arXiv subjects

Yi Yang

Publications and source records attributed to Yi Yang.

At least 253 records · Page 14Linked to original sources

arXiv2025

DeMoGen: Towards Decompositional Human Motion Generation with Energy-Based Diffusion Models

Human motions are compositional: complex behaviors can be described as combinations of simpler primitives. However, existing approaches primarily focus on forward modeling, e.g., learning holistic mappings from text to motion or composing a complex motion from a set of motion concepts. In this paper, we consider the inverse perspective: decomposing a holistic motion into semantically meaningful sub-components. We propose DeMoGen, a compositional training paradigm for decompositional learning that employs an energy-based diffusion model. This energy formulation directly captures the composed distribution of multiple motion concepts, enabling the model to discover them without relying on ground-truth motions for individual concepts. Within this paradigm, we introduce three training variants to encourage a decompositional understanding of motion: 1. DeMoGen-Exp explicitly trains on decomposed text prompts; 2. DeMoGen-OSS performs orthogonal self-supervised decomposition; 3. DeMoGen-SC enforces semantic consistency between original and decomposed text embeddings. These variants enable our approach to disentangle reusable motion primitives from complex motion sequences. We also demonstrate that the decomposed motion concepts can be flexibly recombined to generate diverse and novel motions, generalizing beyond the training distribution. Additionally, we construct a text-decomposed dataset to support compositional training, serving as an extended resource to facilitate text-to-motion generation and motion composition.

cs.CV↗

arXiv2025

Optical parametric free-electron--photon quantum interaction

Optical parametric processes underpin quantum photonics, while free-electron--photon interactions offer agile pathways to generate nontrivial quantum photonic states. These threads have so far largely progressed independently, whereas placing free electrons in a driven nonlinear system can potentially activate coherent parametric interaction channels for joint state engineering of both types of particles. Here we unify these paradigms by developing a general theoretical framework for parametric free-electron--photon interactions in a nonlinear optical system driven by degenerate parametric down-conversion. Unlike free electrons in a linear bath, here they can couple to Bogoliubov quasiparticles through two detuned phase-matching channels, where the parametric process and free-electron interactions can quantum amplify each other. Seeding the interaction with squeezed vacuum yields gain-only or loss-only electron energy spectra, and enables electron-heralded squeezed Fock states; with bare vacuum, postselecting electron energy sidebands generates high-fidelity Schrödinger cat states. Our results show how optical parametric interactions can quantum shape free electrons and photons, potentially enabling a quantum parametric dielectric laser accelerator that mitigates the need for temporal phase synchronization, thereby allowing acceleration probabilities to approach unity even for phase-random electrons.

physics.optics↗

arXiv2025

A graphical framework for proving holographic entanglement entropy inequalities in multipartite systems

We present a graphical method for proving holographic entanglement entropy inequalities (HEIs) in general multipartite systems. By introducing a geometric representation of the entanglement structure, we develop a systematic approach that enables one to visualize and verify the validity of HEIs for any number of subsystems $n$. Several theorems are established to formalize this method, and explicit examples are provided for systems with $n = 4$ to $7$ entangled regions.

quant-ph↗

arXiv2025

ACE-Sync: An Adaptive Cloud-Edge Synchronization Framework for Communication-Efficient Large-Scale Distributed Model Training

