expect that they are useful in the classification program of (braided) fusion (higher) categories and in the study of gapless edges of 2+1D topological orders.

math.QA↗

arXiv2025

Every Step Evolves: Scaling Reinforcement Learning for Trillion-Scale Thinking Model

We present Ring-1T, the first open-source, state-of-the-art thinking model with a trillion-scale parameter. It features 1 trillion total parameters and activates approximately 50 billion per token. Training such models at a trillion-parameter scale introduces unprecedented challenges, including train-inference misalignment, inefficiencies in rollout processing, and bottlenecks in the RL system. To address these, we pioneer three interconnected innovations: (1) IcePop stabilizes RL training via token-level discrepancy masking and clipping, resolving instability from training-inference mismatches; (2) C3PO++ improves resource utilization for long rollouts under a token budget by dynamically partitioning them, thereby obtaining high time efficiency; and (3) ASystem, a high-performance RL framework designed to overcome the systemic bottlenecks that impede trillion-parameter model training. Ring-1T delivers breakthrough results across critical benchmarks: 93.4 on AIME-2025, 86.72 on HMMT-2025, 2088 on CodeForces, and 55.94 on ARC-AGI-1. Notably, it attains a silver medal-level result on the IMO-2025, underscoring its exceptional reasoning capabilities. By releasing the complete 1T parameter MoE model to the community, we provide the research community with direct access to cutting-edge reasoning capabilities. This contribution marks a significant milestone in democratizing large-scale reasoning intelligence and establishes a new baseline for open-source model performance.

cs.CL↗

arXiv2025

Accretion with two-phase gas supply and its application in black hole X-ray binaries

Accretion in black hole X-ray binaries is commonly believed to be supplied by the Roche lobe overflow or the stellar wind. The former is thought to form a geometrically thin disc while the diffuse wind could form a geometrically thick hot accretion flow. In this paper, we instead consider a more generalised case, i.e., accretion with both cold and hot gas supplies, which feed a disc and a corona respectively. We investigate the interaction of disc and corona by analysing the energy coupling and matter exchange, i.e. corona condensation/disc evaporation, with a semi-analytical method. It is found that the accretion geometry in the radial direction and the resultant emission spectrum depend strongly on both the total gas supply rate and the ratio of cold and hot gases. For gas supply rates of a few percent of the Eddington value, diverse geometries and spectral shapes are possible, depending on the fraction of cold gas supply. This provides an interpretation for the various spectra observed in intermediate states. However, at higher accretion rates, regardless of the form of the feeding gas, the inner accretion flow is always disc-dominated, implying an inevitable transition to the soft state, while at very low gas supply rates, hard state spectrum dominated by the hot flow is expected. We also present the predicted hardness-intensity correlation of Cygnus X-1, and constrain the value of the viscosity parameter of the accretion flow to the range of 0.25--0.35 by comparing our results with MAXI observations.

astro-ph.HE↗

arXiv2025

Harmonic balance-automatic differentiation method: an out-of-the-box and efficient solver for general nonlinear dynamics simulation

The Harmonic Balance-Alternating Frequency-Time domain (HB-AFT) method is extensively employed for dynamic response analysis of nonlinear systems. However, its application to high-dimensional complex systems is constrained by the manual derivation of Jacobian matrices during Newton-Raphson iterations, which become computationally intractable or error-prone for intricate nonlinearities. The Harmonic Balance-Automatic Differentiation (HB-AD) method is proposed to address this limitation, in which AD is integrated with the harmonic balance framework. This approach eliminates all manual derivations by leveraging AD to compute exact Jacobians numerically, enabling generic and efficient analysis of high-dimensional complex nonlinear systems. The implementation utilizes advanced deep learning frameworks for native parallel computing and CUDA acceleration, and combines AD with arc-length continuation, establishing an out-of-the-box and high efficiency computational architecture. Users need only supply the system's dynamic equations, HB-AD then autonomously trace the complete panorama of periodic responses -- including stable/unstable solution branches. Computational experiments on a rotor system with squeeze-film damper (SFD) demonstrate HB-AD's capability in handling complex nonlinear expressions with automated Jacobian calculations. For a high-dimensional aero-engine rotor-bearing-casing system with complex bearing nonlinearities, HB-AD achieves 17-fold higher efficiency than traditional HB-AFT and 144-fold acceleration over the Newmark method. The HB-AD method is a synergistic merger of computational mechanics and machine learning primitives, delivers an easy to use, general-purpose, high efficiency platform for high-fidelity dynamic characterization of high-dimensional engineering systems.

