arXiv subjects

Xin Wang

Publications and source records attributed to Xin Wang.

At least 379 records · Page 21Linked to original sources

arXiv2026

Input-Adaptive Generative Dynamics in Diffusion Models

Diffusion models typically generate data through a fixed denoising trajectory that is shared across all samples. However, generation targets can differ in complexity, suggesting that a single pre-defined diffusion process may not be optimal for every input. In this work, we investigate input-adaptive generative dynamics for diffusion models, where the generation process itself adapts to the conditions of each sample. Instead of relying on a fixed diffusion trajectory, the proposed framework allows the generative dynamics to adjust across inputs according to their generation requirements. To enable this behavior, we train the diffusion backbone under varying horizons and noise schedules, so that it can operate consistently under different input-adaptive trajectories. Experiments on conditional image generation show that diffusion trajectories can vary across inputs while maintaining generation quality and reducing the average number of sampling steps. These results provide a proof of the concept that diffusion processes can benefit from input-adaptive generative dynamics rather than relying on a single fixed trajectory.

cs.CV↗

arXiv2026

CanoVerse: 3D Object Scalable Canonicalization and Dataset for Generation and Pose

3D learning systems implicitly assume that objects occupy a coherent reference frame. Nonetheless, in practice, every asset arrives with an arbitrary global rotation, and models are left to resolve directional ambiguity on their own. This persistent misalignment suppresses pose-consistent generation, and blocks the emergence of stable directional semantics. To address this issue, we construct \methodName{}, a massive canonical 3D dataset of 320K objects over 1,156 categories -- an order-of-magnitude increase over prior work. At this scale, directional semantics become statistically learnable: Canoverse improves 3D generation stability, enables precise cross-modal 3D shape retrieval, and unlocks zero-shot point-cloud orientation estimation even for out-of-distribution data. This is achieved by a new canonicalization framework that reduces alignment from minutes to seconds per object via compact hypothesis generation and lightweight human discrimination, transforming canonicalization from manual curation into a high-throughput data generation pipeline. The Canoverse dataset will be publicly released upon acceptance. Project page: https://github.com/123321456-gif/Canoverse

cs.CV↗

arXiv2026

N-Tree Diffusion for Long-Horizon Wildfire Risk Forecasting

Long-horizon wildfire risk forecasting requires generating probabilistic spatial fields under sparse event supervision while maintaining computational efficiency across multiple prediction horizons. Extending diffusion models to multi-step forecasting typically repeats the denoising process independently for each horizon, leading to redundant computation. We introduce N-Tree Diffusion (NT-Diffusion), a hierarchical diffusion model designed for long-horizon wildfire risk forecasting. Fire occurrences are represented as continuous Fire Risk Maps (FRMs), which provide a smoothed spatial risk field suitable for probabilistic modeling. Instead of running separate diffusion trajectories for each predicted timestamp, NT-Diffusion shares early denoising stages and branches at later levels, allowing horizon-specific refinement while reducing redundant sampling. We evaluate the proposed framework on a newly collected real-world wildfire dataset constructed for long-horizon probabilistic prediction. Results indicate that NT-Diffusion achieves consistent accuracy improvements and reduced inference cost compared to baseline forecasting approaches.

cs.LG↗

arXiv2026

Beyond Geometry: Artistic Disparity Synthesis for Immersive 2D-to-3D

Current 2D-to-3D conversion methods achieve geometric accuracy but are artistically deficient, failing to replicate the immersive and emotionally resonant experience of professional 3D cinema. This is because geometric reconstruction paradigms mistake deliberate artistic intent, such as strategic zero-plane shifts for pop-out effects and local depth sculpting, for data noise or ambiguity. This paper argues for a new paradigm: Artistic Disparity Synthesis, shifting the goal from physically accurate disparity estimation to artistically coherent disparity synthesis. We propose Art3D, a preliminary framework exploring this paradigm. Art3D uses a dual-path architecture to decouple global depth parameters (macro-intent) from local artistic effects (visual brushstrokes) and learns from professional 3D film data via indirect supervision. We also introduce a preliminary evaluation method to quantify cinematic alignment. Experiments show our approach demonstrates potential in replicating key local out-of-screen effects and aligning with the global depth styles of cinematic 3D content, laying the groundwork for a new class of artistically-driven conversion tools.

