arXiv subjects

Jun Cheng

Publications and source records attributed to Jun Cheng.

At least 55 records · Page 3Linked to original sources

arXiv2025

HMAD: Advancing E2E Driving with Anchored Offset Proposals and Simulation-Supervised Multi-target Scoring

End-to-end autonomous driving faces persistent challenges in both generating diverse, rule-compliant trajectories and robustly selecting the optimal path from these options via learned, multi-faceted evaluation. To address these challenges, we introduce HMAD, a framework integrating a distinctive Bird's-Eye-View (BEV) based trajectory proposal mechanism with learned multi-criteria scoring. HMAD leverages BEVFormer and employs learnable anchored queries, initialized from a trajectory dictionary and refined via iterative offset decoding (inspired by DiffusionDrive), to produce numerous diverse and stable candidate trajectories. A key innovation, our simulation-supervised scorer module, then evaluates these proposals against critical metrics including no at-fault collisions, drivable area compliance, comfortableness, and overall driving quality (i.e., extended PDM score). Demonstrating its efficacy, HMAD achieves a 44.5% driving score on the CVPR 2025 private test set. This work highlights the benefits of effectively decoupling robust trajectory generation from comprehensive, safety-aware learned scoring for advanced autonomous driving.

cs.CV↗

arXiv2025

Compensate Quantization Errors+: Quantized Models Are Inquisitive Learners

The quantization of large language models (LLMs) has been a prominent research area aimed at enabling their lightweight deployment in practice. Existing research about LLM's quantization has mainly explored the interplay between weights and activations, or employing auxiliary components while neglecting the necessity of adjusting weights during quantization. Consequently, original weight distributions frequently fail to yield desired results after round-to-nearest (RTN) quantization. Even though incorporating techniques such as mixed precision and low-rank error approximation in LLM's quantization can yield improved results, they inevitably introduce additional computational overhead. On the other hand, traditional techniques for weight quantization, such as Generative Post-Training Quantization, rely on manually tweaking weight distributions to minimize local errors, but they fall short of achieving globally optimal outcomes. Although the recently proposed Learnable Singular-value Increment improves global weight quantization by modifying weight distributions, it disrupts the original distribution considerably. This introduces pronounced bias toward the training data and can degrade downstream task performance. In this paper, we introduce Singular-value Diagonal Expansion, a more nuanced approach to refining weight distributions to achieve better quantization alignment. Furthermore, we introduce Cross-layer Learning that improves overall quantization outcomes by distributing errors more evenly across layers. Our plug-and-play weight-quantization methods demonstrate substantial performance improvements over state-of-the-art approaches, including OmniQuant, DuQuant, and PrefixQuant.

cs.CL↗

arXiv2025

LiftFeat: 3D Geometry-Aware Local Feature Matching

Robust and efficient local feature matching plays a crucial role in applications such as SLAM and visual localization for robotics. Despite great progress, it is still very challenging to extract robust and discriminative visual features in scenarios with drastic lighting changes, low texture areas, or repetitive patterns. In this paper, we propose a new lightweight network called \textit{LiftFeat}, which lifts the robustness of raw descriptor by aggregating 3D geometric feature. Specifically, we first adopt a pre-trained monocular depth estimation model to generate pseudo surface normal label, supervising the extraction of 3D geometric feature in terms of predicted surface normal. We then design a 3D geometry-aware feature lifting module to fuse surface normal feature with raw 2D descriptor feature. Integrating such 3D geometric feature enhances the discriminative ability of 2D feature description in extreme conditions. Extensive experimental results on relative pose estimation, homography estimation, and visual localization tasks, demonstrate that our LiftFeat outperforms some lightweight state-of-the-art methods. Code will be released at : https://github.com/lyp-deeplearning/LiftFeat.

cs.CV↗

arXiv2025

A Physics-Inspired Deep Learning Framework with Polar Coordinate Attention for Ptychographic Imaging