cs.CE↗

arXiv2025

ViPSN 2.0: A Reconfigurable Battery-free IoT Platform for Vibration Energy Harvesting

Vibration energy harvesting is a promising solution for powering battery-free IoT systems; however, the instability of ambient vibrations presents significant challenges, such as limited harvested energy, intermittent power supply, and poor adaptability to various applications. To address these challenges, this paper proposes ViPSN2.0, a modular and reconfigurable IoT platform that supports multiple vibration energy harvesters (piezoelectric, electromagnetic, and triboelectric) and accommodates sensing tasks with varying application requirements through standardized hot-swappable interfaces. ViPSN~2.0 incorporates an energy-indication power management framework tailored to various application demands, including light-duty discrete sampling, heavy-duty high-power sensing, and complex-duty streaming tasks, thereby effectively managing fluctuating energy availability. The platform's versatility and robustness are validated through three representative applications: ViPSN-Beacon, enabling ultra-low-power wireless beacon transmission from a single transient fingertip press; ViPSN-LoRa, supporting high-power, long-range wireless communication powered by wave vibrations in actual marine environments; and ViPSN-Cam, enabling intermittent image capture and wireless transfer. Experimental results demonstrate that ViPSN~2.0 can reliably meet a wide range of requirements in practical battery-free IoT deployments under energy-constrained conditions.

cs.AR↗

arXiv2025

The effectively optically thin accretion flow and its implication in supermassive black holes

Based on a unified description of various accretion flows, we find a long-ignored solution - the effectively optically thin accretion flow, occurring at accretion rates around Eddington value. As a consequence of radiation-pressure dominance, the density in a standard thin disc (SSD) decreases with the increase of accretion rates, making the innermost region effectively optically thin. Further increase in accretion rate leads to a rise of the temperature so that the Compton cooling is able to balance the accretion released energy. We demonstrate that the effectively optically thin flow is characterized by moderate temperature and large scattering optical depth, producing a multi-color Wien spectrum. For an appropriate accretion rate, the accretion flow transforms from an outer SSD to an inner effectively optically thin flow. Thus, the spectra of the whole accretion flow exhibit two components, i.e., a multi-color Wien spectrum at higher frequency and a multi-color blackbody, the former could provide an alternative origin of soft X-ray excess or formation of warm corona in active galactic nuclei (AGNs). Our stability analysis proves it is thermally stable and viscously unstable, indicating its existence in accreting systems. We show that effectively optically thin accretion flow exists in supermassive black holes for accretion rates around 0.1 to 10 times Eddington value, bridging the SSD at low accretion rates and slim disc at high rates. By comparing the predictions and average spectra of AGN, we constrain the viscosity parameter to be $α\sim 0.03$, in good agreement with that derived from observed variability.

astro-ph.HE↗

arXiv2025

Generalizations of Frobenius-Schur indicators from Kuperberg invariants

We introduce an approach to produce gauge invariants of any finite-dimensional Hopf algebras from the Kuperberg invariants of framed 3-manifolds. These invariants are generalizations of Frobenius-Schur indicators of Hopf algebras. The computation of Kuperberg invariants is based on a presentation of the framed 3-manifold in terms of Heegaard diagram with combings satisfying certain admissibility conditions. We provide framed Heegaard diagrams for two infinite families of small genus 3-manifolds, which include all the lens spaces, and some homology spheres. In particular, the invariants of the lens spaces $L(n,1)$ coincide with the higher Frobenius-Schur indicators of Hopf algebras. We compute the Kuperberg invariants of all these framed 3-manifolds, and prove that they are invariants of the tensor category of representations of the underlying Hopf algebra, or simply gauge invariants.

