Searcharxiv⌕ Search

arXiv subjects

Le Zhang

Publications and source records attributed to Le Zhang.

At least 55 records · Page 3Linked to original sources
arXiv2026

DynFrame: Adaptive Reasoning-Driven Multimodal Framework with Dynamic Frame Augmentation for Complex Video Understanding

Recent video multimodal large language models (MLLMs) increasingly couple step-by-step reasoning with on-demand visual evidence retrieval, allowing models to revisit relevant video segments during inference. However, two structural gaps remain in existing thinking-with-video systems. (i) Sampling density is not a learnable decision: existing methods may let the model decide where to look, but the per-window frame rate is largely fixed. As a result, fine-grained evidence is often recovered through repeated retrieval calls, which increases inference context length and training difficulty. (ii) Retrieval and answer generation are usually optimized with a single trajectory-level advantage, so the "where to look" tokens and the "how to answer" tokens receive the same credit even when one is correct and the other is not. To address these gaps, we present DynFrame, a framework that emits the temporal window and the sampling density as native tokens within a single autoregressive pass. This learnable span-density retrieval enables acquiring multi-granularity evidence with a single retrieval step. Based on the above tokenized retrieval interface, we further introduce Segment-Decoupled GRPO (SD-GRPO), which splits each rollout at the retrieval boundary and assigns role-specific token-level advantages, separately crediting the sampling decision and the answer. Trained on the curated DM-CoT-74k and DM-RL-45k, DynFrame-4B is competitive with strong 7B-8B baselines across six benchmarks (NExT-GQA, Charades-STA, ActivityNet-MR, Video-MME, MLVU, LVBench), and DynFrame-8B sets new state-of-the-art on most metrics. Code is available at https://github.com/zhangguanghao523/DynFrame.

cs.CV↗
arXiv2026

How Far Has AI Come in Liver Fibrosis Staging? A Large-Scale Real-World Dataset and Benchmark

Despite years of methodological progress, how far AI has come in liver fibrosis staging has never been systematically evaluated under the heterogeneous, multi-center conditions that define clinical practice. To address this gap, we introduce LiFS, a large-scale dataset and benchmark derived from the MICCAI 2025 CARE-Liver challenge, comprising 610 patients across multiple centers and scanners with multi-sequence MRI. To the best of our knowledge, LiFS is the first benchmark providing complete gadoxetic acid-enhanced sequences with histopathology-confirmed annotations from diverse real-world scanners. Through systematic evaluation of 9 independently developed methods selected from 96 registered teams against in-cohort radiologist reference results, our findings address how far current AI has progressed toward clinical-level liver fibrosis staging from three complementary perspectives. First, against radiologists, the best AI methods were broadly comparable to the senior radiologist and significantly exceeded the junior radiologist in selected settings, while median AI performance generally approached junior-radiologist levels. Second, from a data perspective, cross-center heterogeneity, label imbalance, and contrast-enhanced sequence variability emerge as the dominant challenges for AI methods. Third, from a technical perspective, methodological design choices, including spatial registration, input dimensionality, multi-modal fusion strategy, and backbone architecture, appear to modulate cross-center robustness, although no single choice alone closes the gap. Overall, LiFS provides a rigorous real-world benchmark for positioning the current state of AI in liver fibrosis staging and for enabling future research on the key challenges that limit clinically reliable deployment.

