arXiv subjects

Ziang Yan

Publications and source records attributed to Ziang Yan.

At least 37 records · Page 2Linked to original sources

arXiv2026

KiDS-Legacy: WIMP dark matter constraints from the cross-correlation of weak lensing and Fermi-LAT gamma rays

Dark matter dominates the matter content of the Universe, and its properties can be constrained through large-scale structure probes such as the cross-correlation between the unresolved gamma-ray background (UGRB) and weak gravitational lensing. We analysed 15 years of Fermi-LAT data, constructing UGRB intensity maps in ten energy bins (0.5-1000 GeV), and cross-correlated them with KiDS-Legacy shear in six tomographic bins. The measurements were performed using angular power spectra estimated with the pseudo-$C_\ell$ method. No significant cross-correlation is found. Based on this non-detection, we present 95% upper bounds on the weakly interacting massive particle (WIMP) decay rate $Γ_{\rm dec}$ and velocity-averaged annihilation cross-section $\langleσ_{\rm ann} v\rangle$ as functions of mass. We compare our results with bounds from other cosmological tracers and from local probes, and found them to be complementary, particularly at low masses ($\rm GeV/TeV$). In addition, using a Euclid-like lensing survey cross-correlated with Fermi-LAT, we forecast $\sim$2 times tighter limits, highlighting the potential of forthcoming data to strengthen constraints on dark matter annihilation and decay.

astro-ph.CO↗

arXiv2026

Diagnosing the Effects of Spectroscopic Training Set Imperfection on Photometric Redshift Performance

Most LSST extragalactic science will rely on photometric redshifts (photo-$z$) to extract distance information for the galaxies. However, an incomplete or non-representative training set can introduce bias into photo-$z$ estimation. It is necessary to understand how various forms of training set imperfection, such as incompleteness and non-trivial spectroscopic target selection, affect photo-$z$ estimation algorithms, and to identify metrics best-suited to quantify the impact. This work aims to systematically study metrics for diagnosing how various photo-$z$ methods react to certain types of training set incompleteness and non-representativeness. We use methods available through the open-source Python library Redshift Assessment Infrastructure Layers (RAIL) to systematically test the algorithms CMNN, GPz, FlexZBoost, and PZFlow on mock training data degraded in accordance with several existing spectroscopic sky surveys, as well as under conditions of inverse redshift incompleteness, which approximately mimics observed patterns of incompleteness at high redshift. We employ the algorithm TrainZ as a control. Finally, we quantify photo-$z$ algorithm performance using a variety of statistical metrics implemented externally to RAIL. We determine that the Kullback-Liebler Divergence, Wasserstein Distance, and Probability Integral Transform are particularly informative metrics with which to assess the impact of training set imperfection on algorithmic performance. We also find that inverse redshift incompleteness effects alone lack the complexity to realistically represent anticipated training data.

astro-ph.IM↗

arXiv2025

VideoChat-R1: Enhancing Spatio-Temporal Perception via Reinforcement Fine-Tuning

Reinforcement Learning (RL) benefits Large Language Models (LLMs) for complex reasoning. Inspired by this, we explore integrating spatio-temporal specific rewards into Multimodal Large Language Models (MLLMs) to address the unique challenges of video understanding, such as long-range temporal associations. This paper investigates how rule-based rewards, particularly temporal ones, can improve video reasoning and their generalizability. Our study proposes Reinforcement Fine-Tuning (RFT) as a data-efficient method to enhance video reasoning on specific tasks without sacrificing original capabilities. Through joint RFT on multiple spatio-temporal perception tasks, we developed VideoChat-R1, a powerful Video MLLM. VideoChat-R1 achieves state-of-the-art spatio-temporal perception, demonstrating significant improvements in tasks like temporal grounding (+31.8) and object tracking (+31.2), while also improving general QA benchmarks. The enhanced perception and preserved chat abilities contribute to a more reliable video dialogue system, leading to our ``Temporal Clue-driven Reasoning" inference schema. This work provides a foundation for developing robust, real-world video comprehension agents.

cs.CV↗

arXiv2025

VideoTG-R1: Boosting Video Temporal Grounding via Curriculum Reinforcement Learning on Reflected Boundary Annotations