Ptychographic imaging confronts inherent challenges in applying deep learning for phase retrieval from diffraction patterns. Conventional neural architectures, both convolutional neural networks and Transformer-based methods, are optimized for natural images with Euclidean spatial neighborhood-based inductive biases that exhibit geometric mismatch with the concentric coherent patterns characteristic of diffraction data in reciprocal space. In this paper, we present PPN, a physics-inspired deep learning network with Polar Coordinate Attention (PoCA) for ptychographic imaging, that aligns neural inductive biases with diffraction physics through a dual-branch architecture separating local feature extraction from non-local coherence modeling. It consists of a PoCA mechanism that replaces Euclidean spatial priors with physically consistent radial-angular correlations. PPN outperforms existing end-to-end models, with spectral and spatial analysis confirming its greater preservation of high-frequency details. Notably, PPN maintains robust performance compared to iterative methods even at low overlap ratios, making it well suited for high-throughput imaging in real-world acquisition scenarios for samples with consistent structural characteristics.

physics.optics↗

arXiv2025

Redox chemistry meets semiconductor defect physics

Understanding how the electronic structure of electrodes influences electrocatalytic reactions has been a longstanding topic in the electrochemistry community, with predominant attention paid to metallic electrodes. In this work, we present a defect physics perspective on the effect of semiconductor band structure on electrochemical redox reactions. Specifically, the Haldane-Anderson model, originally developed to study multiple charge states of transition-metal defects in semiconductors, is extended to describe electrochemical redox reactions by incorporating the solvent effect, inspired by the Holstein model. The solvent coordinate and the actual charge on the redox species in reduced and oxidized states are assumed to be instant equilibrium, and the transitions between these states are defined by the framework of Green's function. With these treatments, we treat the charge state transition in a self-consistent manner. We first confirm that this self-consistent approach is essential to accurately depict the hybridization effect of band structure by comparing the model-calculated ionization potential (IP), electron affinity (EA), and redox potential of the species with those obtained from density functional theory (DFT) calculations. Next, we illustrate how this self-consistent treatment enhances our understanding of the catalytic activities of semiconductor electrodes and the source of asymmetry in reorganization energies, which is often observed in prior ab initio molecular dynamics (AIMD) simulations. Additionally, we discuss how band structure impacts redox reactions in the strong coupling limit. Finally, we compare our work with other relevant studies in the literature.

physics.chem-ph↗

arXiv2025

Decoding the Competing Effects of Dynamic Solvation Structures on Nuclear Magnetic Resonance Chemical Shifts of Battery Electrolytes via Machine Learning

Understanding the solvation structure of electrolytes is critical for optimizing the electrochemical performance of rechargeable batteries, as it directly influences properties such as ionic conductivity, viscosity, and electrochemical stability. The highly complex structures and strong interactions in high-concentration electrolytes make accurate modeling and interpretation of their ``structure-property" relationships even more challenging with spectroscopic methods. In this study, we present a machine learning-based approach to predict dynamic $^7$Li NMR chemical shifts in LiFSI/DME electrolyte solutions. Additionally, we provide a comprehensive structural analysis to interpret the observed chemical shift behavior in our experiments, particularly the abrupt changes in $^7$Li chemical shifts at high concentrations. Using advanced modeling techniques, we quantitatively establish the relationship between molecular structure and NMR spectra, offering critical insights into solvation structure assignments. Our findings reveal the coexistence of two competing local solvation structures that shift in dominance as electrolyte concentration approaches the concentrated limit, leading to anomalous reverse of $^7$Li NMR chemical shift in our experiment. This work provides a detailed molecular-level understanding of the intricate solvation structures probed by NMR spectroscopy, leading the way for enhanced electrolyte design.

physics.chem-ph↗

arXiv2025

The Tenth NTIRE 2025 Image Denoising Challenge Report

This paper presents an overview of the NTIRE 2025 Image Denoising Challenge (σ = 50), highlighting the proposed methodologies and corresponding results. The primary objective is to develop a network architecture capable of achieving high-quality denoising performance, quantitatively evaluated using PSNR, without constraints on computational complexity or model size. The task assumes independent additive white Gaussian noise (AWGN) with a fixed noise level of 50. A total of 290 participants registered for the challenge, with 20 teams successfully submitting valid results, providing insights into the current state-of-the-art in image denoising.