math.QA↗

arXiv2025

Are You Using Reliable Graph Prompts? Trojan Prompt Attacks on Graph Neural Networks

Graph Prompt Learning (GPL) has been introduced as a promising approach that uses prompts to adapt pre-trained GNN models to specific downstream tasks without requiring fine-tuning of the entire model. Despite the advantages of GPL, little attention has been given to its vulnerability to backdoor attacks, where an adversary can manipulate the model's behavior by embedding hidden triggers. Existing graph backdoor attacks rely on modifying model parameters during training, but this approach is impractical in GPL as GNN encoder parameters are frozen after pre-training. Moreover, downstream users may fine-tune their own task models on clean datasets, further complicating the attack. In this paper, we propose TGPA, a backdoor attack framework designed specifically for GPL. TGPA injects backdoors into graph prompts without modifying pre-trained GNN encoders and ensures high attack success rates and clean accuracy. To address the challenge of model fine-tuning by users, we introduce a finetuning-resistant poisoning approach that maintains the effectiveness of the backdoor even after downstream model adjustments. Extensive experiments on multiple datasets under various settings demonstrate the effectiveness of TGPA in compromising GPL models with fixed GNN encoders.

cs.LG↗

arXiv2025

Bridging Source and Target Domains via Link Prediction for Unsupervised Domain Adaptation on Graphs

Graph neural networks (GNNs) have shown great ability for node classification on graphs. However, the success of GNNs relies on abundant labeled data, while obtaining high-quality labels is costly and challenging, especially for newly emerging domains. Hence, unsupervised domain adaptation (UDA), which trains a classifier on the labeled source graph and adapts it to the unlabeled target graph, is attracting increasing attention. Various approaches have been proposed to alleviate the distribution shift between the source and target graphs to facilitate the classifier adaptation. However, most of them simply adopt existing UDA techniques developed for independent and identically distributed data to gain domain-invariant node embeddings for graphs, which do not fully consider the graph structure and message-passing mechanism of GNNs during the adaptation and will fail when label distribution shift exists among domains. In this paper, we proposed a novel framework that adopts link prediction to connect nodes between source and target graphs, which can facilitate message-passing between the source and target graphs and augment the target nodes to have ``in-distribution'' neighborhoods with the source domain. This strategy modified the target graph on the input level to reduce its deviation from the source domain in the embedding space and is insensitive to disproportional label distributions across domains. To prevent the loss of discriminative information in the target graph, we further design a novel identity-preserving learning objective, which guides the learning of the edge insertion module together with reconstruction and adaptation losses. Experimental results on real-world datasets demonstrate the effectiveness of our framework.

cs.LG↗

arXiv2025

A generalized solution for advection-dominated accretion flow, standard disc, and slim disc

Aiming at a general description of four basic solutions describing the accretion processes, i.e., the Shakura-Sunyaev thin disc (SSD), the Shapiro-Lightman-Eardley solution (SLE), the slim disc, and the advection-dominated accretion flow (ADAF), we present generalized axisymmetric height-averaged equations, where the entropy advection, the radiation pressure, and photon trapping effect are all included self-consistently. Our generalized solution can reproduce the ADAF, SLE, SSD, and slim disc branches in a wide range of accretion rates from sub- to super-Eddington accretion. An S-curve in the $\dot{m}-Σ$ plane is also reproduced, representing the SSD branch, the radiation-pressure-dominant branch, and the slim disc branch. The solution gives a natural transition between SSD and slim disc when photon trapping occurs in the accretion flow, producing a radially hybrid SSD-slim disc structure in a certain range of accretion rates. The coexistence of ADAF and SSD below a critical accretion rate is clearly shown with distinct advection fraction of accretion energy. We also present the luminosity, radiation efficiency, and spectrum from the generalized solutions for a large range of accretion rates in stellar mass black holes.

astro-ph.HE↗

arXiv2025

Low latency global carbon budget reveals a continuous decline of the land carbon sink during the 2023/24 El Nino event