Large-scale deep learning models impose substantial communication overh ead in distributed training, particularly in bandwidth-constrained or heterogeneous clo ud-edge environments. Conventional synchronous or fixed-compression techniques o ften struggle to balance communication cost, convergence stability, and model accura cy. To address these challenges, we propose ACE-Sync, an Adaptive Cloud-Edge Sy nchronization Framework that integrates (1) an attention-based gradient importance p redictor, (2) a differentiated parameter compression strategy, and (3) a hierarchical cl oud-edge coordination mechanism. ACE-Sync dynamically selects which parameter groups to synchronize and determines appropriate compression levels under per-devic e bandwidth budgets. A knapsack-based optimization strategy is adopted to maximize important gradient preservation while reducing redundant communication. Furthermo re, residual-based error compensation and device clustering ensure long-term converg ence and cross-device personalization. Experiments show that ACE-Sync substantiall y reduces communication overhead while maintaining competitive accuracy. Compar ed with FullSync, ACE-Sync lowers communication cost from 112.5 GB to 44.7 GB (a 60% reduction) and shortens convergence from 41 to 39 epochs. Despite aggressiv e communication reduction, ACE-Sync preserves high model quality, achieving 82. 1% Top-1 accuracy-only 0.3% below the full-synchronization baseline-demonstrating its efficiency and scalability for large-scale distributed training. These results indicate that ACE-Sync provides a scalable, communication-efficient, and accuracy-preservin g solution for large-scale cloud-edge distributed model training.

cs.DC↗

arXiv2025

Strong well-filteredness of upper topology on sup-complete posets

We first introduce and investigate a new class of $T_0$ spaces -- strong R-spaces, which are stronger than both R-spaces and strongly well-filtered spaces. It is proved that any sup-complete poset equipped with the upper topology is a strong R-space and the Hoare power space of a $T_0$-space is a strong R-space. Hence the upper topology on a sup-complete poset is strongly well-filtered and the Hoare power space of a $T_0$-space is strongly well-filtered, which answers two problems recently posed by Xu.

math.GN↗

arXiv2025

See Once, Then Act: Vision-Language-Action Model with Task Learning from One-Shot Video Demonstrations

Developing robust and general-purpose manipulation policies represents a fundamental objective in robotics research. While Vision-Language-Action (VLA) models have demonstrated promising capabilities for end-to-end robot control, existing approaches still exhibit limited generalization to tasks beyond their training distributions. In contrast, humans possess remarkable proficiency in acquiring novel skills by simply observing others performing them once. Inspired by this capability, we propose ViVLA, a generalist robotic manipulation policy that achieves efficient task learning from a single expert demonstration video at test time. Our approach jointly processes an expert demonstration video alongside the robot's visual observations to predict both the demonstrated action sequences and subsequent robot actions, effectively distilling fine-grained manipulation knowledge from expert behavior and transferring it seamlessly to the agent. To enhance the performance of ViVLA, we develop a scalable expert-agent pair data generation pipeline capable of synthesizing paired trajectories from easily accessible human videos, further augmented by curated pairs from publicly available datasets. This pipeline produces a total of 892,911 expert-agent samples for training ViVLA. Experimental results demonstrate that our ViVLA is able to acquire novel manipulation skills from only a single expert demonstration video at test time. Our approach achieves over 30% improvement on unseen LIBERO tasks and maintains above 35% gains with cross-embodiment videos. Real-world experiments demonstrate effective learning from human videos, yielding more than 38% improvement on unseen tasks.

cs.RO↗

arXiv2025

3DID: Direct 3D Inverse Design for Aerodynamics with Physics-Aware Optimization

Inverse design aims to design the input variables of a physical system to optimize a specified objective function, typically formulated as a search or optimization problem. However, in 3D domains, the design space grows exponentially, rendering exhaustive grid-based searches infeasible. Recent advances in deep learning have accelerated inverse design by providing powerful generative priors and differentiable surrogate models. Nevertheless, current methods tend to approximate the 3D design space using 2D projections or fine-tune existing 3D shapes. These approaches sacrifice volumetric detail and constrain design exploration, preventing true 3D design from scratch. In this paper, we propose a 3D Inverse Design (3DID) framework that directly navigates the 3D design space by coupling a continuous latent representation with a physics-aware optimization strategy. We first learn a unified physics-geometry embedding that compactly captures shape and physical field data in a continuous latent space. Then, we introduce a two-stage strategy to perform physics-aware optimization. In the first stage, a gradient-guided diffusion sampler explores the global latent manifold. In the second stage, an objective-driven, topology-preserving refinement further sculpts each candidate toward the target objective. This enables 3DID to generate high-fidelity 3D geometries, outperforming existing methods in both solution quality and design versatility.