cs.CV↗

arXiv2025

EchoONE: Segmenting Multiple echocardiography Planes in One Model

In clinical practice of echocardiography examinations, multiple planes containing the heart structures of different view are usually required in screening, diagnosis and treatment of cardiac disease. AI models for echocardiography have to be tailored for each specific plane due to the dramatic structure differences, thus resulting in repetition development and extra complexity. Effective solution for such a multi-plane segmentation (MPS) problem is highly demanded for medical images, yet has not been well investigated. In this paper, we propose a novel solution, EchoONE, for this problem with a SAM-based segmentation architecture, a prior-composable mask learning (PC-Mask) module for semantic-aware dense prompt generation, and a learnable CNN-branch with a simple yet effective local feature fusion and adaption (LFFA) module for SAM adapting. We extensively evaluated our method on multiple internal and external echocardiography datasets, and achieved consistently state-of-the-art performance for multi-source datasets with different heart planes. This is the first time that the MPS problem is solved in one model for echocardiography data. The code will be available at https://github.com/a2502503/EchoONE.

cs.CV↗

arXiv2025

Auto-resolving the atomic structure at van der Waals interfaces using a generative model

The high-resolution visualization of atomic structures is significant for understanding the relationship between the microscopic configurations and macroscopic properties of materials. However, a rapid, accurate, and robust approach to automatically resolve complex patterns in atomic-resolution microscopy remains difficult to implement. Here, we present a Trident strategy-enhanced disentangled representation learning method (a generative model), which utilizes a few unlabelled experimental images with abundant low-cost simulated images to generate a large corpus of annotated simulation data that closely resembles experimental results, producing a high-quality large-volume training dataset. A structural inference model is then trained via a residual neural network which can directly deduce the interlayer slip and rotation of diversified and complicated stacking patterns at van der Waals interfaces with picometer-scale accuracy across various materials (e.g. MoS2, WS2, ReS2, ReSe2, and 1T'-MoTe2) with different layer numbers , demonstrating robustness to defects, imaging quality, and surface contaminations. The framework can also identify pattern transition interfaces, quantify subtle motif variations, and discriminate moire patterns that are difficult to distinguish in frequency domains. Finally, the high-throughput processing ability of our method provides insights into a van der Waals epitaxy mode where various thermodynamically favorable slip stackings can coexist.

cond-mat.mtrl-sci↗

arXiv2025

Beyond Existance: Fulfill 3D Reconstructed Scenes with Pseudo Details

The emergence of 3D Gaussian Splatting (3D-GS) has significantly advanced 3D reconstruction by providing high fidelity and fast training speeds across various scenarios. While recent efforts have mainly focused on improving model structures to compress data volume or reduce artifacts during zoom-in and zoom-out operations, they often overlook an underlying issue: training sampling deficiency. In zoomed-in views, Gaussian primitives can appear unregulated and distorted due to their dilation limitations and the insufficient availability of scale-specific training samples. Consequently, incorporating pseudo-details that ensure the completeness and alignment of the scene becomes essential. In this paper, we introduce a new training method that integrates diffusion models and multi-scale training using pseudo-ground-truth data. This approach not only notably mitigates the dilation and zoomed-in artifacts but also enriches reconstructed scenes with precise details out of existing scenarios. Our method achieves state-of-the-art performance across various benchmarks and extends the capabilities of 3D reconstruction beyond training datasets.

cs.GR↗

arXiv2025

Bootstrap-GS: Self-Supervised Augmentation for High-Fidelity Gaussian Splatting

Recent advancements in 3D Gaussian Splatting (3D-GS) have established new benchmarks for rendering quality and efficiency in 3D reconstruction. However, 3D-GS faces critical limitations when generating novel views that significantly deviate from those encountered during training. Moreover, issues such as dilation and aliasing arise during zoom operations. These challenges stem from a fundamental issue: training sampling deficiency. In this paper, we introduce a bootstrapping framework to address this problem. Our approach synthesizes pseudo-ground truth from novel views that align with the limited training set and reintegrates these synthesized views into the training pipeline. Experimental results demonstrate that our bootstrapping technique not only reduces artifacts but also improves quantitative metrics. Furthermore, our technique is highly adaptable, allowing various Gaussian-based method to benefit from its integration.