Video temporal grounding (VTG) aims to locate precise segments in videos based on language queries, which is a fundamental challenge in video understanding. While recent Multimodal Large Language Models (MLLMs) have shown promise in tackling VTG through reinforcement learning (RL), they overlook the challenges arising from both the quality and difficulty of training samples. (1) Partially annotated samples. Many samples contain relevant segments beyond the annotated interval, introducing ambiguous supervision. (2) Hard-to-ground samples. Samples with poor zero-shot performance produce consistently low and indistinguishable rewards during RL training, exhibiting no clear preference among multiple outputs and thus hindering learning efficiency. To address these challenges, we propose VideoTG-R1, a novel curriculum RL framework with reflected boundary annotations, enabling data-efficient training. Specifically, we propose a Boundary Reflection Agent that utilizes MLLMs to predict query-relevant timestamps outside the annotated intervals, allowing us to identify and filter out partially annotated samples, thereby reducing ambiguity. Furthermore, we introduce a Difficulty Estimation Agent to assess the training difficulty of each sample and design a curriculum RL strategy that dynamically masks the videos of hard-to-ground samples according to the training steps, easing the training difficulty and providing clearer preference. Experiments on the VTG and grounded VideoQA tasks demonstrate the effectiveness of our method. Remarkably, with only 10% of the training samples and 21% of the computational budget, VideoTG-R1 outperforms full-data counterparts under both group relative policy optimization (GRPO) and supervised fine-tuning (SFT). The code is available at https://github.com/ldong1111/VideoTG-R1.

cs.CV↗

arXiv2025

DSI-Bench: A Benchmark for Dynamic Spatial Intelligence

Reasoning about dynamic spatial relationships is essential, as both observers and objects often move simultaneously. Although vision-language models (VLMs) and visual expertise models excel in 2D tasks and static scenarios, their ability to fully understand dynamic 3D scenarios remains limited. We introduce Dynamic Spatial Intelligence and propose DSI-Bench, a benchmark with nearly 1,000 dynamic videos and over 1,700 manually annotated questions covering nine decoupled motion patterns of observers and objects. Spatially and temporally symmetric designs reduce biases and enable systematic evaluation of models' reasoning about self-motion and object motion. Our evaluation of 14 VLMs and expert models reveals key limitations: models often conflate observer and object motion, exhibit semantic biases, and fail to accurately infer relative relationships in dynamic scenarios. Our DSI-Bench provides valuable findings and insights about the future development of general and expertise models with dynamic spatial intelligence.

cs.CV↗

arXiv2025

KiDS-Legacy: Cosmological constraints from cosmic shear with the complete Kilo-Degree Survey

We present cosmic shear constraints from the completed Kilo-Degree Survey (KiDS), where the cosmological parameter $S_8\equivσ_8\sqrt{Ω_{\rm m}/0.3} = 0.815^{+0.016}_{-0.021}$, is found to be in agreement ($0.73σ$) with results from the Planck Legacy cosmic microwave background experiment. The final KiDS footprint spans $1347$ square degrees of deep nine-band imaging across the optical and near-infrared, along with an extra $23$ square degrees of KiDS-like calibration observations of deep spectroscopic surveys. Improvements in our redshift distribution estimation methodology, combined with our enhanced calibration data and multi-band image simulations, allow us to extend our lensed sample out to a photometric redshift of $z_{\rm B}\leq2.0$. Compared to previous KiDS analyses, the increased survey area and redshift depth results in a $\sim32\%$ improvement in constraining power in terms of $Σ_8\equivσ_8\left(Ω_{\rm m}/0.3\right)^α= 0.821^{+0.014}_{-0.016}$, where $α= 0.58$ has been optimised to match the revised degeneracy direction of $σ_8$ and $Ω_{\rm m}$. We adopt a new physically motivated intrinsic alignment model that depends jointly on the galaxy sample's halo mass and spectral type distributions, and that is informed by previous direct alignment measurements. We also marginalise over our uncertainty on the impact of baryon feedback on the non-linear matter power spectrum. Comparing to previous KiDS analyses, we conclude that the increase seen in $S_8$ primarily results from our improved redshift distribution estimation and calibration, as well as new survey area and improved image reduction. Our companion paper Stölzner et al. (submitted) presents a full suite of internal and external consistency tests, finding the KiDS-Legacy data set to be the most internally robust sample produced by KiDS to date.