The high growth rate of atmospheric CO2 in 2023 was found to be caused by a severe reduction of the global net land carbon sink. Here we update the global CO2 budget from January 1st to July 1st 2024, during which El Niño drought conditions continued to prevail in the Tropics but ceased by March 2024. We used three dynamic global vegetation models (DGVMs), machine learning emulators of ocean models, three atmospheric inversions driven by observations from the second Orbiting Carbon Observatory (OCO-2) satellite, and near-real-time fossil CO2 emissions estimates. In a one-year period from July 2023 to July 2024 covering the El Niño 2023/24 event, we found a record-high CO2 growth rate of 3.66~$\pm$~0.09 ppm~yr$^{-1}$ ($\pm$~1 standard deviation) since 1979. Yet, the CO2 growth rate anomaly obtained after removing the long term trend is 1.1 ppm~yr$^{-1}$, which is marginally smaller than the July--July growth rate anomalies of the two major previous El Niño events in 1997/98 and 2015/16. The atmospheric CO2 growth rate anomaly was primarily driven by a 2.24 GtC~yr$^{-1}$ reduction in the net land sink including 0.3 GtC~yr$^{-1}$ of fire emissions, partly offset by a 0.38 GtC~yr$^{-1}$ increase in the ocean sink relative to the 2015--2022 July--July mean. The tropics accounted for 97.5\% of the land CO2 flux anomaly, led by the Amazon (50.6\%), central Africa (34\%), and Southeast Asia (8.2\%), with extra-tropical sources in South Africa and southern Brazil during April--July 2024. Our three DGVMs suggest greater tropical CO2 losses in 2023/2024 than during the two previous large El Niño in 1997/98 and 2015/16, whereas inversions indicate losses more comparable to 2015/16. Overall, this update of the low latency budget highlights the impact of recent El Niño droughts in explaining the high CO2 growth rate until July 2024.

physics.ao-ph↗

arXiv2025

TAMT: Temporal-Aware Model Tuning for Cross-Domain Few-Shot Action Recognition

Going beyond few-shot action recognition (FSAR), cross-domain FSAR (CDFSAR) has attracted recent research interests by solving the domain gap lying in source-to-target transfer learning. Existing CDFSAR methods mainly focus on joint training of source and target data to mitigate the side effect of domain gap. However, such kind of methods suffer from two limitations: First, pair-wise joint training requires retraining deep models in case of one source data and multiple target ones, which incurs heavy computation cost, especially for large source and small target data. Second, pre-trained models after joint training are adopted to target domain in a straightforward manner, hardly taking full potential of pre-trained models and then limiting recognition performance. To overcome above limitations, this paper proposes a simple yet effective baseline, namely Temporal-Aware Model Tuning (TAMT) for CDFSAR. Specifically, our TAMT involves a decoupled paradigm by performing pre-training on source data and fine-tuning target data, which avoids retraining for multiple target data with single source. To effectively and efficiently explore the potential of pre-trained models in transferring to target domain, our TAMT proposes a Hierarchical Temporal Tuning Network (HTTN), whose core involves local temporal-aware adapters (TAA) and a global temporal-aware moment tuning (GTMT). Particularly, TAA learns few parameters to recalibrate the intermediate features of frozen pre-trained models, enabling efficient adaptation to target domains. Furthermore, GTMT helps to generate powerful video representations, improving match performance on the target domain. Experiments on several widely used video benchmarks show our TAMT outperforms the recently proposed counterparts by 13%$\sim$31%, achieving new state-of-the-art CDFSAR results.

cs.CV↗

arXiv2025

Enhance GNNs with Reliable Confidence Estimation via Adversarial Calibration Learning

Despite their impressive predictive performance, GNNs often exhibit poor confidence calibration, i.e., their predicted confidence scores do not accurately reflect true correctness likelihood. This issue raises concerns about their reliability in high-stakes domains such as fraud detection, and risk assessment, where well-calibrated predictions are essential for decision-making. To ensure trustworthy predictions, several GNN calibration methods are proposed. Though they can improve global calibration, our experiments reveal that they often fail to generalize across different node groups, leading to inaccurate confidence in node groups with different degree levels, classes, and local structures. In certain cases, they even degrade calibration compared to the original uncalibrated GNN. To address this challenge, we propose a novel AdvCali framework that adaptively enhances calibration across different node groups. Our method leverages adversarial training to automatically identify mis-calibrated node groups and applies a differentiable Group Expected Calibration Error (ECE) loss term to refine confidence estimation within these groups. This allows the model to dynamically adjust its calibration strategy without relying on dataset-specific prior knowledge about miscalibrated subgroups. Extensive experiments on real-world datasets demonstrate that our approach not only improves global calibration but also significantly enhances calibration within groups defined by feature similarity, topology, and connectivity, outperforming previous methods and demonstrating its effectiveness in practical scenarios.