cs.CV↗

arXiv2025

Development of a 15-Degree-of-Freedom Bionic Hand with Cable-Driven Transmission and Distributed Actuation

In robotic hand research, minimizing the number of actuators while maintaining human-hand-consistent dimensions and degrees of freedom constitutes a fundamental challenge. Drawing bio-inspiration from human hand kinematic configurations and muscle distribution strategies, this work proposes a novel 15-DoF dexterous robotic hand, with detailed analysis of its mechanical architecture, electrical system, and control system. The bionic hand employs a new tendon-driven mechanism, significantly reducing the number of motors required by traditional tendon-driven systems while enhancing motion performance and simplifying the mechanical structure. This design integrates five motors in the forearm to provide strong gripping force, while ten small motors are installed in the palm to support fine manipulation tasks. Additionally, a corresponding joint sensing and motor driving electrical system was developed to ensure efficient control and feedback. The entire system weighs only 1.4kg, combining lightweight and high-performance features. Through experiments, the bionic hand exhibited exceptional dexterity and robust grasping capabilities, demonstrating significant potential for robotic manipulation tasks.

cs.RO↗

arXiv2025

Line Polarization of Si II $λ$6355 Å in Type Ia Supernovae: A New Statistical Approach to Probe the Explosion Physics and Diversity

Spectropolarimetry provides a unique probe of ejecta asphericities, offering direct insights into the underlying explosion physics of Type Ia supernovae (SNe Ia). We analyze the statistical properties of pre-maximum spectropolarimetric data for 24 SNe Ia observed with VLT/FORS, focusing on the Si II $λ$6355 Åline. Previous studies have revealed a correlation between the peak Si II polarization degree and the expansion velocity. Here, we combine these observations with multi-dimensional non-LTE radiative transfer simulations. We consider two asphericity classes: (i) lopsided abundance distributions produced by off-center delayed-detonation transitions in near-$M_{Ch}$ white dwarfs or, for example, WD collisions (Class I), and (ii) global, axisymmetric density asphericities such as those arising from explosions of rapidly rotating WDs or mergers (Class II). Our model grid spans normal to subluminous SNe Ia and successfully reproduces the observed Si II velocity-polarization trend, with higher velocities associated with stronger asphericities. Consistent with observations, transitional SNe Ia and the faint end of the normal SNe Ia population show the highest Si II polarization and are best explained by Class I scenarios. In contrast, subluminous SNe Ia are dominated by Class II asphericities, characterized by lower Si II polarization but significant continuum polarization. The observed distribution of Si II polarization depends on both the observer's viewing angle $θ$ and the intrinsic asphericity. Statistical analysis of these spectropolarimetric snapshots enables the separation of Class I and Class II contributions and highlights the intrinsic diversity among SNe Ia. Our results imply viewing-angle-dependent luminosities in our local sample, which may have implications when using high-redshift SNe Ia as evidence for the need of non-standard cosmology.

astro-ph.HE↗

arXiv2025

3DIS: Depth-Driven Decoupled Instance Synthesis for Text-to-Image Generation

The increasing demand for controllable outputs in text-to-image generation has spurred advancements in multi-instance generation (MIG), allowing users to define both instance layouts and attributes. However, unlike image-conditional generation methods such as ControlNet, MIG techniques have not been widely adopted in state-of-the-art models like SD2 and SDXL, primarily due to the challenge of building robust renderers that simultaneously handle instance positioning and attribute rendering. In this paper, we introduce Depth-Driven Decoupled Instance Synthesis (3DIS), a novel framework that decouples the MIG process into two stages: (i) generating a coarse scene depth map for accurate instance positioning and scene composition, and (ii) rendering fine-grained attributes using pre-trained ControlNet on any foundational model, without additional training. Our 3DIS framework integrates a custom adapter into LDM3D for precise depth-based layouts and employs a finetuning-free method for enhanced instance-level attribute rendering. Extensive experiments on COCO-Position and COCO-MIG benchmarks demonstrate that 3DIS significantly outperforms existing methods in both layout precision and attribute rendering. Notably, 3DIS offers seamless compatibility with diverse foundational models, providing a robust, adaptable solution for advanced multi-instance generation. The code is available at: https://github.com/limuloo/3DIS.