cs.CV↗

arXiv2026

AV-Unified: A Unified Framework for Audio-visual Scene Understanding

When humans perceive the world, they naturally integrate multiple audio-visual tasks within dynamic, real-world scenes. However, current works such as event localization, parsing, segmentation and question answering are mostly explored individually, making it challenging to comprehensively understand complex audio-visual scenes and explore inter-task relationships. Hence, we propose \textbf{AV-Unified}, a unified framework that enables joint learning across a wide range of audio-visual scene understanding tasks. AV-Unified standardizes the diverse input-output formats of each task and incorporates a multi-scale spatiotemporal perception network to effectively capture audio-visual associations. Specifically, we unify the inputs and outputs of all supported tasks by converting them into sequences of discrete tokens, establishing a shared representation that allows a single architecture to be trained jointly across heterogeneous varied datasets. Considering the varying temporal granularity of audio-visual events, a multi-scale temporal perception module is designed to capture key cues. Meanwhile, to overcome the lack of auditory supervision in the visual domain, we design a cross-modal guidance-based spatial perception module that models spatial audio-visual associations. Furthermore, task-specific text prompts are employed to enhance the model's adaptability and task-awareness. Extensive experiments on benchmark datasets (e.g., AVE, LLP, MUSIC-AVQA, VGG-SS and AVS) demonstrate the effectiveness of AV-Unified across temporal, spatial, and spatiotemporal tasks.

cs.CV↗

arXiv2026

Resurgent structure of 2d Yang-Mills theory on a torus

We study the resurgent structure of the topological string dual to 2d $U(N)$ Yang-Mills on torus. We find closed form formulas for instanton amplitudes up to arbitrarily high instanton orders, based on which we propose the non-perturbative partition function including contributions from all the real instantons, which is real for positive modulus and string coupling. We also explore complex instantons and find two infinite towers of them. We expect them to correspond to BPS states in type II string.

hep-th↗

arXiv2026

Does Fine-tuning by Reinforcement Learning Improve Generalization in Binary Speech Deepfake Detection?

Building speech deepfake detection models that are generalizable to unseen attacks remains a challenging problem. Although the field has shifted toward a pre-training and fine-tuning paradigm using speech foundation models, most approaches rely solely on supervised fine-tuning (SFT). Inspired by the field of large language models, wherein reinforcement learning (RL) is used for model fine-tuning, we investigate the impact of RL, specifically Group Relative Policy Optimization (GRPO). The results from experiments using multiple detectors and test sets indicate that pure GRPO-based fine-tuning improves performance on out-of-domain test sets while maintaining performance on target-domain test data. This approach outperforms both SFT-only and hybrid setups. Our ablation studies further suggest that the negative reward in GRPO may be a key factor in this improvement.

eess.AS↗

arXiv2026

GNN Based Joint Beamforming Design for Extremely Large-Scale RIS Assisted Near-Field ISAC Systems

This paper investigates an extremely large-scale reconfigurable intelligent surface (XL-RIS) assisted near-field integrated sensing and communication (ISAC) system, where a multi-antenna base station (BS) simultaneously sends unicast data to multiple single-antenna communication users (CUs) and senses multiple targets (TGTs). The BS, CUs and TGTs are \emph{all} assumed to be located in the near-field region of the XL-RIS. We aim to maximize the weighted sum rate (WSR) of all CUs, subject to the sensing beampattern gain constraint for each TGT, the transmit power constraint for the BS, and the unit modulus constraints on the XL-RIS phase shift. First, we develop a fractional programming (FP) based block coordinate descent (BCD) algorithm to obtain a locally optimal solution for such a non-convex joint design problem. Secondly, to address the high-dimensional spatial correlations and scalability of the XL-RIS near-field channels, we propose a customized graph neural network (GNN) scheme to generate the BS transmit beamforming variables and the XL-RIS reflecting coefficient vector for ISAC, where the near-field ISAC system is modeled as a heterogeneous graph comprising XL-RIS/CU/TGT nodes. The proposed GNN scheme can effectively learn the near-field channel state information (CSI) features, in which the message passing mechanism is employed to exchange CSI among these directly connected nodes in the graph. Furthermore, each XL-RIS/CU/TGT node maintains a feature vector for mapping to the BS transmit beamforming variables or the XL-RIS reflecting coefficient vector. Numerical results show that the proposed GNN-based beamforming design scheme achieves a better performance than the existing baselines, in terms of computational efficiency, feasibility, robustness, and the ability of generalization.

cs.IT↗

arXiv2026

Revisiting Global Token Mixing in Task-Dependent MRI Restoration: Insights from Minimal Gated CNN Baselines