astro-ph.CO↗

arXiv2025

KiDS-Legacy: Redshift distributions and their calibration

We present the redshift calibration methodology and bias estimates for the cosmic shear analysis of the fifth and final data release (DR5) of the Kilo-Degree Survey (KiDS). KiDS-DR5 includes a greatly expanded compilation of calibrating spectra, drawn from $27$ square degrees of dedicated optical and near-IR imaging taken over deep spectroscopic fields. The redshift distribution calibration leverages a range of new methods and updated simulations to produce the most precise $N(z)$ bias estimates used by KiDS to date. Improvements to our colour-based redshift distribution measurement method (SOM) mean that we are able to use many more sources per tomographic bin for our cosmological analyses, and better estimate the representation of our source sample given the available spec-$z$. We validate our colour-based redshift distribution estimates with spectroscopic cross-correlations (CC). We find that improvements to our cross-correlation redshift distribution measurement methods mean that redshift distribution biases estimated between the SOM and CC methods are fully consistent on simulations, and the data calibration is consistent to better than $2σ$ in all tomographic bins.

astro-ph.CO↗

arXiv2025

KiDS-Legacy: Consistency of cosmic shear measurements and joint cosmological constraints with external probes

We present a cosmic shear consistency analysis of the final data release from the Kilo-Degree Survey (KiDS-Legacy). By adopting three tiers of consistency metrics, we compare cosmological constraints between subsets of the KiDS-Legacy dataset split by redshift, angular scale, galaxy colour and spatial region. We also review a range of two-point cosmic shear statistics. With the data passing all our consistency metric tests, we demonstrate that KiDS-Legacy is the most internally consistent KiDS catalogue to date. In a joint cosmological analysis of KiDS-Legacy and DES Y3 cosmic shear, combined with data from the Pantheon+ Type Ia supernovae compilation and baryon acoustic oscillations from DESI Y1, we find constraints consistent with Planck measurements of the cosmic microwave background with $S_8\equiv σ_8\sqrt{Ω_{\rm m}/0.3} = 0.814^{+0.011}_{-0.012}$ and $σ_8 = 0.802^{+0.022}_{-0.018}$.

astro-ph.CO↗

arXiv2025

ExpVid: A Benchmark for Experiment Video Understanding & Reasoning

Multimodal Large Language Models (MLLMs) hold promise for accelerating scientific discovery by interpreting complex experimental procedures. However, their true capabilities are poorly understood, as existing benchmarks neglect the fine-grained and long-horizon nature of authentic laboratory work, especially in wet-lab settings. To bridge this gap, we introduce ExpVid, the first benchmark designed to systematically evaluate MLLMs on scientific experiment videos. Curated from peer-reviewed video publications, ExpVid features a new three-level task hierarchy that mirrors the scientific process: (1) Fine-grained Perception of tools, materials, and actions; (2) Procedural Understanding of step order and completeness; and (3) Scientific Reasoning that connects the full experiment to its published conclusions. Our vision-centric annotation pipeline, combining automated generation with multi-disciplinary expert validation, ensures that tasks require visual grounding. We evaluate 19 leading MLLMs on ExpVid and find that while they excel at coarse-grained recognition, they struggle with disambiguating fine details, tracking state changes over time, and linking experimental procedures to scientific outcomes. Our results reveal a notable performance gap between proprietary and open-source models, particularly in high-order reasoning. ExpVid not only provides a diagnostic tool but also charts a roadmap for developing MLLMs capable of becoming trustworthy partners in scientific experimentation.