cs.CV↗

arXiv2025

SR-GRPO: Stable Rank as an Intrinsic Geometric Reward for Large Language Model Alignment

Aligning Large Language Models (LLMs) with human preferences typically relies on external supervision, which faces critical limitations: human annotations are scarce and subjective, reward models are vulnerable to reward hacking, and self-evaluation methods suffer from prompt sensitivity and biases. In this work, we propose stable rank, an intrinsic, annotation-free quality signal derived from model representations. Stable rank measures the effective dimensionality of hidden states by computing the ratio of total variance to dominant-direction variance, capturing quality through how information distributes across representation dimensions. Empirically, stable rank achieves 84.04% accuracy on RewardBench and improves task accuracy by an average of 11.3 percentage points over greedy decoding via Best-of-N sampling. Leveraging this insight, we introduce Stable Rank Group Relative Policy Optimization (SR-GRPO), which uses stable rank as a reward signal for reinforcement learning. Without external supervision, SR-GRPO improves Qwen2.5-1.5B-Instruct by 10% on STEM and 19% on mathematical reasoning, outperforming both learned reward models and self-evaluation baselines. Our findings demonstrate that quality signals can be extracted from internal model geometry, offering a path toward scalable alignment without external supervision.

cs.CL↗

arXiv2025

HiMo: High-Speed Objects Motion Compensation in Point Clouds

LiDAR point cloud is essential for autonomous vehicles, but motion distortions from dynamic objects degrade the data quality. While previous work has considered distortions caused by ego motion, distortions caused by other moving objects remain largely overlooked, leading to errors in object shape and position. This distortion is particularly pronounced in high-speed environments such as highways and in multi-LiDAR configurations, a common setup for heavy vehicles. To address this challenge, we introduce HiMo, a pipeline that repurposes scene flow estimation for non-ego motion compensation, correcting the representation of dynamic objects in point clouds. During the development of HiMo, we observed that existing self-supervised scene flow estimators often produce degenerate or inconsistent estimates under high-speed distortion. We further propose SeFlow++, a real-time scene flow estimator that achieves state-of-the-art performance on both scene flow and motion compensation. Since well-established motion distortion metrics are absent in the literature, we introduce two evaluation metrics: compensation accuracy at a point level and shape similarity of objects. We validate HiMo through extensive experiments on Argoverse 2, ZOD, and a newly collected real-world dataset featuring highway driving and multi-LiDAR-equipped heavy vehicles. Our findings show that HiMo improves the geometric consistency and visual fidelity of dynamic objects in LiDAR point clouds, benefiting downstream tasks such as semantic segmentation and 3D detection. See https://kin-zhang.github.io/HiMo for more details.

cs.CV↗

arXiv2025

SN 2021ukt: A Transitional Supernova with a Short Plateau and Persistent Interaction

