arXiv subjects

Yuhao Wan

Publications and source records attributed to Yuhao Wan.

11 recordsLinked to original sources

arXiv2026

Direct 3D-Aware Object Insertion via Decomposed Visual Proxies

Object insertion aims to seamlessly composite a reference object into a specified region of a background image. Recent diffusion-based methods achieve high visual quality but formulate insertion as a simple 2D inpainting task, providing no explicit control over the object's 3D pose and limiting their practical applicability. We propose DIRECT (Decomposed Injection for Reference Composition and Target-integration), a novel framework that integrates interactive pose manipulation with high-fidelity 2D image synthesis to enable pose-controllable object insertion. Our method decomposes the insertion conditions into three complementary components: appearance guidance capturing visual details from the reference object, geometry guidance derived from the user-adjusted 3D proxy, and context guidance from the target background. By injecting them through separate pathways, DIRECT avoids feature entanglement and simultaneously preserves reference appearance, follows the user-specified pose, and adapts the object to the target scene. We also introduce an automated data construction pipeline to improve the diversity and quality of training data. Experiments show that DIRECT outperforms previous methods in both geometric controllability and visual quality.

cs.CV↗

arXiv2026

Pose-Aware Diffusion for 3D Generation

Generating pose-aligned 3D objects is challenging due to the spatial mismatches and transformation ambiguities inherent in decoupled canonical-then-rotate paradigms. To this end, we introduce Pose-Aware Diffusion (PAD), a novel end-to-end diffusion framework that synthesizes 3D geometry directly within the observation space. By unprojecting monocular depth into a partial point cloud and explicitly injecting it as a 3D geometric anchor, PAD abandons canonical assumptions to enforce rigorous spatial supervision. This native generation intrinsically resolves pose ambiguity, producing high-fidelity pose-aligned assets. Extensive experiments demonstrate that PAD achieves superior geometric alignment and image-to-3D correspondence compared to state-of-the-art methods. Additionally, PAD naturally extends to compositional 3D scene reconstruction via a simple union of independently generated objects, highlighting its robust ability to preserve precise spatial layouts.

cs.CV↗

arXiv2026

Trust but Verify: Adaptive Conditioning for Reference-Based Diffusion Super-Resolution via Implicit Reference Correlation Modeling

Recent works have explored reference-based super-resolution (RefSR) to mitigate hallucinations in diffusion-based image restoration. A key challenge is that real-world degradations make correspondences between low-quality (LQ) inputs and reference (Ref) images unreliable, requiring adaptive control of reference usage. Existing methods either ignore LQ-Ref correlations or rely on brittle explicit matching, leading to over-reliance on misleading references or under-utilization of valuable cues. To address this, we propose Ada-RefSR, a single-step diffusion framework guided by a "Trust but Verify" principle: reference information is leveraged when reliable and suppressed otherwise. Its core component, Adaptive Implicit Correlation Gating (AICG), employs learnable summary tokens to distill dominant reference patterns and capture implicit correlations with LQ features. Integrated into the attention backbone, AICG provides lightweight, adaptive regulation of reference guidance, serving as a built-in safeguard against erroneous fusion. Experiments on multiple datasets demonstrate that Ada-RefSR achieves a strong balance of fidelity, naturalness, and efficiency, while remaining robust under varying reference alignment.

cs.CV↗

arXiv2025

GeoWorld: Providing Full-frame Geometry Features to Facilitate 3D Scene Generation

Previous works that leverage video models for image-to-3D scene generation often suffer from geometric distortions and blurry content. Using video generation models to implicitly maintain geometric consistency according to a single-frame input is ineffective. In this paper, we present a two-stage method, named $\textbf{GeoWorld}$, that renovates the image-to-3D scene generation pipeline by providing full-frame geometry features. The first-stage video generation model, followed by a multi-view geometry model, produces $\textbf{full-frame}$ geometry features, which are then used as a mental draft of geometric conditions to aid the second-stage video-generation model. A geometric loss is proposed to impose real-world geometric constraints, and a geometry adaptation module is introduced to ensure the effective utilization of geometry features. Thanks to full-frame geometric modeling, the two smaller video models in our two-stage method can generate higher-fidelity 3D scenes than SOTA methods, while being even faster, e.g. 7.5$\times$ faster than Hunyuan-Voyager. Project page: https://peaes.github.io/GeoWorld.

cs.CV↗

arXiv2024

ControlSR: Taming Diffusion Models for Consistent Real-World Image Super Resolution

We present ControlSR, a new method that can tame Diffusion Models for consistent real-world image super-resolution (Real-ISR). Previous Real-ISR models mostly focus on how to activate more generative priors of text-to-image diffusion models to make the output high-resolution (HR) images look better. However, since these methods rely too much on the generative priors, the content of the output images is often inconsistent with the input LR ones. To mitigate the above issue, in this work, we tame Diffusion Models by effectively utilizing LR information to impose stronger constraints on the control signals from ControlNet in the latent space. We show that our method can produce higher-quality control signals, which enables the super-resolution results to be more consistent with the LR image and leads to clearer visual results. In addition, we also propose an inference strategy that imposes constraints in the latent space using LR information, allowing for the simultaneous improvement of fidelity and generative ability. Experiments demonstrate that our model can achieve better performance across multiple metrics on several test sets and generate more consistent SR results with LR images than existing methods. Our code is available at https://github.com/HVision-NKU/ControlSR.