Global token mixing, implemented via self-attention or state-space sequence models, has become a popular model design choice for MRI restoration. However, MRI restoration tasks differ substantially in how their degradations vary over image and k-space domains, and in the degree to which global coupling is already imposed by physics-driven data consistency terms. In this work, we ask the question whether global token mixing is actually beneficial in each individual task across three representative settings: accelerated MRI reconstruction with explicit data consistency, MRI super-resolution with k-space center cropping, and denoising of clinical carotid MRI data with spatially heteroscedastic noise. To reduce confounding factors, we establish a controlled testbed comparing a minimal local gated CNN and its large-field variant, benchmarking them directly against state-of-the-art global models under aligned training and evaluation protocols. For accelerated MRI reconstruction, the minimal unrolled gated-CNN baseline is already highly competitive compared to recent token-mixing approaches in public reconstruction benchmarks, suggesting limited additional benefits when the forward model and data-consistency steps provide strong global constraints. For super-resolution, where low-frequency k-space data are largely preserved by the controlled low-pass degradation, local gated models remain competitive, and a lightweight large-field variant yields only modest improvements. In contrast, for denoising with pronounced spatially heteroscedastic noise, token-mixing models achieve the strongest overall performance, consistent with the need to estimate spatially varying reliability. In conclusion, our results demonstrate that the utility of global token mixing in MRI restoration is task-dependent, and it should be tailored to the underlying imaging physics and degradation structure.

eess.IV↗

arXiv2026

Shape-Interpretable Visual Self-Modeling Enables Geometry-Aware Continuum Robot Control

Continuum robots possess high flexibility and redundancy, making them well suited for safe interaction in complex environments, yet their continuous deformation and nonlinear dynamics pose fundamental challenges to perception, modeling, and control. Existing vision-based control approaches often rely on end-to-end learning, achieving shape regulation without explicit awareness of robot geometry or its interaction with the environment. Here, we introduce a shape-interpretable visual self-modeling framework for continuum robots that enables geometry-aware control. Robot shapes are encoded from multi-view planar images using a Bezier-curve representation, transforming visual observations into a compact and physically meaningful shape space that uniquely characterizes the robot's three-dimensional configuration. Based on this representation, neural ordinary differential equations are employed to self-model both shape and end-effector dynamics directly from data, enabling hybrid shape-position control without analytical models or dense body markers. The explicit geometric structure of the learned shape space allows the robot to reason about its body and surroundings, supporting environment-aware behaviors such as obstacle avoidance and self-motion while maintaining end-effector objectives. Experiments on a cable-driven continuum robot demonstrate accurate shape-position regulation and tracking, with shape errors within 1.56% of image resolution and end-effector errors within 2% of robot length, as well as robust performance in constrained environments. By elevating visual shape representations from two-dimensional observations to an interpretable three-dimensional self-model, this work establishes a principled alternative to vision-based end-to-end control and advances autonomous, geometry-aware manipulation for continuum robots.

cs.RO↗

arXiv2026

CoSMo3D: Open-World Promptable 3D Semantic Part Segmentation through LLM-Guided Canonical Spatial Modeling

Open-world promptable 3D semantic segmentation remains brittle as semantics are inferred in the input sensor coordinates. Yet, humans, in contrast, interpret parts via functional roles in a canonical space -- wings extend laterally, handles protrude to the side, and legs support from below. Psychophysical evidence shows that we mentally rotate objects into canonical frames to reveal these roles. To fill this gap, we propose \methodName{}, which attains canonical space perception by inducing a latent canonical reference frame learned directly from data. By construction, we create a unified canonical dataset through LLM-guided intra- and cross-category alignment, exposing canonical spatial regularities across 200 categories. By induction, we realize canonicality inside the model through a dual-branch architecture with canonical map anchoring and canonical box calibration, collapsing pose variation and symmetry into a stable canonical embedding. This shift from input pose space to canonical embedding yields far more stable and transferable part semantics. Experimental results show that \methodName{} establishes new state of the art in open-world promptable 3D segmentation.

cs.CV↗

arXiv2026

FaithCoT-Bench: Benchmarking Instance-Level Faithfulness of Chain-of-Thought Reasoning