We present spectroscopic and photometric observations of supernova (SN) 2021ukt, a peculiar short-plateau object that was originally identified as a Type IIn SN and later underwent an unprecedented transition to a Type Ib (possibly Type IIb) SN. The early-time light curves of SN 2021ukt exhibit a ~25 day plateau. Such a short phase of hydrogen recombination suggests a rather thin H-rich outer envelope of the progenitor star. The relatively narrow Balmer emission lines in spectra of SN 2021ukt during the first week indicate the interaction between the expanding ejecta and the immediate circumstellar material (CSM). This Hα line is observed throughout its helium-rich ejecta-dominated phase and nebular phase, suggesting persistent interaction with a radially extended CSM profile. We explore the synthetic light-curve model among grids of parameters generated by MESA+STELLA. We also compare the spectrophotometric evolution of SN 2021ukt with several well-sampled supernovae that exhibit a short plateau and persistent ejecta-CSM interaction. An estimate of the progenitor mass of SN 2021ukt is made based on the flux ratio between [Ca II] λλ 7291, 7324 and [O I] λλ 6300, 6364 during its nebular phase. Our analysis suggests that the progenitor star of SN 2021ukt has a zero-age main-sequence (ZAMS) mass of about 12 solar masses, a mass of radioactive nickel-56 synthesized in the SN ejecta of about 0.04 solar masses, and a mass of the H-rich envelope of about 0.5 solar masses. This study adds to the growing sample of transitional supernovae, reinforcing evidence for a continuum of underrepresented progenitors whose evolutionary pathways lie between those of standard SN models.

astro-ph.SR↗

arXiv2025

AnchorFlow: Training-Free 3D Editing via Latent Anchor-Aligned Flows

Training-free 3D editing aims to modify 3D shapes based on human instructions without model finetuning. It plays a crucial role in 3D content creation. However, existing approaches often struggle to produce strong or geometrically stable edits, largely due to inconsistent latent anchors introduced by timestep-dependent noise during diffusion sampling. To address these limitations, we introduce AnchorFlow, which is built upon the principle of latent anchor consistency. Specifically, AnchorFlow establishes a global latent anchor shared between the source and target trajectories, and enforces coherence using a relaxed anchor-alignment loss together with an anchor-aligned update rule. This design ensures that transformations remain stable and semantically faithful throughout the editing process. By stabilizing the latent reference space, AnchorFlow enables more pronounced semantic modifications. Moreover, AnchorFlow is mask-free. Without mask supervision, it effectively preserves geometric fidelity. Experiments on the Eval3DEdit benchmark show that AnchorFlow consistently delivers semantically aligned and structurally robust edits across diverse editing types. Code is at https://github.com/ZhenglinZhou/AnchorFlow.

cs.CV↗

arXiv2025

ITS3D: Inference-Time Scaling for Text-Guided 3D Diffusion Models

We explore inference-time scaling in text-guided 3D diffusion models to enhance generative quality without additional training. To this end, we introduce ITS3D, a framework that formulates the task as an optimization problem to identify the most effective Gaussian noise input. The framework is driven by a verifier-guided search algorithm, where the search algorithm iteratively refines noise candidates based on verifier feedback. To address the inherent challenges of 3D generation, we introduce three techniques for improved stability, efficiency, and exploration capability. 1) Gaussian normalization is applied to stabilize the search process. It corrects distribution shifts when noise candidates deviate from a standard Gaussian distribution during iterative updates. 2) The high-dimensional nature of the 3D search space increases computational complexity. To mitigate this, a singular value decomposition-based compression technique is employed to reduce dimensionality while preserving effective search directions. 3) To further prevent convergence to suboptimal local minima, a singular space reset mechanism dynamically updates the search space based on diversity measures. Extensive experiments demonstrate that ITS3D enhances text-to-3D generation quality, which shows the potential of computationally efficient search methods in generative processes. The source code is available at https://github.com/ZhenglinZhou/ITS3D.

cs.CV↗

arXiv2025

The Faintest, Extremely Variable X-ray Tidal Disruption Event from a Supermassive Black Hole Binary?

