arXiv subjects

Bin Yang

Publications and source records attributed to Bin Yang.

At least 163 records · Page 9Linked to original sources

arXiv2025

$K^2$VAE: A Koopman-Kalman Enhanced Variational AutoEncoder for Probabilistic Time Series Forecasting

Probabilistic Time Series Forecasting (PTSF) plays a crucial role in decision-making across various fields, including economics, energy, and transportation. Most existing methods excell at short-term forecasting, while overlooking the hurdles of Long-term Probabilistic Time Series Forecasting (LPTSF). As the forecast horizon extends, the inherent nonlinear dynamics have a significant adverse effect on prediction accuracy, and make generative models inefficient by increasing the cost of each iteration. To overcome these limitations, we introduce $K^2$VAE, an efficient VAE-based generative model that leverages a KoopmanNet to transform nonlinear time series into a linear dynamical system, and devises a KalmanNet to refine predictions and model uncertainty in such linear system, which reduces error accumulation in long-term forecasting. Extensive experiments demonstrate that $K^2$VAE outperforms state-of-the-art methods in both short- and long-term PTSF, providing a more efficient and accurate solution.

cs.LG↗

arXiv2025

Discovery and Preliminary Characterization of a Third Interstellar Object: 3I/ATLAS

We report initial observations aimed at the characterization of a third interstellar object. This object, 3I/ATLAS or C/2025 N1 (ATLAS), was discovered on 2025 July 1 UT and has an orbital eccentricity of $e\sim6.1$, perihelion of $q\sim 1.36$ au, inclination of $\sim175^\circ$, and hyperbolic velocity of $V_\infty\sim 58$ km s$^{-1}$. We report deep stacked images obtained using the Canada-France-Hawaii Telescope and the Very Large Telescope that resolve a compact coma. Using images obtained from several smaller ground-based telescopes, we find minimal light curve variation for the object over a $\sim4$ day time span. The visible/near-infrared spectral slope of the object is 17.1$\pm$0.2 %/100 nm, comparable to other interstellar objects and primitive solar system small bodies (comets and D-type asteroids). 3I/ATLAS will be observable through early September 2025, then unobservable by Earth-based observatories near perihelion due to low solar elongation. It will be observable again from the ground in late November 2025. Although this limitation unfortunately prohibits detailed Earth-based observations at perihelion when the activity of 3I/ATLAS is likely to peak, spacecraft at Mars could be used to make valuable observations at this time.

astro-ph.EP↗

arXiv2025

Positive Style Accumulation: A Style Screening and Continuous Utilization Framework for Federated DG-ReID

The Federated Domain Generalization for Person re-identification (FedDG-ReID) aims to learn a global server model that can be effectively generalized to source and target domains through distributed source domain data. Existing methods mainly improve the diversity of samples through style transformation, which to some extent enhances the generalization performance of the model. However, we discover that not all styles contribute to the generalization performance. Therefore, we define styles that are beneficial or harmful to the model's generalization performance as positive or negative styles. Based on this, new issues arise: How to effectively screen and continuously utilize the positive styles. To solve these problems, we propose a Style Screening and Continuous Utilization (SSCU) framework. Firstly, we design a Generalization Gain-guided Dynamic Style Memory (GGDSM) for each client model to screen and accumulate generated positive styles. Meanwhile, we propose a style memory recognition loss to fully leverage the positive styles memorized by Memory. Furthermore, we propose a Collaborative Style Training (CST) strategy to make full use of positive styles. Unlike traditional learning strategies, our approach leverages both newly generated styles and the accumulated positive styles stored in memory to train client models on two distinct branches. This training strategy is designed to effectively promote the rapid acquisition of new styles by the client models, and guarantees the continuous and thorough utilization of positive styles, which is highly beneficial for the model's generalization performance. Extensive experimental results demonstrate that our method outperforms existing methods in both the source domain and the target domain.