cs.CV↗

arXiv2025

StreamForest: Efficient Online Video Understanding with Persistent Event Memory

Multimodal Large Language Models (MLLMs) have recently achieved remarkable progress in video understanding. However, their effectiveness in real-time streaming scenarios remains limited due to storage constraints of historical visual features and insufficient real-time spatiotemporal reasoning. To address these challenges, we propose StreamForest, a novel architecture specifically designed for streaming video understanding. Central to StreamForest is the Persistent Event Memory Forest, a memory mechanism that adaptively organizes video frames into multiple event-level tree structures. This process is guided by penalty functions based on temporal distance, content similarity, and merge frequency, enabling efficient long-term memory retention under limited computational resources. To enhance real-time perception, we introduce a Fine-grained Spatiotemporal Window, which captures detailed short-term visual cues to improve current scene perception. Additionally, we present OnlineIT, an instruction-tuning dataset tailored for streaming video tasks. OnlineIT significantly boosts MLLM performance in both real-time perception and future prediction. To evaluate generalization in practical applications, we introduce ODV-Bench, a new benchmark focused on real-time streaming video understanding in autonomous driving scenarios. Experimental results demonstrate that StreamForest achieves the state-of-the-art performance, with accuracies of 77.3% on StreamingBench, 60.5% on OVBench, and 55.6% on OVO-Bench. In particular, even under extreme visual token compression (limited to 1024 tokens), the model retains 96.8% of its average accuracy in eight benchmarks relative to the default setting. These results underscore the robustness, efficiency, and generalizability of StreamForest for streaming video understanding.

cs.CV↗

arXiv2025

VideoChat-R1.5: Visual Test-Time Scaling to Reinforce Multimodal Reasoning by Iterative Perception

Inducing reasoning in multimodal large language models (MLLMs) is critical for achieving human-level perception and understanding. Existing methods mainly leverage LLM reasoning to analyze parsed visuals, often limited by static perception stages. This paper introduces Visual Test-Time Scaling (VTTS), a novel approach to enhance MLLMs' reasoning via iterative perception during inference. VTTS mimics humans' hierarchical attention by progressively refining focus on high-confidence spatio-temporal regions, guided by updated textual predictions. Specifically, VTTS employs an Iterative Perception (ITP) mechanism, incorporating reinforcement learning with spatio-temporal supervision to optimize reasoning. To support this paradigm, we also present VTTS-80K, a dataset tailored for iterative perception. These designs allows a MLLM to enhance its performance by increasing its perceptual compute. Extensive experiments validate VTTS's effectiveness and generalization across diverse tasks and benchmarks. Our newly introduced Videochat-R1.5 model has achieved remarkable improvements, with an average increase of over 5\%, compared to robust baselines such as Qwen2.5VL-3B and -7B, across more than 15 benchmarks that encompass video conversation, video reasoning, and spatio-temporal perception.

cs.CV↗

arXiv2025

KiDS-Legacy: Covariance validation and the unified OneCovariance framework for projected large-scale structure observables

We introduce OneCovariance, an open-source software designed to accurately compute covariance matrices for an arbitrary set of two-point summary statistics across a variety of large-scale structure tracers. Utilising the halo model, we estimated the statistical properties of matter and biased tracer fields, incorporating all Gaussian, non-Gaussian, and super-sample covariance terms. The flexible configuration permits user-specific parameters, such as the complexity of survey geometry, the halo occupation distribution employed to define each galaxy sample, or the form of the real-space and/or Fourier space statistics to be analysed. We illustrate the capabilities of OneCovariance within the context of a cosmic shear analysis of the final data release of the Kilo-Degree Survey (KiDS-Legacy). Upon comparing our estimated covariance with measurements from mock data and calculations from independent software, we ascertain that OneCovariance achieves accuracy at the per cent level. When assessing the impact of ignoring complex survey geometry in the cosmic shear covariance computation, we discover misestimations at approximately the $10\%$ level for cosmic variance terms. Nonetheless, these discrepancies do not significantly affect the KiDS-Legacy recovery of cosmological parameters. We derive the cross-covariance between real-space correlation functions, bandpowers, and COSEBIs, facilitating future consistency tests among these three cosmic shear statistics. Additionally, we calculate the covariance matrix of photometric-spectroscopic galaxy clustering measurements, validating the jackknife covariance estimates for calibrating KiDS-Legacy redshift distributions. The OneCovariance can be found on GitHub, together with comprehensive documentation and examples.

