Searcharxiv⌕ Search

arXiv subjects

Yufeng Wang

Publications and source records attributed to Yufeng Wang.

86 records · Page 5Linked to original sources
arXiv2023

PI-VEGAN: Physics Informed Variational Embedding Generative Adversarial Networks for Stochastic Differential Equations

We present a new category of physics-informed neural networks called physics informed variational embedding generative adversarial network (PI-VEGAN), that effectively tackles the forward, inverse, and mixed problems of stochastic differential equations. In these scenarios, the governing equations are known, but only a limited number of sensor measurements of the system parameters are available. We integrate the governing physical laws into PI-VEGAN with automatic differentiation, while introducing a variational encoder for approximating the latent variables of the actual distribution of the measurements. These latent variables are integrated into the generator to facilitate accurate learning of the characteristics of the stochastic partial equations. Our model consists of three components, namely the encoder, generator, and discriminator, each of which is updated alternatively employing the stochastic gradient descent algorithm. We evaluate the effectiveness of PI-VEGAN in addressing forward, inverse, and mixed problems that require the concurrent calculation of system parameters and solutions. Numerical results demonstrate that the proposed method achieves satisfactory stability and accuracy in comparison with the previous physics-informed generative adversarial network (PI-WGAN).

cs.LG↗
arXiv2023

Proton Collective Quantum Tunneling Induces Anomalous Thermal Conductivity of Ice under Pressure

Proton tunneling is believed to be non-local in ice but has never been shown experimentally. Here we measured thermal conductivity of ice under pressure up to 50 GPa and found it to increase with pressure until 20 GPa but decrease at higher pressures. We attribute this anomalous drop of thermal conductivity to the collective tunneling of protons at high pressures, supported by large-scale quantum molecular dynamics simulations. The collective tunneling loops span several picoseconds in time and are as large as nanometers in space, which match the phonon periods and wavelengths, leading to strong phonon scattering at high pressures. Our results show direct evidence of collective quantum motion existing in high-pressure ice and provide a new perspective to understanding the coupling between phonon propagation and atomic tunneling.

physics.chem-ph↗
arXiv2023

Missingness Augmentation: A General Approach for Improving Generative Imputation Models

Missing data imputation is a fundamental problem in data analysis, and many studies have been conducted to improve its performance by exploring model structures and learning procedures. However, data augmentation, as a simple yet effective method, has not received enough attention in this area. In this paper, we propose a novel data augmentation method called Missingness Augmentation (MisA) for generative imputation models. Our approach dynamically produces incomplete samples at each epoch by utilizing the generator's output, constraining the augmented samples using a simple reconstruction loss, and combining this loss with the original loss to form the final optimization objective. As a general augmentation technique, MisA can be easily integrated into generative imputation frameworks, providing a simple yet effective way to enhance their performance. Experimental results demonstrate that MisA significantly improves the performance of many recently proposed generative imputation models on a variety of tabular and image datasets. The code is available at \url{https://github.com/WYu-Feng/Missingness-Augmentation}.

cs.LG↗
arXiv2023

Less Emphasis on Difficult Layer Regions: Curriculum Learning for Singularly Perturbed Convection-Diffusion-Reaction Problems

Although Physics-Informed Neural Networks (PINNs) have been successfully applied in a wide variety of science and engineering fields, they can fail to accurately predict the underlying solution in slightly challenging convection-diffusion-reaction problems. In this paper, we investigate the reason of this failure from a domain distribution perspective, and identify that learning multi-scale fields simultaneously makes the network unable to advance its training and easily get stuck in poor local minima. We show that the widespread experience of sampling more collocation points in high-loss layer regions hardly help optimize and may even worsen the results. These findings motivate the development of a novel curriculum learning method that encourages neural networks to prioritize learning on easier non-layer regions while downplaying learning on harder layer regions. The proposed method helps PINNs automatically adjust the learning emphasis and thereby facilitate the optimization procedure. Numerical results on typical benchmark equations show that the proposed curriculum learning approach mitigates the failure modes of PINNs and can produce accurate results for very sharp boundary and interior layers. Our work reveals that for equations whose solutions have large scale differences, paying less attention to high-loss regions can be an effective strategy for learning them accurately.

cs.LG↗
arXiv2023

One is All: Bridging the Gap Between Neural Radiance Fields Architectures with Progressive Volume Distillation