cs.CV↗

arXiv2025

Recent Chemo-morphological Coma Evolution of Comet 67P/Churyumov-Gerasimenko

We present VLT/MUSE observations of comet 67P/Churyumov-Gerasimenko during its 2021 perihelion passage, from which we generated simultaneous maps of dust, [OI], C$_2$, NH$_2$, and CN comae across 12 pre- and post-perihelion epochs. These maps reveal the evolutionary and compositional trends of 67P's coma and further enrich the context and findings of ESA's Rosetta mission. Dust and gas species displayed distinct structures, where NH$_2$ and CN signals were uniquely associated with known dust fans, raising the question of possible correlation to the dust and contributions of extended sources. Localised fitted NH$_2$ scale lengths were 1.5-1.9$\times$ larger than those fitted for the rest of the coma, which is consistent with an extended source component for northern pre-perihelion emissions. In the southern hemisphere, CN was correlated with a prominent and sharp dust structure, potentially revealing an extended source origin via larger dust particles that preserve the CN parent species, as evidenced by higher spectral slopes in the region. Gas maps depicted two distinct evolutionary regimes: (1) evolving H$_2$O ([OI]$^{1}$D) and C$_2$ emissions driven by nucleus sublimation and subsolar insolation, and (2) stable NH$_2$ and CN emissions associated with seasonal dynamics and possible distributed sources. Dust spectral slope maps revealed spectral slope trends consistent with Rosetta findings, while green/red [OI] ratios generally indicate a coma dominated by H$_2$O.

astro-ph.EP↗

arXiv2025

TAB: Unified Benchmarking of Time Series Anomaly Detection Methods

Time series anomaly detection (TSAD) plays an important role in many domains such as finance, transportation, and healthcare. With the ongoing instrumentation of reality, more time series data will be available, leading also to growing demands for TSAD. While many TSAD methods already exist, new and better methods are still desirable. However, effective progress hinges on the availability of reliable means of evaluating new methods and comparing them with existing methods. We address deficiencies in current evaluation procedures related to datasets and experimental settings and protocols. Specifically, we propose a new time series anomaly detection benchmark, called TAB. First, TAB encompasses 29 public multivariate datasets and 1,635 univariate time series from different domains to facilitate more comprehensive evaluations on diverse datasets. Second, TAB covers a variety of TSAD methods, including Non-learning, Machine learning, Deep learning, LLM-based, and Time-series pre-trained methods. Third, TAB features a unified and automated evaluation pipeline that enables fair and easy evaluation of TSAD methods. Finally, we employ TAB to evaluate existing TSAD methods and report on the outcomes, thereby offering a deeper insight into the performance of these methods. Besides, all datasets and code are available at https://github.com/decisionintelligence/TAB.

cs.LG↗

arXiv2025

Tailoring deep learning for real-time brain-computer interfaces: From offline models to calibration-free online decoding

Despite the growing success of deep learning (DL) in offline brain-computer interfaces (BCIs), its adoption in real-time applications remains limited due to three primary challenges. First, most DL solutions are designed for offline decoding, making the transition to online decoding unclear. Second, the use of sliding windows in online decoding substantially increases computational complexity. Third, DL models typically require large amounts of training data, which are often scarce in BCI applications. To address these challenges and enable real-time, cross-subject decoding without subject-specific calibration, we introduce realtime adaptive pooling (RAP), a novel parameter-free method. RAP seamlessly modifies the pooling layers of existing offline DL models to meet online decoding requirements. It also reduces computational complexity during training by jointly decoding consecutive sliding windows. To further alleviate data requirements, our method leverages source-free domain adaptation, enabling privacy-preserving adaptation across varying amounts of target data. Our results demonstrate that RAP provides a robust and efficient framework for real-time BCI applications. It preserves privacy, reduces calibration demands, and supports co-adaptive BCI systems, paving the way for broader adoption of DL in online BCIs. These findings lay a strong foundation for developing user-centered, high-performance BCIs that facilitate immediate feedback and user learning.