cs.GR↗

arXiv2025

Pay Attention to What You Need

Although large language models (LLMs) have achieved significant success in natural language processing, they still struggle with long-context comprehension. Traditional approaches to mitigating this issue typically rely on fine-tuning or retraining, which is both resource-intensive and challenging to deploy in lightweight industrial settings. In this paper, we investigate the potential to accomplish this without any additional resources. Through an in-depth study of the attention mechanism in LLMs, we propose a method called Scaled ReAttention (SRA) to strengthen LLMs' ability to interpret and retrieve information by strategically manipulating their attention scores during inference. Through extensive experiments, we demonstrate that integrating SRA significantly boosts LLMs' performance on a variety of downstream tasks, highlighting its practical potential for enhancing language understanding without incurring the overhead of traditional training.

cs.CL↗

arXiv2024

Scaling policy iteration based reinforcement learning for unknown discrete-time linear systems

In optimal control problem, policy iteration (PI) is a powerful reinforcement learning (RL) tool used for designing optimal controller for the linear systems. However, the need for an initial stabilizing control policy significantly limits its applicability. To address this constraint, this paper proposes a novel scaling technique, which progressively brings a sequence of stable scaled systems closer to the original system, enabling the acquisition of stable control gain. Based on the designed scaling update law, we develop model-based and model-free scaling policy iteration (SPI) algorithms for solving the optimal control problem for discrete-time linear systems, in both known and completely unknown system dynamics scenarios. Unlike existing works on PI based RL, the SPI algorithms do not necessitate an initial stabilizing gain to initialize the algorithms, they can achieve the optimal control under any initial control gain. Finally, the numerical results validate the theoretical findings and confirm the effectiveness of the algorithms.

math.OC↗

arXiv2024

Diffusion Priors for Variational Likelihood Estimation and Image Denoising

Real-world noise removal is crucial in low-level computer vision. Due to the remarkable generation capabilities of diffusion models, recent attention has shifted towards leveraging diffusion priors for image restoration tasks. However, existing diffusion priors-based methods either consider simple noise types or rely on approximate posterior estimation, limiting their effectiveness in addressing structured and signal-dependent noise commonly found in real-world images. In this paper, we build upon diffusion priors and propose adaptive likelihood estimation and MAP inference during the reverse diffusion process to tackle real-world noise. We introduce an independent, non-identically distributed likelihood combined with the noise precision (inverse variance) prior and dynamically infer the precision posterior using variational Bayes during the generation process. Meanwhile, we rectify the estimated noise variance through local Gaussian convolution. The final denoised image is obtained by propagating intermediate MAP solutions that balance the updated likelihood and diffusion prior. Additionally, we explore the local diffusion prior inherent in low-resolution diffusion models, enabling direct handling of high-resolution noisy images. Extensive experiments and analyses on diverse real-world datasets demonstrate the effectiveness of our method. Code is available at https://github.com/HUST-Tan/DiffusionVI.

cs.CV↗

arXiv2024

Progressive Retinal Image Registration via Global and Local Deformable Transformations