cs.CV↗
arXiv2026

RiT: Vanilla Diffusion Transformers Suffice in Representation Space

Flow matching with $x$-prediction -- regressing the clean data point rather than the ambient velocity -- is known to exploit low-dimensional manifold structure effectively in pixel space \cite{li2025back}. We ask whether a pretrained representation space, while containing a low-dimensional data manifold of comparable intrinsic dimensionality, offers a distribution more favorable for flow-matching learning. Comparing pixel, SD-VAE, and DINOv2 features along four geometric axes, we find that pixel and DINOv2 share nearly identical intrinsic dimensionalities (both $\hat{d}\!\approx\!33$) yet DINOv2 exhibits $7.3\times$ higher effective rank, $35\times$ better covariance conditioning, $11.5\times$ lower excess kurtosis, and $1.7\times$ lower on-manifold interpolation error; SD-VAE latents are consistently intermediate, indicating that the advantage stems from representation-learning objectives rather than mere compression. These statistical properties render the flow-matching regression well-conditioned and remove the need for the specialized prediction heads or Riemannian transport used by prior DINOv2 diffusion methods. We propose the \emph{Representation Image Transformer} (RiT): a vanilla Diffusion Transformer trained by $x$-prediction on frozen DINOv2 features, augmented only by a dimension-aware noise schedule and joint \texttt{[CLS]}-patch modeling. On ImageNet $256{\times}256$, RiT attains FID 1.45 without guidance and 1.14 with classifier-free guidance, outperforming DiT$^\text{DH}$-XL with $19\%$ fewer parameters (676M vs.\ 839M). The resulting ODE is efficiently solvable at coarse discretizations: with classifier-free guidance, $5$ Heun steps already reach FID 2.0 and $10$ steps reach 1.25, without distillation or consistency training. Code at https://github.com/lezhang7/RiT.

cs.CV↗
arXiv2026

The Shear-to-Cosmology Paradigm I. Hybrid Field-Level and Simulation-Based Framework for Weak Lensing Surveys

Precise cosmological inference from next-generation weak lensing surveys requires extracting non-Gaussian information beyond standard two-point statistics. We present a hybrid machine-learning (ML) framework that integrates field-level inference (FLI) with simulation-based inference (SBI) to map observed shear fields directly to cosmological parameters, eliminating the need for convergence reconstruction. The FLI network extracts rich non-Gaussian information from the shear field to produce informative features, which are then used by SBI to model the resulting complex posteriors. To mitigate noise from intrinsic galaxy shapes, we develop a blind, training-free, PCA-based shear denoising method. Tests on CSST-like mock catalogs reveal significant performance gains. The shear-based inference achieves approximately twice the cosmological constraining power in Figure of Merit (FoM) compared to the conventional convergence-based approach. Moreover, the combination of PCA denoising and ML compression can deliver a 36.4% improvement in FoM over standard shear two-point statistics. This work establishes a scalable and robust pathway for cosmological inference, unlocking the full potential of Stage-IV weak-lensing surveys.

astro-ph.CO↗
arXiv2026

DuIVRS-2: An LLM-based Interactive Voice Response System for Large-scale POI Attribute Acquisition

Accurate Point of Interest (POI) attribute acquisition is essential for location-based services, yet traditional modular Interactive Voice Response (IVR) systems suffer from error accumulation and high maintenance overhead. We present DuIVRS-2, a large language model (LLM)-based end-to-end framework designed for large-scale POI attribute acquisition at Baidu Maps. To address the long-tail distribution of real-world interactions, our methodology first employs a finite state machine (FSM)-guided data augmentation strategy to synthesize a balanced and diverse training dataset. We then streamline dialogue management via a selective generation scheme combined with a Chain-of-Thought (CoT) mechanism, which ensures output stability and effectively eliminates hallucinations in industrial settings. To facilitate continuous policy refinement with minimal manual effort, we design a cooperative iterative learning framework that leverages a dual-evaluator voting system. Deployed in production for two months, DuIVRS-2 processed 0.4 million calls daily and achieved a 83.9\% Task Success Rate (TSR), outperforming its predecessor by 4 percentage points while maintaining a low reaction time of 130ms. This work provides a production-proven reference for developing robust, cost-effective LLM agents for large-scale industrial dialogue applications.

cs.AI↗
arXiv2026

Hi-GaTA: Hierarchical Gated Temporal Aggregation Adapter for Surgical Video Report Generation

Automated, clinician-grade assessment reports for surgical procedures could reduce documentation burden and provide objective feedback, yet remain challenging due to the difficulty of aligning dense spatio-temporal video representations with language-based reasoning and the scarcity of high-quality, privacy-preserving datasets. To address this gap, we establish a benchmark comprising 214 high-quality simulated surgical videos paired with surgeon-authored evaluation reports. Building on this resource, we propose a Perception-Alignment-Reasoning framework for surgical video report generation, featuring Hi-GaTA, a novel lightweight temporal adapter that efficiently compresses long video sequences into compact, LLM-compatible visual prefix tokens through short-to-long-range temporal aggregation. For robust visual perception, we pretrain Sur40k, a surgical-specific ViViT-style video encoder on 40,000 minutes of public surgical videos to capture fine-grained spatio-temporal procedural priors. Hi-GaTA employs a temporal pyramid with text-conditioned dual cross-attention, and improves multi-scale consistency through cross-level gated fusion and an increasing-depth strategy. Finally, we fine-tune the LLM backbone using LoRA to enable coherent and stylistically consistent surgical report generation under limited supervision. Experiments show our approach achieves the best overall performance, with consistent gains over strong Multimodal Large Language Model (MLLM) baselines. Ablation studies further validate the effectiveness of each proposed component.