cs.HC↗

arXiv2025

Numerical Analysis of the Parallel Orbital-Updating Approach for Eigenvalue Problems

The parallel orbital-updating approach is an orbital/eigenfunction iteration based approach for solving eigenvalue problems when many eigenpairs are required. It has been proven to be efficient, for instance, in electronic structure calculations. In this paper, based on the investigation of a quasi-orthogonality, we present the numerical analysis of the parallel orbital-updating approach for linear eigenvalue problems, including convergence and error estimates of the numerical approximations.

math.NA↗

arXiv2025

A portrait throughout perihelion of the NH$_2$-rich interstellar comet 2I/Borisov

The interstellar comet 2I/Borisov is the first interstellar object where compositional characterisation was possible throughout its entire perihelion passage. We report all 16 epochs of a comprehensive optical observation campaign with ESO VLT's integral field spectrograph MUSE, spanning 126 days from 2019 November 14 to 2020 March 19. The spatial dust emission of 2I/Borisov was predominantly smooth, with no seasonal effect. A jet-like feature was consistently visible. The gas production morphology of its coma was also smooth and similar for C$_2$, NH$_2$, and CN: symmetric around the photocentre. The production rates of these species gently declined into and beyond perihelion, until 2I's outburst and splitting event in early 2020 March. C$_2$, NH$_2$, and CN production rates all increased, with NH$_2$ being the most significant; the dust emission also slightly reddened. 2I/Borisov is a carbon-depleted, relatively NH$_2$-rich comet when compared to those comets yet measured in the Solar System.

astro-ph.EP↗

arXiv2025

Optical and Near-Infrared Spectroscopy of the Outbursting Comet 12P/Pons-Brooks

We present optical and near-infrared (NIR) observations of the outbursting, Halley-type comet 12P/Pons-Brooks. Three NIR spectra were obtained during two outbursts in October and November 2023, with the 3-meter Infrared Telescope Facility and the Palomar 200-inch Telescope, respectively. The NIR spectra exhibited absorption features at 1.5 and 2.0 $μ$m, consistent with the diagnostic absorption bands of water ice, superimposed on a red dust-scattering continuum. We find that the absorption bands and the red continuum can be well explained by micrometer-sized crystalline ice at 140--170 K, along with sub-micrometer-sized refractory grains (e.g., amorphous carbon). In addition, an optical spectrum was obtained with the Lijiang 2.4-meter Telescope during the November 2023 outburst, which exhibited the emission bands of gaseous CN, C$_3$, C$_2$ and NH$_2$. The C$_3$/CN and C$_2$/CN ratios suggest that 12P/Pons-Brooks was ''typical'' in C$_3$ abundance but somewhat depleted in C$_2$. The specific kinetic energy of the 2023 November outburst is estimated to be $\sim8\times10^3$ J kg$^{-1}$, suggesting a likely triggering mechanism similar to 332P/Ikeya--Murakami and 17P/Holmes, i.e., the crystallization of amorphous water ice. A refractory-to-ice ratio of $\sim$1.7--3.2 is derived from the total mass loss of dust and gas, aligning with the lower-end estimates for 67P/Churyumov-Gerasimenko and 1P/Halley. This suggests either a less evolved nucleus or an outburst region enriched in icy materials relative to the bulk nucleus.