astro-ph.CO↗

arXiv2025

InternVideo2.5: Empowering Video MLLMs with Long and Rich Context Modeling

This paper aims to improve the performance of video multimodal large language models (MLLM) via long and rich context (LRC) modeling. As a result, we develop a new version of InternVideo2.5 with a focus on enhancing the original MLLMs' ability to perceive fine-grained details and capture long-form temporal structure in videos. Specifically, our approach incorporates dense vision task annotations into MLLMs using direct preference optimization and develops compact spatiotemporal representations through adaptive hierarchical token compression. Experimental results demonstrate this unique design of LRC greatly improves the results of video MLLM in mainstream video understanding benchmarks (short & long), enabling the MLLM to memorize significantly longer video inputs (at least 6x longer than the original), and master specialized vision capabilities like object tracking and segmentation. Our work highlights the importance of multimodal context richness (length and fineness) in empowering MLLM's innate abilites (focus and memory), providing new insights for future research on video MLLM. Code and models are available at https://github.com/OpenGVLab/InternVideo/tree/main/InternVideo2.5

cs.CV↗

arXiv2025

Task Preference Optimization: Improving Multimodal Large Language Models with Vision Task Alignment

Current multimodal large language models (MLLMs) struggle with fine-grained or precise understanding of visuals although they give comprehensive perception and reasoning in a spectrum of vision applications. Recent studies either develop tool-using or unify specific visual tasks into the autoregressive framework, often at the expense of overall multimodal performance. To address this issue and enhance MLLMs with visual tasks in a scalable fashion, we propose Task Preference Optimization (TPO), a novel method that utilizes differentiable task preferences derived from typical fine-grained visual tasks. TPO introduces learnable task tokens that establish connections between multiple task-specific heads and the MLLM. By leveraging rich visual labels during training, TPO significantly enhances the MLLM's multimodal capabilities and task-specific performance. Through multi-task co-training within TPO, we observe synergistic benefits that elevate individual task performance beyond what is achievable through single-task training methodologies. Our instantiation of this approach with VideoChat and LLaVA demonstrates an overall 14.6% improvement in multimodal performance compared to baseline models. Additionally, MLLM-TPO demonstrates robust zero-shot capabilities across various tasks, performing comparably to state-of-the-art supervised models. The code will be released at https://github.com/OpenGVLab/TPO

cs.CV↗

arXiv2025

Quantifying the Detectability of Milky Way Satellites with Image Simulations: a Case Study with KiDS

Ultra-faint dwarf galaxies, which can be detected as resolved satellite systems of the Milky Way, are critical to understanding galaxy formation, evolution, and the nature of dark matter, as they are the oldest, smallest, most metal-poor, and most dark matter-dominated stellar systems known. Quantifying the sensitivity of surveys is essential for understanding their capability and limitations in searching for ultra-faint satellites. In this paper, we present the first study of the image-level observational selection function for Kilo-Degree Survey (KiDS) based on the Synthetic UniveRses For Surveys (surfs)-based KiDS-Legacy-Like Simulations. We generate mock satellites and simulate images that include resolved stellar populations of the mock satellites and the background galaxies, capturing realistic observational effects such as source blending, photometric uncertainties, and star-galaxy separation. The matched-filter method is applied to recover the injected satellites. We derive the observational selection function of the survey in terms of the luminosity, half-light radius, and heliocentric distance of the satellites. Compared to a catalogue-level simulation as used in previous studies, the image-level simulation provides a more realistic assessment of survey sensitivity, accounting for observational limitations that are neglected in catalogue-level simulations. The image-level simulation shows a detection loss for compact sources with a distance $d \gtrsim 100~\rm kpc$. We argue that this is because compact sources are more likely to be identified as single sources rather than being resolved during the source extraction process.

astro-ph.CO↗

arXiv2025

Correlation-Attention Masked Temporal Transformer for User Identity Linkage Using Heterogeneous Mobility Data