cs.CV↗
arXiv2026

A Spatially Resolved HI Survey of Seyfert Galaxies: the Role of AGN Feedback in Shaping Atomic Gas Reservoirs

Active galactic nucleus (AGN) feedback is a key ingredient in galaxy evolution, yet its impact on the cold atomic gas reservoir -- the neutral hydrogen (HI) phase -- remains poorly constrained. We present the most extensive spatially resolved HI 21-cm survey of Seyfert AGN hosts to date, based on observations with the Giant Metrewave Radio Telescope (GMRT). Our high-resolution HI maps of eight Seyfert galaxies reveal detailed kinematics and surface density distributions of their atomic gas disks. We find that AGN-host galaxies exhibit a slightly shallower HI mass-size relation than the canonical relation or the SIMBA simulation predictions; however, the measured slope remains consistent with the canonical value within $2σ$ uncertainties. This result suggests that AGN feedback does not significantly disrupt the global extent or large-scale structure of atomic gas reservoirs. To investigate the internal HI kinematics in greater detail, we perform a 3D kinematic forward modeling of the HI disk in UGC 4503. Our analysis reveals an elevated intrinsic velocity dispersion of $σ= 14.9^{+6.1}_{-3.8}$ km/s and a reduced level of rotational support, with $V/σ= 14.28_{-4.17}^{+4.97}$, compared to large-sample star-forming spirals. These kinematic signatures, together with localized residuals in the velocity field, indicate that AGN-driven outflows or jets may inject or indirectly affect the turbulence in the atomic gas disk, potentially regulating the cold gas reservoir. Future GMRT observations, combined with optical integral-field spectroscopy from MaNGA, will enable quantitative constraints on the role of AGN feedback in regulating star formation efficiency across a larger and more representative galaxy sample.

astro-ph.GA↗
arXiv2026

DRNet: All-in-One Image Restoration via Prior-Guided Dynamic Reparameterization

All-in-one image restoration aims to handle diverse degradations within a single model. However, existing methods often suffer from three key limitations: 1) per-input computational overhead from dynamic degradation estimation; 2) optimization challenges due to task heterogeneity; and 3) inefficient, frequency-agnostic encoder designs. To overcome these, we introduce the Dynamic Reparameterization Network (DRNet), a novel framework operating on an initialization-stage reconfiguration paradigm that fundamentally eliminates per-input overhead. At its core, a Dynamic Reparameterization MLP (DRMLP) guided by a Task-Specific Modulator (TSM), which effectively mitigates task heterogeneity by orchestrating both specific restoration goals and a versatile general-purpose mode within a unified architecture. Furthermore, we incorporate a Continuous Wavelet Transform Encoder (CWTE) that explicitly leverages frequency characteristics via wavelet decomposition for a lightweight yet powerful design. Extensive experiments demonstrate that DRNet achieves state-of-the-art performance across five restoration tasks with superior parameter efficiency. Crucially, it showcases unique flexibility, excelling as both a highly competitive foundation model for blind restoration and a top-performing user-guided specialist.

cs.CV↗
arXiv2026

Diffusion-APO: Trajectory-Aware Direct Preference Alignment for Video Diffusion Transformers

Efficiently aligning large-scale video diffusion models with human intent requires a scalable and trajectory-aware pathway that bridges the inherent discrepancy between training noise distributions and practical inference trajectories. While existing paradigms such as Direct Preference Optimization (DPO) and Group Relative Policy Optimization (GRPO) attempt to address this, they are often hindered by either reliance on bias-prone, complex reward models or suboptimal timestep sampling. In this paper, we propose Diffusion-APO (Aligned Preference Optimization), a trajectory-aware algorithm that resolves this misalignment by synchronizing training noise with inference-time denoising paths to maximize gradient signal efficacy. To translate this algorithmic innovation into a practical solution, we introduce a unified and modular RLHF framework that integrates online ranking, half-online anchoring, offline refinement, and distillation-aware drift correction. This framework enables flexible, multi-stage preference alignment across diverse data and computational constraints without relying on scalar-reward-based policy gradients. Through extensive experiments, we demonstrate that Diffusion-APO consistently outperforms standard baselines in visual quality and instruction following, while effectively preserving generative fidelity during model acceleration, providing a robust, end-to-end pathway for scalable video diffusion alignment.