astro-ph.EP↗

arXiv2025

Towards 3D Semantic Image Synthesis for Medical Imaging

In the medical domain, acquiring large datasets is challenging due to both accessibility issues and stringent privacy regulations. Consequently, data availability and privacy protection are major obstacles to applying machine learning in medical imaging. To address this, our study proposes the Med-LSDM (Latent Semantic Diffusion Model), which operates directly in the 3D domain and leverages de-identified semantic maps to generate synthetic data as a method of privacy preservation and data augmentation. Unlike many existing methods that focus on generating 2D slices, Med-LSDM is designed specifically for 3D semantic image synthesis, making it well-suited for applications requiring full volumetric data. Med-LSDM incorporates a guiding mechanism that controls the 3D image generation process by applying a diffusion model within the latent space of a pre-trained VQ-GAN. By operating in the compressed latent space, the model significantly reduces computational complexity while still preserving critical 3D spatial details. Our approach demonstrates strong performance in 3D semantic medical image synthesis, achieving a 3D-FID score of 0.0054 on the conditional Duke Breast dataset and similar Dice scores (0.70964) to those of real images (0.71496). These results demonstrate that the synthetic data from our model have a small domain gap with real data and are useful for data augmentation.

eess.IV↗

arXiv2025

A Large Outburst, Coma Asymmetries, and the Color of Comet 243P/NEAT

Water ice is a fundamental building material of comets and other bodies in the outer solar system. Yet, the properties of cometary water ice are challenging to study, due to its volatility and the typical distances at which comets are observed. Cometary outbursts, impulsive mass-loss events that can liberate large amounts of material, offer opportunities to directly observe and characterize cometary water ice. We present a study of comet 243P/NEAT, instigated by a $-3$ mag outburst that occurred in December 2018. Optical images and a 251-day lightcurve were examined to characterize the outburst and the comet's quiescent activity. Variations in the quiescent lightcurve appear to be dominated by coma asymmetries, rather than changing activity levels as the comet approached and receded from the Sun. Furthermore, the lightcurve shows evidence for 1 to 2 additional small outbursts ($-0.3$ mag) occurring in September 2018. The large December 2018 outburst likely ejected water ice grains, yet no signatures of ice were found in color photometry, a color map, nor a near-infrared spectrum. We discuss possible dynamical and thermal reasons for this non-detection. In this context, we examined the comae of comets 103P/Hartley 2 and C/2013 US$_{10}$ (Catalina), and show that a one-to-one mapping between continuum color and the presence of water ice cannot be supported. We also discuss possible causes for the large outburst, and find that there is an apparent grouping in the kinetic energy per mass estimates for the outbursts of 5 comets.

astro-ph.EP↗

arXiv2025

Gaussian Process Latent Variable Modeling for Few-shot Time Series Forecasting

Accurate time series forecasting is crucial for optimizing resource allocation, industrial production, and urban management, particularly with the growth of cyber-physical and IoT systems. However, limited training sample availability in fields like physics and biology poses significant challenges. Existing models struggle to capture long-term dependencies and to model diverse meta-knowledge explicitly in few-shot scenarios. To address these issues, we propose MetaGP, a meta-learning-based Gaussian process latent variable model that uses a Gaussian process kernel function to capture long-term dependencies and to maintain strong correlations in time series. We also introduce Kernel Association Search (KAS) as a novel meta-learning component to explicitly model meta-knowledge, thereby enhancing both interpretability and prediction accuracy. We study MetaGP on simulated and real-world few-shot datasets, showing that it is capable of state-of-the-art prediction accuracy. We also find that MetaGP can capture long-term dependencies and can model meta-knowledge, thereby providing valuable insights into complex time series patterns.

cs.LG↗

arXiv2025

TSFM-Bench: A Comprehensive and Unified Benchmark of Foundation Models for Time Series Forecasting

Time Series Forecasting (TSF) is key functionality in numerous fields, such as financial investment, weather services, and energy management. Although increasingly capable TSF methods occur, many of them require domain-specific data collection and model training and do not generalize well when applied in other domains. Time Series Foundation Models (TSFMs) that are pre-trained on massive heterogeneous time series data aim to overcome these limitations. The prospects for generalizability have spurred the development of a new generation of TSFMs. This study proposes a benchmark, TSFM-Bench, to facilitate comprehensive and unified evaluation of TSFMs. TSFM-Bench covers a wide range of TSFMs, including those based on large language models and those pre-trained on time series data. TSFM-Bench supports multiple forecasting scenarios, including zero-shot, few-shot, and full-shot, enabling assessment across the full range of adaptation strategies. TSFM-Bench also provides a standardized experimental protocols for critical evaluation processes such as dataset splitting, loading, normalization, and few-shot sampling, facilitating consistency and fairness. We report on an extensive evaluation of TSFMs across a diverse range of datasets spanning multiple domains and exhibiting varied statistical characteristics. Specifically, we identify pros and cons and inherent limitations of existing TSFMs, and we propose potential directions for new model designs.