With the rise of social media and Location-Based Social Networks (LBSN), check-in data across platforms has become crucial for User Identity Linkage (UIL). These data not only reveal users' spatio-temporal information but also provide insights into their behavior patterns and interests. However, cross-platform identity linkage faces challenges like poor data quality, high sparsity, and noise interference, which hinder existing methods from extracting cross-platform user information. To address these issues, we propose a Correlation-Attention Masked Transformer for User Identity Linkage Network (MT-Link), a transformer-based framework to enhance model performance by learning spatio-temporal co-occurrence patterns of cross-platform users. Our model effectively captures spatio-temporal co-occurrence in cross-platform user check-in sequences. It employs a correlation attention mechanism to detect the spatio-temporal co-occurrence between user check-in sequences. Guided by attention weight maps, the model focuses on co-occurrence points while filtering out noise, ultimately improving classification performance. Experimental results show that our model significantly outperforms state-of-the-art baselines by 12.92%~17.76% and 5.80%~8.38% improvements in terms of Macro-F1 and Area Under Curve (AUC).

cs.SI↗

arXiv2025

The fifth data release of the Kilo Degree Survey: Multi-epoch optical/NIR imaging covering wide and legacy-calibration fields

We present the final data release of the Kilo-Degree Survey (KiDS-DR5), a public European Southern Observatory (ESO) wide-field imaging survey optimised for weak gravitational lensing studies. We combined matched-depth multi-wavelength observations from the VLT Survey Telescope and the VISTA Kilo-degree INfrared Galaxy (VIKING) survey to create a nine-band optical-to-near-infrared survey spanning $1347$ deg$^2$. The median $r$-band $5σ$ limiting magnitude is 24.8 with median seeing $0.7^{\prime\prime}$. The main survey footprint includes $4$ deg$^2$ of overlap with existing deep spectroscopic surveys. We complemented these data in DR5 with a targeted campaign to secure an additional $23$ deg$^2$ of KiDS- and VIKING-like imaging over a range of additional deep spectroscopic survey fields. From these fields, we extracted a catalogue of $126\,085$ sources with both spectroscopic and photometric redshift information, which enables the robust calibration of photometric redshifts across the full survey footprint. In comparison to previous releases, DR5 represents a $34\%$ areal extension and includes an $i$-band re-observation of the full footprint, thereby increasing the effective $i$-band depth by $0.4$ magnitudes and enabling multi-epoch science. Our processed nine-band imaging, single- and multi-band catalogues with masks, and homogenised photometry and photometric redshifts can be accessed through the ESO Archive Science Portal.

astro-ph.GA↗

arXiv2025

TimeSuite: Improving MLLMs for Long Video Understanding via Grounded Tuning

Multimodal Large Language Models (MLLMs) have demonstrated impressive performance in short video understanding. However, understanding long-form videos still remains challenging for MLLMs. This paper proposes TimeSuite, a collection of new designs to adapt the existing short-form video MLLMs for long video understanding, including a simple yet efficient framework to process long video sequence, a high-quality video dataset for grounded tuning of MLLMs, and a carefully-designed instruction tuning task to explicitly incorporate the grounding supervision in the traditional QA format. Specifically, based on VideoChat, we propose our long-video MLLM, coined as VideoChat-T, by implementing a token shuffling to compress long video tokens and introducing Temporal Adaptive Position Encoding (TAPE) to enhance the temporal awareness of visual representation. Meanwhile, we introduce the TimePro, a comprehensive grounding-centric instruction tuning dataset composed of 9 tasks and 349k high-quality grounded annotations. Notably, we design a new instruction tuning task type, called Temporal Grounded Caption, to peform detailed video descriptions with the corresponding time stamps prediction. This explicit temporal location prediction will guide MLLM to correctly attend on the visual content when generating description, and thus reduce the hallucination risk caused by the LLMs. Experimental results demonstrate that our TimeSuite provides a successful solution to enhance the long video understanding capability of short-form MLLM, achieving improvement of 5.6% and 6.8% on the benchmarks of Egoschema and VideoMME, respectively. In addition, VideoChat-T exhibits robust zero-shot temporal grounding capabilities, significantly outperforming the existing state-of-the-art MLLMs. After fine-tuning, it performs on par with the traditional supervised expert models.

cs.CV↗