cs.CV↗
arXiv2026

ReportMedSAM: Guiding Segmentation Through Radiology Reports

Free-form radiology reports contain rich clinical descriptions, yet converting them for reliable segmentation remains challenging due to the inherent variability of natural language. Existing pipelines often rely on predefined organ phrases or brittle rule-based inference-time extraction, which limits their scalability to novel anatomical structures and makes them sensitive to linguistic variations. To address this, we propose ReportMedSAM, a report-driven framework that replaces discrete extraction with a learnable concept bank. By leveraging a frozen medical vision-language encoder (BiomedCLIP), we align organ-level concept embeddings with large-scale clinical corpora through contrastive learning, establishing mutually orthogonal semantic anchors. Our approach explicitly mitigates organ-level semantic collapse and ensures high robustness against diverse clinical synonyms (e.g., "renal" vs. "kidney" ). During inference, a clinical report is embedded and matched against this concept bank to dynamically activate task-specific Mixture-of-Experts (MoE) modules. This decoupled design allows new concepts and experts to be added without retraining existing components, providing a parameter-isolated extension mechanism while keeping previously learned experts unchanged. Evaluated on the AbdomenAtlas 3.0 dataset, ReportMedSAM effectively interprets free-form reports, achieves competitive segmentation accuracy, and demonstrates seamless, non-interfering extension to novel clinical tasks.

cs.CL↗
arXiv2026

MedFlowSeg: Flow Matching for Medical Image Segmentation with Frequency-Aware Attention

Flow matching has recently emerged as a principled framework for learning continuous-time transport maps, enabling efficient ODE-based sampling without relying on stochastic diffusion processes. While generative modeling has shown promise for medical image segmentation, particularly in capturing uncertainty and complex anatomical variability, existing approaches are predominantly based on diffusion models, which require iterative sampling and incur substantial computational overhead. In this work, we propose MedFlowSeg, a conditional flow matching framework that formulates medical image segmentation as learning a time-dependent vector field that transports a simple prior distribution to the target segmentation distribution. Compared to diffusion-based methods, our formulation enables more efficient inference through solving an ordinary differential equation, while preserving the flexibility of generative modeling. To effectively incorporate conditional information, we introduce a dual-conditioning mechanism. Specifically, we propose a Dual-Branch Spatial Attention (DB-SA) module to inject multi-frequency structural priors, and a Frequency-Aware Attention (FA-Attention) module to model interactions between spatial and spectral representations via discrepancy-aware fusion and time-dependent modulation. These components improve the alignment between noisy intermediate states and clean semantic features, leading to better structural consistency and boundary delineation. We conduct extensive experiments across multiple medical imaging modalities, where MedFlowSeg consistently outperforms prior state-of-the-art (SOTA) baselines, including diffusion-based and flow-based methods.

cs.CV↗
arXiv2026

From Where Things Are to What They Are For: Benchmarking Spatial-Functional Intelligence in Multimodal LLMs

Human-level agentic intelligence extends beyond low-level geometric perception, evolving from recognizing where things are to understanding what they are for. While existing benchmarks effectively evaluate the geometric perception capabilities of multimodal large language models (MLLMs), they fall short of probing the higher-order cognitive abilities required for grounded intelligence. To address this gap, we introduce the Spatial-Functional Intelligence Benchmark (SFI-Bench), a video-based benchmark with over 1,500 expert-annotated questions derived from diverse egocentric indoor video scans. SFI-Bench systematically evaluates two complementary dimensions of advanced reasoning: (1) Structured Spatial Reasoning, which requires understanding complex layouts and forming coherent spatial representations, and (2) Functional Reasoning, which involves inferring object affordances and their context-dependent utility. The benchmark includes tasks such as conditional counting, multi-hop relational reasoning, functional pairing, and knowledge-grounded troubleshooting, directly challenging models to integrate perception, memory, and inference. Our experiments reveal that current MLLMs consistently struggle to combine spatial memory with functional reasoning and external knowledge, highlighting a critical bottleneck in achieving grounded intelligence. SFI-Bench therefore provides a diagnostic tool for measuring progress toward more cognitively capable and truly grounded multimodal agents.