Large language models (LLMs) increasingly rely on Chain-of-Thought (CoT) prompting to improve problem-solving and provide seemingly transparent explanations. However, growing evidence shows that CoT often fail to faithfully represent the underlying reasoning process, raising concerns about their reliability in high-risk applications. Although prior studies have focused on mechanism-level analyses showing that CoTs can be unfaithful, they leave open the practical challenge of deciding whether a specific trajectory is faithful to the internal reasoning of the model. To address this gap, we introduce FaithCoT-Bench, a unified benchmark for instance-level CoT unfaithfulness detection. Our framework establishes a rigorous task formulation that formulates unfaithfulness detection as a discriminative decision problem, and provides FINE-CoT (Faithfulness instance evaluation for Chain-of-Thought), an expert-annotated collection of over 1,000 trajectories generated by four representative LLMs across four domains, including more than 300 unfaithful instances with fine-grained causes and step-level evidence. We further conduct a systematic evaluation of eleven representative detection methods spanning counterfactual, logit-based, and LLM-as-judge paradigms, deriving empirical insights that clarify the strengths and weaknesses of existing approaches and reveal the increased challenges of detection in knowledge-intensive domains and with more advanced models. To the best of our knowledge, FaithCoT-Bench establishes the first comprehensive benchmark for instance-level CoT faithfulness, setting a solid basis for future research toward more interpretable and trustworthy reasoning in LLMs.

cs.AI↗

arXiv2026

Unleashing Low-Bit Inference on Ascend NPUs: A Comprehensive Evaluation of HiFloat Formats

As LLMs scale, low-bit floating-point formats like MXFP and NVFP4 offer new opportunities for precision and efficiency. In this work, we evaluate HiFloat (HiF8 and HiF4), a family of formats tailored for Ascend NPUs. Through rigorous comparison across weight-activation and KV-cache tasks, we provide three key insights: (1) INT8 suits narrow-range data, while floating-point formats excel with high-variance data; (2) in 4-bit regimes, HiF4's hierarchical scaling prevents the accuracy collapse seen in integer formats; and (3) HiFloat is fully compatible with state-of-the-art post-training quantization frameworks. Overall, HiFloat provides a solution for high-efficiency LLM inference on NPUs.

cs.CL↗

arXiv2026

Deepfake Word Detection by Next-token Prediction using Fine-tuned Whisper

Deepfake speech utterances can be forged by replacing one or more words in a bona fide utterance with semantically different words synthesized with speech-generative models. While a dedicated synthetic word detector could be developed, we developed a cost-effective method that fine-tunes a pre-trained Whisper model to detect synthetic words while transcribing the input utterance via next-token prediction. We further investigate using partially vocoded utterances as the fine-tuning data, thus reducing the cost of data collection. Our experiments demonstrate that, on in-domain test data, the fine-tuned Whisper yields low synthetic-word detection error rates and transcription error rates. On out-of-domain test data with synthetic words produced with unseen speech-generative models, the fine-tuned Whisper remains on par with a dedicated ResNet-based detection model; however, the overall performance degradation calls for strategies to improve its generalization capability.

eess.AS↗

arXiv2026

TokenCLIP: Token-wise Prompt Learning for Zero-shot Anomaly Detection

Adapting CLIP for anomaly detection on unseen objects has shown strong potential in a zero-shot manner. However, existing methods typically rely on a single textual space to align with visual semantics across diverse objects and domains. The indiscriminate alignment hinders the model from accurately capturing varied anomaly semantics. We propose TokenCLIP, a token-wise adaptation framework that enables dynamic alignment between visual and learnable textual spaces for fine-grained anomaly learning. Rather than mapping all visual tokens to a single, token-agnostic textual space, TokenCLIP aligns each token with a customized textual subspace that represents its visual characteristics. Explicitly assigning a unique learnable textual space to each token is computationally intractable and prone to insufficient optimization. We instead expand the token-agnostic textual space into a set of orthogonal subspaces, and then dynamically assign each token to a subspace combination guided by semantic affinity, which jointly supports customized and efficient token-wise adaptation. To this end, we formulate dynamic alignment as an optimal transport problem, where all visual tokens in an image are transported to textual subspaces based on semantic similarity. The transport constraints of OT ensure sufficient optimization across subspaces and encourage them to focus on different semantics. Solving the problem yields a transport plan that adaptively assigns each token to semantically relevant subspaces. A top-k masking is then applied to sparsify the plan and specialize subspaces for distinct visual regions. Extensive experiments demonstrate the superiority of TokenCLIP.

cs.CV↗

arXiv2026

RhythmBERT: A Self-Supervised Language Model Based on Latent Representations of ECG Waveforms for Heart Disease Detection