cs.LG↗

arXiv2025

Stability analysis of discrete Boltzmann simulation for supersonic flows: Influencing factors, coupling mechanisms and optimization strategies

Supersonic flow simulations face challenges in trans-scale modeling, numerical stability, and complex field analysis due to inherent nonlinear, nonequilibrium, and multiscale characteristics. The discrete Boltzmann method (DBM) provides a multiscale kinetic modeling framework and analysis tool to capture complex discrete/nonequilibrium effects. While the numerical scheme plays a fundamental role in DBM simulations, a comprehensive stability analysis remains lacking. Similar to LBM, complexity mainly lies in the intrinsic coupling between velocity and spatiotemporal discretizations, compared with CFD. This study conducts von Neumann stability analysis to investigate key factors influencing DBM simulation stability, including phase-space discretization, thermodynamic nonequilibrium (TNE) levels, spatiotemporal schemes, initial conditions, and model parameters. Key findings include: (i) the moment-matching approach outperforms the expansion- and weighting-based methods in the test simulations; (ii) increased TNE enhances system nonlinearity and the intrinsic nonlinearity embedded in the model equations, amplifying instabilities; (iii) additional viscous dissipation based on distribution functions improves stability but distorts flow fields and alters constitutive relations; (iv) larger CFL numbers and relative time steps degrade stability, necessitating appropriate time-stepping strategies. To assess the stability regulation capability of DBMs across TNE levels, stability-phase diagrams and probability curves are constructed via morphological analysis within the moment-matching framework. These diagrams identify common stable parameter regions across model orders. This study reveals key factors and coupling mechanisms affecting DBM stability and proposes strategies for optimizing equilibrium distribution discretization, velocity design, and parameter selection in supersonic regimes.

physics.flu-dyn↗

arXiv2025

LightGTS: A Lightweight General Time Series Forecasting Model

Existing works on general time series forecasting build foundation models with heavy model parameters through large-scale multi-source pre-training. These models achieve superior generalization ability across various datasets at the cost of significant computational burdens and limitations in resource-constrained scenarios. This paper introduces LightGTS, a lightweight general time series forecasting model designed from the perspective of consistent periodical modeling. To handle diverse scales and intrinsic periods in multi-source pre-training, we introduce Periodical Tokenization, which extracts consistent periodic patterns across different datasets with varying scales. To better utilize the periodicity in the decoding process, we further introduce Periodical Parallel Decoding, which leverages historical tokens to improve forecasting. Based on the two techniques above which fully leverage the inductive bias of periods inherent in time series, LightGTS uses a lightweight model to achieve outstanding performance on general time series forecasting. It achieves state-of-the-art forecasting performance on 9 real-world benchmarks in both zero-shot and full-shot settings with much better efficiency compared with existing time series foundation models.