cs.CV↗

arXiv2023

Anomalous universal quantum transport in 2D asymptotic quasiperiodic system

Quasiperiodic systems extend the concept of the Anderson transition to quasi-random and low-dimensional realms and have garnered widespread attention. Here, we propose the asymptotic quasiperiodic two-dimensional systems characterized by a sequence of rational magnetic fluxes, which have an irrational limit, and predict exotic universal wave-packet dynamics and transport phenomena associated with the asymptotic quasiperiodicity (AQP). The predictions unveil a class of multiple metal-insulator transitions driven by a novel interplay effect between AQP, relaxation, and finite temperature, which further reveals a unified and profound mechanism. Specifically, all the transport phenomena, including the wave-packet dynamics, the bulk and edge transport, are unified in the universal scaling laws unveiled in the asymptotic quasiperiodic regime, which demonstrate a nontrivial asymptotic connection to quantum phases in the quasiperiodic limit. Our work enriches the universal quantum transport phenomena, adds to the basic mechanisms underlying metal-insulator transitions, and opens up an avenue to study the exotic transport physics with AQP in high dimensions.

cond-mat.mes-hall↗

arXiv2023

Chain-of-Thought Hub: A Continuous Effort to Measure Large Language Models' Reasoning Performance

As large language models (LLMs) are continuously being developed, their evaluation becomes increasingly important yet challenging. This work proposes Chain-of-Thought Hub, an open-source evaluation suite on the multi-step reasoning capabilities of large language models. We are interested in this setting for two reasons: (1) from the behavior of GPT and PaLM model family, we observe that complex reasoning is likely to be a key differentiator between weaker and stronger LLMs; (2) we envisage large language models to become the next-generation computational platform and foster an ecosystem of LLM-based new applications, this naturally requires the foundation models to perform complex tasks that often involve the composition of linguistic and logical operations. Our approach is to compile a suite of challenging reasoning benchmarks to track the progress of LLMs. Our current results show that: (1) model scale clearly correlates with reasoning capabilities; (2) As of May 2023, Claude-v1.3 and PaLM-2 are the only two models that are comparable with GPT-4, while open-sourced models still lag behind; (3) LLaMA-65B performs closely to code-davinci-002, indicating that with successful further development such as reinforcement learning from human feedback (RLHF), it has great potential to be close to GPT-3.5-Turbo. Our results also suggest that for the open-source efforts to catch up, the community may focus more on building better base models and exploring RLHF.

cs.CL↗

arXiv2022

Two dimensional vertex-decorated Lieb lattice with exact mobility edges and robust flat bands

The mobility edge (ME) that marks the energy separating extended and localized states is a central concept in understanding the metal-insulator transition induced by disordered or quasiperiodic potentials. MEs have been extensively studied in three dimensional disorder systems and one-dimensional quasiperiodic systems. However, the studies of MEs in two dimensional (2D) systems are rare. Here we propose a class of 2D vertex-decorated Lieb lattice models with quasiperiodic potentials only acting on the vertices of a (extended) Lieb lattice. By mapping these models to the 2D Aubry-Andr\'{e} model, we obtain exact expressions of MEs and the localization lengths of localized states, and further demonstrate that the flat bands remain unaffected by the quasiperiodic potentials. Finally, we propose a highly feasible scheme to experimentally realize our model in a quantum dot array. Our results open the door to studying and realizing exact MEs and robust flat bands in 2D systems.

cond-mat.dis-nn↗

arXiv2022

Data Determines Distributional Robustness in Contrastive Language Image Pre-training (CLIP)

Contrastively trained language-image models such as CLIP, ALIGN, and BASIC have demonstrated unprecedented robustness to multiple challenging natural distribution shifts. Since these language-image models differ from previous training approaches in several ways, an important question is what causes the large robustness gains. We answer this question via a systematic experimental investigation. Concretely, we study five different possible causes for the robustness gains: (i) the training set size, (ii) the training distribution, (iii) language supervision at training time, (iv) language supervision at test time, and (v) the contrastive loss function. Our experiments show that the more diverse training distribution is the main cause for the robustness gains, with the other factors contributing little to no robustness. Beyond our experimental results, we also introduce ImageNet-Captions, a version of ImageNet with original text annotations from Flickr, to enable further controlled experiments of language-image training.

cs.CV↗

arXiv2021

Topological Magnetoelectric Response in Ferromagnetic Axion Insulators

Topological magnetoelectric effect (TME) is a hallmark response of the topological field theory, which provides a paradigm shift in the study of emergent topological phenomena. However, its direct observation is yet to be realized due to the demanding magnetic configuration required to gap all the surface states. Here, we theoretically propose that the axion insulators with a simple ferromagnetic configuration, such as MnBi2Te4/(Bi2Te3)n family, provide an ideal playground to realize TME. In a designed triangular prism geometry, all the surface states are magnetically gapped. Under a vertical electric field, the surface Hall currents give rise to a nearly half-quantized orbital moment, accompanied with a gapless chiral hinge mode circulating parallelly. Thus, the orbital magnetization from the two topological origins can be easily distinguished by reversing the electric field. Our work paves a new avenue towards the direct observation of TME in realistic axion-insulator materials.

cond-mat.mes-hall↗

arXiv2019

Mix and Match: Markov Chains & Mixing Times for Matching in Rideshare

Rideshare platforms such as Uber and Lyft dynamically dispatch drivers to match riders' requests. We model the dispatching process in rideshare as a Markov chain that takes into account the geographic mobility of both drivers and riders over time. Prior work explores dispatch policies in the limit of such Markov chains; we characterize when this limit assumption is valid, under a variety of natural dispatch policies. We give explicit bounds on convergence in general, and exact (including constants) convergence rates for special cases. Then, on simulated and real transit data, we show that our bounds characterize convergence rates -- even when the necessary theoretical assumptions are relaxed. Additionally these policies compare well against a standard reinforcement learning algorithm which optimizes for profit without any convergence properties.

cs.DS↗