Tidal disruption events (TDEs), which occur when stars enter the tidal radii of supermassive black holes (SMBHs) and are subsequently torn apart by their tidal forces, represent intriguing phenomena that stimulate growing research interest and pose an increasing number of puzzles in the era of time-domain astronomy. Here we report an unusual X-ray transient, XID 935, discovered in the 7 Ms Chandra Deep Field-South, the deepest X-ray survey ever. XID 935 experienced an overall X-ray dimming by a factor of more than 40 between 1999 and 2016. Not monotonically decreasing during this period, its X-ray luminosity increased by a factor $> 27$ within 2 months, from $L_{\rm 0.5-7\ keV}<10^{40.87}$ erg s$^{-1}$ (10 October 2014 -- 4 January 2015) to $L_{\rm 0.5-7\ keV}=10^{42.31\pm 0.20}$ erg s$^{-1}$ (16 March 2015). The X-ray position of XID 935 is located at the center of its host galaxy with a spectroscopic redshift of 0.251, whose optical spectra do not display emission characteristics associated with an active galactic nucleus. The peak 0.5--2.0 keV flux is the faintest among all the X-ray-selected TDE candidates to date. Thanks to a total exposure of $\sim 9.5$ Ms in the X-ray bands, we manage to secure relatively well-sampled, 20-year-long X-ray light curves of this deepest X-ray-selected TDE candidate. We find that a partial TDE model could not explain the main declining trend. An SMBH binary TDE model is in acceptable accordance with the light curves of XID 935; however, it fails to match short-timescale fluctuations exactly. Therefore, the exceptional observational features of XID 935 provide a key benchmark for refining quantitative TDE models and simulations.

astro-ph.HE↗

arXiv2025

Are Image-to-Video Models Good Zero-Shot Image Editors?

Large-scale video diffusion models show strong world simulation and temporal reasoning abilities, but their use as zero-shot image editors remains underexplored. We introduce IF-Edit, a tuning-free framework that repurposes pretrained image-to-video diffusion models for instruction-driven image editing. IF-Edit addresses three key challenges: prompt misalignment, redundant temporal latents, and blurry late-stage frames. It includes (1) a chain-of-thought prompt enhancement module that transforms static editing instructions into temporally grounded reasoning prompts; (2) a temporal latent dropout strategy that compresses frame latents after the expert-switch point, accelerating denoising while preserving semantic and temporal coherence; and (3) a self-consistent post-refinement step that sharpens late-stage frames using a short still-video trajectory. Experiments on four public benchmarks, covering non-rigid editing, physical and temporal reasoning, and general instruction edits, show that IF-Edit performs strongly on reasoning-centric tasks while remaining competitive on general-purpose edits. Our study provides a systematic view of video diffusion models as image editors and highlights a simple recipe for unified video-image generative reasoning.

cs.CV↗

arXiv2025

BideDPO: Conditional Image Generation with Simultaneous Text and Condition Alignment

Conditional image generation enhances text-to-image synthesis with structural, spatial, or stylistic priors, but current methods face challenges in handling conflicts between sources. These include 1) input-level conflicts, where the conditioning image contradicts the text prompt, and 2) model-bias conflicts, where generative biases disrupt alignment even when conditions match the text. Addressing these conflicts requires nuanced solutions, which standard supervised fine-tuning struggles to provide. Preference-based optimization techniques like Direct Preference Optimization (DPO) show promise but are limited by gradient entanglement between text and condition signals and lack disentangled training data for multi-constraint tasks. To overcome this, we propose a bidirectionally decoupled DPO framework (BideDPO). Our method creates two disentangled preference pairs-one for the condition and one for the text-to reduce gradient entanglement. The influence of pairs is managed using an Adaptive Loss Balancing strategy for balanced optimization. We introduce an automated data pipeline to sample model outputs and generate conflict-aware data. This process is embedded in an iterative optimization strategy that refines both the model and the data. We construct a DualAlign benchmark to evaluate conflict resolution between text and condition. Experiments show BideDPO significantly improves text success rates (e.g., +35%) and condition adherence. We also validate our approach using the COCO dataset. Project Pages: https://limuloo.github.io/BideDPO/.

cs.CV↗