cs.LG↗

arXiv2025

Analysis of Pseudo-Labeling for Online Source-Free Universal Domain Adaptation

A domain (distribution) shift between training and test data often hinders the real-world performance of deep neural networks, necessitating unsupervised domain adaptation (UDA) to bridge this gap. Online source-free UDA has emerged as a solution for practical scenarios where access to source data is restricted and target data is received as a continuous stream. However, the open-world nature of many real-world applications additionally introduces category shifts meaning that the source and target label spaces may differ. Online source-free universal domain adaptation (SF-UniDA) addresses this challenge. Existing methods mainly rely on self-training with pseudo-labels, yet the relationship between pseudo-labeling and adaptation outcomes has not been studied yet. To bridge this gap, we conduct a systematic analysis through controlled experiments with simulated pseudo-labeling, offering valuable insights into pseudo-labeling for online SF-UniDA. Our findings reveal a substantial gap between the current state-of-the-art and the upper bound of adaptation achieved with perfect pseudo-labeling. Moreover, we show that a contrastive loss enables effective adaptation even with moderate pseudo-label accuracy, while a cross-entropy (CE) loss, though less robust to pseudo-label errors, achieves superior results when pseudo-labeling approaches perfection. Lastly, our findings indicate that pseudo-label accuracy is in general more crucial than quantity, suggesting that prioritizing fewer but high-confidence pseudo-labels is beneficial. Overall, our study highlights the critical role of pseudo-labeling in (online) SF-UniDA and provides actionable insights to drive future advancements in the field. Our code is available at https://github.com/pascalschlachter/PLAnalysis.

cs.LG↗

arXiv2025

QwenLong-CPRS: Towards $\infty$-LLMs with Dynamic Context Optimization

This technical report presents QwenLong-CPRS, a context compression framework designed for explicit long-context optimization, addressing prohibitive computation overhead during the prefill stage and the "lost in the middle" performance degradation of large language models (LLMs) during long sequence processing. Implemented through a novel dynamic context optimization mechanism, QwenLong-CPRS enables multi-granularity context compression guided by natural language instructions, achieving both efficiency gains and improved performance. Evolved from the Qwen architecture series, QwenLong-CPRS introduces four key innovations: (1) Natural language-guided dynamic optimization, (2) Bidirectional reasoning layers for enhanced boundary awareness, (3) Token critic mechanisms with language modeling heads, and (4) Window-parallel inference. Comprehensive evaluations across five benchmarks (4K-2M word contexts) demonstrate QwenLong-CPRS's threefold effectiveness: (1) Consistent superiority over other context management methods like RAG and sparse attention in both accuracy and efficiency. (2) Architecture-agnostic integration with all flagship LLMs, including GPT-4o, Gemini2.0-pro, Claude3.7-sonnet, DeepSeek-v3, and Qwen2.5-max, achieves 21.59$\times$ context compression alongside 19.15-point average performance gains; (3) Deployed with Qwen2.5-32B-Instruct, QwenLong-CPRS surpasses leading proprietary LLMs by 4.85 and 10.88 points on Ruler-128K and InfiniteBench, establishing new SOTA performance.

cs.CL↗

arXiv2025

RadarRGBD A Multi-Sensor Fusion Dataset for Perception with RGB-D and mmWave Radar

Multi-sensor fusion has significant potential in perception tasks for both indoor and outdoor environments. Especially under challenging conditions such as adverse weather and low-light environments, the combined use of millimeter-wave radar and RGB-D sensors has shown distinct advantages. However, existing multi-sensor datasets in the fields of autonomous driving and robotics often lack high-quality millimeter-wave radar data. To address this gap, we present a new multi-sensor dataset:RadarRGBD. This dataset includes RGB-D data, millimeter-wave radar point clouds, and raw radar matrices, covering various indoor and outdoor scenes, as well as low-light environments. Compared to existing datasets, RadarRGBD employs higher-resolution millimeter-wave radar and provides raw data, offering a new research foundation for the fusion of millimeter-wave radar and visual sensors. Furthermore, to tackle the noise and gaps in depth maps captured by Kinect V2 due to occlusions and mismatches, we fine-tune an open-source relative depth estimation framework, incorporating the absolute depth information from the dataset for depth supervision. We also introduce pseudo-relative depth scale information to further optimize the global depth scale estimation. Experimental results demonstrate that the proposed method effectively fills in missing regions in sensor data. Our dataset and related documentation will be publicly available at: https://github.com/song4399/RadarRGBD.

eess.IV↗