Retinal image registration plays an important role in the ophthalmological diagnosis process. Since there exist variances in viewing angles and anatomical structures across different retinal images, keypoint-based approaches become the mainstream methods for retinal image registration thanks to their robustness and low latency. These methods typically assume the retinal surfaces are planar, and adopt feature matching to obtain the homography matrix that represents the global transformation between images. Yet, such a planar hypothesis inevitably introduces registration errors since retinal surface is approximately curved. This limitation is more prominent when registering image pairs with significant differences in viewing angles. To address this problem, we propose a hybrid registration framework called HybridRetina, which progressively registers retinal images with global and local deformable transformations. For that, we use a keypoint detector and a deformation network called GAMorph to estimate the global transformation and local deformable transformation, respectively. Specifically, we integrate multi-level pixel relation knowledge to guide the training of GAMorph. Additionally, we utilize an edge attention module that includes the geometric priors of the images, ensuring the deformation field focuses more on the vascular regions of clinical interest. Experiments on two widely-used datasets, FIRE and FLoRI21, show that our proposed HybridRetina significantly outperforms some state-of-the-art methods. The code is available at https://github.com/lyp-deeplearning/awesome-retinal-registration.

cs.CV↗

arXiv2024

The Sampling-Gaussian for stereo matching

The soft-argmax operation is widely adopted in neural network-based stereo matching methods to enable differentiable regression of disparity. However, network trained with soft-argmax is prone to being multimodal due to absence of explicit constraint to the shape of the probability distribution. Previous methods leverages Laplacian distribution and cross-entropy for training but failed to effectively improve the accuracy and even compromises the efficiency of the network. In this paper, we conduct a detailed analysis of the previous distribution-based methods and propose a novel supervision method for stereo matching, Sampling-Gaussian. We sample from the Gaussian distribution for supervision. Moreover, we interpret the training as minimizing the distance in vector space and propose a combined loss of L1 loss and cosine similarity loss. Additionally, we leveraged bilinear interpolation to upsample the cost volume. Our method can be directly applied to any soft-argmax-based stereo matching method without a reduction in efficiency. We have conducted comprehensive experiments to demonstrate the superior performance of our Sampling-Gaussian. The experimental results prove that we have achieved better accuracy on five baseline methods and two datasets. Our method is easy to implement, and the code is available online.

cs.CV↗

arXiv2024

Revealing the molecular structures of a-Al2O3(0001)-water interface by machine learning based computational vibrational spectroscopy

Solid-water interfaces are crucial to many physical and chemical processes and are extensively studied using surface-specific sum-frequency generation (SFG) spectroscopy. To establish clear correlations between specific spectral signatures and distinct interfacial water structures, theoretical calculations using molecular dynamics (MD) simulations are required. These MD simulations typically need relatively long trajectories (a few nanoseconds) to achieve reliable SFG response function calculations via the dipole-polarizability time correlation function. However, the requirement for long trajectories limits the use of computationally expensive techniques such as ab initio MD (AIMD) simulations, particularly for complex solid-water interfaces. In this work, we present a pathway for calculating vibrational spectra (IR, Raman, SFG) of solid-water interfaces using machine learning (ML)-accelerated methods. We employ both the dipole moment-polarizability correlation function and the surface-specific velocity-velocity correlation function approaches to calculate SFG spectra. Our results demonstrate the successful acceleration of AIMD simulations and the calculation of SFG spectra using ML methods. This advancement provides an opportunity to calculate SFG spectra for the complicated solid-water systems more rapidly and at a lower computational cost with the aid of ML.

cond-mat.mtrl-sci↗

arXiv2024

Towards a Unified Benchmark and Framework for Deep Learning-Based Prediction of Nuclear Magnetic Resonance Chemical Shifts

The study of structure-spectrum relationships is essential for spectral interpretation, impacting structural elucidation and material design. Predicting spectra from molecular structures is challenging due to their complex relationships. Herein, we introduce NMRNet, a deep learning framework using the SE(3) Transformer for atomic environment modeling, following a pre-training and fine-tuning paradigm. To support the evaluation of NMR chemical shift prediction models, we have established a comprehensive benchmark based on previous research and databases, covering diverse chemical systems. Applying NMRNet to these benchmark datasets, we achieve state-of-the-art performance in both liquid-state and solid-state NMR datasets, demonstrating its robustness and practical utility in real-world scenarios. This marks the first integration of solid and liquid state NMR within a unified model architecture, highlighting the need for domainspecific handling of different atomic environments. Our work sets a new standard for NMR prediction, advancing deep learning applications in analytical and structural chemistry.

physics.comp-ph↗