cs.CV↗
arXiv2026

The Swarm Intelligence Freeway-Urban Trajectories (SWIFTraj) Dataset -- Part II: A Graph-Based Approach for Trajectory Connection

In Part I of this companion paper series, we introduced SWIFTraj, a new open-source vehicle trajectory dataset collected using a unmanned aerial vehicle (UAV) swarm. The dataset has two distinctive features. First, by connecting trajectories across consecutive UAV videos, it provides long-distance continuous trajectories, with the longest exceeding 4.5 km. Second, it covers an integrated traffic network consisting of both freeways and their connected urban roads. Obtaining such long-distance continuous trajectories from a UAV swarm is challenging, due to the need for accurate time alignment across multiple videos and the irregular spatial distribution of UAVs. To address these challenges, this paper proposes a novel graph-based approach for connecting vehicle trajectories captured by a UAV swarm. An undirected graph is constructed to represent flexible UAV layouts, and an automatic time alignment method based on trajectory matching cost minimization is developed to estimate optimal time offsets across videos. To associate trajectories of the same vehicle observed in different videos, a vehicle matching table is established using the Hungarian algorithm. The proposed approach is evaluated using both simulated and real-world data. Results from real-world experiments show that the time alignment error is within three video frames, corresponding to approximately 0.1 s, and that the vehicle matching achieves an F1-score of about 0.99. These results demonstrate the effectiveness of the proposed method in addressing key challenges in UAV-based trajectory connection and highlight its potential for large-scale vehicle trajectory collection.

physics.soc-ph↗
arXiv2026

Reconstructing the Thermal Sunyaev Zeldovich Power Spectrum from Planck using the ABS Method

This study employs a novel approach for reconstructing the thermal Sunyaev-Zeldovich (tSZ) effect power spectrum from Planck data using the Analytical Blind Separation (ABS) method. The ABS method improves the recovery of weak signals, by applying eigenmode exclusion for low signal-to-noise ratio regimes and introducing a shift parameter to stabilize calculations. Validation through simulated Planck data demonstrates the robustness of ABS in reconstructing the tSZ power spectrum, even under challenging conditions. In the analysis of the {\it Planck} PR3 full-mission data, ABS shows lower amplitudes at $\ell \gtrsim 300$ compared to the {\it Planck} 2015 band powers using the MILCA and NILC foreground cleaning methods. In our analysis, we include or exclude the trispectrum contribution to the statistical uncertainty to enable comparison with previous studies and to test the robustness of our results. When the trispectrum contribution is included, and after marginalizing over residual foreground components, we find that the overall amplitude of the tSZ power spectrum is 34\% lower than the ``Planck 2015'' best-fit amplitude, 24\% lower than the ``Battaglia 2012'' model, and 12\% lower than the ``Bolliet 2018'' best-fit amplitude. These differences correspond to $3.0σ$, $1.78σ$, and $0.75σ$, respectively, in terms of the associated statistical uncertainties. Across all cases, the tSZ band power remains unaffected by template choice. These findings highlight the potential of the ABS method as a promising alternative for tSZ power spectrum analysis.

astro-ph.CO↗
arXiv2026

Tomographic Alcock-Paczynski Test with Marked Correlation Functions

The tomographic Alcock-Paczynski(AP) method, developed over the past decade, exploits redshift evolution for cosmological determination, aiming to mitigate contamination from redshift distortions and capture nonlinear scale information. Marked Correlation Functions (MCFs) extend information beyond the two-point correlation. For the first time, this study integrated the tomographic AP test with MCFs to constrain the flat $w$CDM cosmology model. Our findings show that multiple density weights in MCFs outperform the traditional two-point correlation function, reducing the uncertainties of the matter density parameter $Ω_m$ and dark energy equation of state $w$ by 48\% and 45\%, respectively. Furthermore, we introduce a novel principal component analysis (PCA) compression scheme that efficiently projects high-dimensional statistical measurements into a compact set of eigenmodes while preserving most of the cosmological information. This approach retains significantly more information than traditional coarse binning methods, which simply average adjacent bins in a lossy manner, yielding an additional $\sim 40\%$ reduction in error margins. To assess robustness, we incorporate realistic redshift errors expected in future spectroscopic surveys. While these errors modestly degrade cosmological constraints, our combined framework, which utilizes MCFs and PCA compression within tomographic AP tests, is less affected and always yields tight cosmological constraints. This scheme remains highly promising for upcoming slitless spectroscopic surveys, such as the Chinese Space Station Telescope (CSST).