Electrocardiogram (ECG) analysis is crucial for diagnosing heart disease, but most self-supervised learning methods treat ECG as a generic time series, overlooking physiologic semantics and rhythm-level structure. Existing contrastive methods utilize augmentations that distort morphology, whereas generative approaches employ fixed-window segmentation, which misaligns cardiac cycles. To address these limitations, we propose RhythmBERT, a generative ECG language model that considers ECG as a language paradigm by encoding P, QRS, and T segments into symbolic tokens via autoencoder-based latent representations. These discrete tokens capture rhythm semantics, while complementary continuous embeddings retain fine-grained morphology, enabling a unified view of waveform structure and rhythm. RhythmBERT is pretrained on approximately 800,000 unlabeled ECG recordings with a masked prediction objective, allowing it to learn contextual representations in a label-efficient manner. Evaluations show that despite using only a single lead, RhythmBERT achieves comparable or superior performance to strong 12-lead baselines. This generalization extends from prevalent conditions such as atrial fibrillation to clinically challenging cases such as subtle ST-T abnormalities and myocardial infarction. Our results suggest that considering ECG as structured language offers a scalable and physiologically aligned pathway for advancing cardiac analysis.

cs.LG↗

arXiv2026

A Morphology Catalog of Galaxies in CEERS: Evolution in the Size and Color Gradients of Galaxies Since Cosmic Dawn

We present measurements of morphological parameters from fitting 53,885 galaxies detected to a magnitude limit of F356W$< 28.5$ in the CEERS NIRCam imaging with galfit in six broadband filters: F115W, F150W, F200W, F277W, F356W, and F444W. We provide a public catalog of Sérsic index, effective semi-major axis, axis ratio, integrated magnitude, and position angle for these galaxies in each of the filters. Uncertainties in the measured parameters are estimated from simulated galaxies that have similar noise and background properties as the observed galaxies. We compare our measurements with those in the CANDELS/EGS field measured with HST/WFC3 and find that the sizes agree to within 0.09 dex and the Sérsic indices agree to within 0.13 dex. We further present the evolution in the size-mass relation, and find that the evolution to $z\sim9$ is consistent with previous results derived at lower redshift. Finally, we look at the color gradients of galaxies at $1 2.5$), the color gradients are nearly flat with no dependence on mass, indicating that the stellar populations are more uniform throughout. The structural measurements presented are accurate to $20\%$ or better for most galaxies with F356W $<27.0$ mag and will enable further studies of galaxy morphology to $z\sim10$.

astro-ph.GA↗

arXiv2026

Feeding the Dead: Neutral Gas Inflow in a Long-Quenched Ancient Massive Galaxy at $z\sim 2.7$ Observed with JWST/NIRSpec

We report the spectroscopic detection of neutral gas inflow into a massive ($M_* \simeq 4\times 10^{10} M_\odot$) quiescent galaxy observed at $z_{\rm{spec}} = 2.6576$ with JWST. From the redshifted absorption of the NaI doublet at $λλ5890, 5896 $ Ang, we estimate an inflow velocity $v=278^{+79}_{-79}$ km s$^{-1}$ and a column density $\log(N_{NaI}/\rm{cm^2}) = 13.02^{+0.03}_{-0.03}$. We derive the inflowing mass of the gas $M_{in} = 1.6^{+0.1}_{-0.1} \times 10^8 M_\odot$ and rate $\dot{M}_{in} = 19^{+6}_{-7} \, M_\odot \, \rm{yr}^{-1}$. The presence of several surrounding galaxies suggests that the galaxy may be accreting gas from nearby companions. However, we cannot confirm it with current data and the intergalactic medium or cosmic filaments are also viable sources of the inflowing gas. Despite the ongoing inflow, the galaxy remains quiescent, with an upper limit to the star formation rate of $0.2 \, M_\odot \, \rm{yr}^{-1}$. Moreover, its star formation history suggests that the galaxy has remained quiescent during the past $\sim1$ Gyr, with half of its stars formed by redshift $z_{50}=11^{+18}_{-3}$. We discuss that the inflow is not massive, dense, or long-lived enough to ignite significant star formation, or it is fueling low-level AGN activity instead. This is direct evidence that quiescent galaxies can accrete cold gas after their quenching while keeping their star formation subdued. Follow-up observations with JWST and ALMA will be needed to constraint the nature of the inflowing gas.

astro-ph.GA↗