cs.LG↗

arXiv2025

Every FLOP Counts: Scaling a 300B Mixture-of-Experts LING LLM without Premium GPUs

In this technical report, we tackle the challenges of training large-scale Mixture of Experts (MoE) models, focusing on overcoming cost inefficiency and resource limitations prevalent in such systems. To address these issues, we present two differently sized MoE large language models (LLMs), namely Ling-Lite and Ling-Plus (referred to as "Bailing" in Chinese, spelled Bǎilíng in Pinyin). Ling-Lite contains 16.8 billion parameters with 2.75 billion activated parameters, while Ling-Plus boasts 290 billion parameters with 28.8 billion activated parameters. Both models exhibit comparable performance to leading industry benchmarks. This report offers actionable insights to improve the efficiency and accessibility of AI development in resource-constrained settings, promoting more scalable and sustainable technologies. Specifically, to reduce training costs for large-scale MoE models, we propose innovative methods for (1) optimization of model architecture and training processes, (2) refinement of training anomaly handling, and (3) enhancement of model evaluation efficiency. Additionally, leveraging high-quality data generated from knowledge graphs, our models demonstrate superior capabilities in tool use compared to other models. Ultimately, our experimental findings demonstrate that a 300B MoE LLM can be effectively trained on lower-performance devices while achieving comparable performance to models of a similar scale, including dense and MoE models. Compared to high-performance devices, utilizing a lower-specification hardware system during the pre-training phase demonstrates significant cost savings, reducing computing costs by approximately 20%. The models can be accessed at https://huggingface.co/inclusionAI.

cs.LG↗

arXiv2025

One-Shot Dual-Arm Imitation Learning

We introduce One-Shot Dual-Arm Imitation Learning (ODIL), which enables dual-arm robots to learn precise and coordinated everyday tasks from just a single demonstration of the task. ODIL uses a new three-stage visual servoing (3-VS) method for precise alignment between the end-effector and target object, after which replay of the demonstration trajectory is sufficient to perform the task. This is achieved without requiring prior task or object knowledge, or additional data collection and training following the single demonstration. Furthermore, we propose a new dual-arm coordination paradigm for learning dual-arm tasks from a single demonstration. ODIL was tested on a real-world dual-arm robot, demonstrating state-of-the-art performance across six precise and coordinated tasks in both 4-DoF and 6-DoF settings, and showing robustness in the presence of distractor objects and partial occlusions. Videos are available at: https://www.robot-learning.uk/one-shot-dual-arm.

cs.RO↗

arXiv2025

LanP: Rethinking the Impact of Language Priors in Large Vision-Language Models

Large Vision-Language Models (LVLMs) have shown impressive performance in various tasks. However, LVLMs suffer from hallucination, which hinders their adoption in the real world. Existing studies emphasized that the strong language priors of LVLMs can overpower visual information, causing hallucinations. However, the positive role of language priors is the key to a powerful LVLM. If the language priors are too weak, LVLMs will struggle to leverage rich parameter knowledge and instruction understanding abilities to complete tasks in challenging visual scenarios where visual information alone is insufficient. Therefore, we propose a benchmark called LanP to rethink the impact of Language Priors in LVLMs. It is designed to investigate how strong language priors are in current LVLMs. LanP consists of 170 images and 340 corresponding well-designed questions. Extensive experiments on 25 popular LVLMs reveal that many LVLMs' language priors are not strong enough to effectively aid question answering when objects are partially hidden. Many models, including GPT-4 Turbo, exhibit an accuracy below 0.5 in such a scenario.

cs.CV↗