astro-ph.CO↗
arXiv2026

Closing the Observational Gap in Cosmic Dynamics: AI-Enabled Reconstruction of the Universe's Vorticity and Rotational Flow Morphology

The cosmic vorticity field, an essential tracer of nonlinear structure formation, has remained observationally inaccessible because transverse galaxy motions are difficult to measure and analytic models struggle to capture shell-crossing. Here we report an empirical reconstruction of this field by applying an artificial intelligence framework trained on simulations of the concordance LambdaCDM model to Sloan Digital Sky Survey galaxies. The recovered three-dimensional velocity and vorticity fields reveal coherent vortical structures, including spiral flows in clusters, filaments, and voids, and the cosmic web inferred from vorticity closely matches that derived from density segmentation. The power spectra of the reconstructed velocity and vorticity fields agree statistically with LambdaCDM predictions, and the inferred velocity field effectively removes redshift-space distortions, yielding an almost isotropic clustering signal. These converging lines of evidence, obtained from an independent perspective, reinforce the concordance cosmological model. By closing a long-standing observational gap, our results highlight the potential of AI-driven reconstruction to access otherwise unobservable quantities and to address fundamental questions in cosmology and galaxy formation.

astro-ph.CO↗
arXiv2026

Make It Up: Fake Images, Real Gains in Generalized Few-shot Semantic Segmentation

Generalized few-shot semantic segmentation (GFSS) is fundamentally limited by the coverage of novel-class appearances under scarce annotations. While diffusion models can synthesize novel-class images at scale, practical gains are often hindered by insufficient coverage and noisy supervision when masks are unavailable or unreliable. We propose Syn4Seg, a generation-enhanced GFSS framework designed to expand novel-class coverage while improving pseudo-label quality. Syn4Seg first maximizes prompt-space coverage by constructing an embedding-deduplicated prompt bank for each novel class, yielding diverse yet class-consistent synthetic images. It then performs support-guided pseudo-label estimation via a two-stage refinement that i) filters low-consistency regions to obtain high-precision seeds and ii) relabels uncertain pixels with image-adaptive prototypes that combine global (support) and local (image) statistics. Finally, we refine only boundary-band and unlabeled pixels using a constrained SAM-based update to improve contour fidelity without overwriting high-confidence interiors. Extensive experiments on PASCAL-$5^i$ and COCO-$20^i$ demonstrate consistent improvements in both 1-shot and 5-shot settings, highlighting synthetic data as a scalable path for GFSS with reliable masks and precise boundaries.

cs.CV↗
arXiv2026

Enhancing cosmological constraints with nonlinear tanh transformations of Hermite-Gaussian Derivative fields

A key goal in large-scale structure analysis is to extract multi-scale information to improve cosmological parameter constraints. In particular, higher-order derivative fields are especially valuable as they capture the geometric and topological information of the cosmic web that is highly sensitive to cosmological parameters. Traditional derivative-based methods, such as finite-difference or Fourier approaches, suffer from noise amplification at small scales and cannot stably capture multi-scale features. We present a robust two-step framework: first, stable multi-scale arbitrary-order derivatives are obtained via Hermite-Gaussian convolutional filters that suppress small-scale noise; second, a tanh nonlinear transformation compresses extreme density contrasts and enhances the visibility of cosmic web structures. Using the Quijote simulations, we show that combining multi-scale first-order spectra yields improvements of 1.2-3.0 times across all seven cosmological parameters, while multi-order spectra at a fixed scale provide 1.3-2.9 times gains. The most comprehensive combination achieves nominal gains of 2.0-5.3 times. Our method offers a robust approach to extracting additional cosmological information for future surveys.

astro-ph.CO↗