Neural Radiance Fields (NeRF) methods have proved effective as compact, high-quality and versatile representations for 3D scenes, and enable downstream tasks such as editing, retrieval, navigation, etc. Various neural architectures are vying for the core structure of NeRF, including the plain Multi-Layer Perceptron (MLP), sparse tensors, low-rank tensors, hashtables and their compositions. Each of these representations has its particular set of trade-offs. For example, the hashtable-based representations admit faster training and rendering but their lack of clear geometric meaning hampers downstream tasks like spatial-relation-aware editing. In this paper, we propose Progressive Volume Distillation (PVD), a systematic distillation method that allows any-to-any conversions between different architectures, including MLP, sparse or low-rank tensors, hashtables and their compositions. PVD consequently empowers downstream applications to optimally adapt the neural representations for the task at hand in a post hoc fashion. The conversions are fast, as distillation is progressively performed on different levels of volume representations, from shallower to deeper. We also employ special treatment of density to deal with its specific numerical instability problem. Empirical evidence is presented to validate our method on the NeRF-Synthetic, LLFF and TanksAndTemples datasets. For example, with PVD, an MLP-based NeRF model can be distilled from a hashtable-based Instant-NGP model at a 10X~20X faster speed than being trained the original NeRF from scratch, while achieving a superior level of synthesis quality. Code is available at https://github.com/megvii-research/AAAI2023-PVD.

cs.CV↗
arXiv2022

Towards Accurate Binary Neural Networks via Modeling Contextual Dependencies

Existing Binary Neural Networks (BNNs) mainly operate on local convolutions with binarization function. However, such simple bit operations lack the ability of modeling contextual dependencies, which is critical for learning discriminative deep representations in vision models. In this work, we tackle this issue by presenting new designs of binary neural modules, which enables BNNs to learn effective contextual dependencies. First, we propose a binary multi-layer perceptron (MLP) block as an alternative to binary convolution blocks to directly model contextual dependencies. Both short-range and long-range feature dependencies are modeled by binary MLPs, where the former provides local inductive bias and the latter breaks limited receptive field in binary convolutions. Second, to improve the robustness of binary models with contextual dependencies, we compute the contextual dynamic embeddings to determine the binarization thresholds in general binary convolutional blocks. Armed with our binary MLP blocks and improved binary convolution, we build the BNNs with explicit Contextual Dependency modeling, termed as BCDNet. On the standard ImageNet-1K classification benchmark, the BCDNet achieves 72.3% Top-1 accuracy and outperforms leading binary methods by a large margin. In particular, the proposed BCDNet exceeds the state-of-the-art ReActNet-A by 2.9% Top-1 accuracy with similar operations. Our code is available at https://github.com/Sense-GVT/BCDN

cs.CV↗
arXiv2022

Improve Deep Image Inpainting by Emphasizing the Complexity of Missing Regions

Deep image inpainting research mainly focuses on constructing various neural network architectures or imposing novel optimization objectives. However, on the one hand, building a state-of-the-art deep inpainting model is an extremely complex task, and on the other hand, the resulting performance gains are sometimes very limited. We believe that besides the frameworks of inpainting models, lightweight traditional image processing techniques, which are often overlooked, can actually be helpful to these deep models. In this paper, we enhance the deep image inpainting models with the help of classical image complexity metrics. A knowledge-assisted index composed of missingness complexity and forward loss is presented to guide the batch selection in the training procedure. This index helps find samples that are more conducive to optimization in each iteration and ultimately boost the overall inpainting performance. The proposed approach is simple and can be plugged into many deep inpainting models by changing only a few lines of code. We experimentally demonstrate the improvements for several recently developed image inpainting models on various datasets.

cs.CV↗
arXiv2021

The Emission Order of Hydrogen Isotopes via Correlation Functions in 30 MeV/u Ar+Au Reactions

The intensity interferometry is applied as a chronometer of the particle emission of hydrogen isotopes from the intermediate velocity source formed in $^{40}$Ar+$^{197}$Au reactions at 30 MeV/u. The dynamic emission order of $τ_{\rm p}>τ_{\rm d}>τ_{\rm t}$ is evidenced via the correlation functions of nonidentical particle pairs. Assuming the similar source size, the same emission order is inferred from the correlation functions of identical particle pairs, where $τ_{\rm p} \approx 100 {\rm ~fm/c}$ is extracted by the fit of Koonin-Pratt equation to p-p correlation function. Transport model simulations demonstrate that the dynamic emission order of light charged particles depends on the stiffness of the nuclear symmetry energy.

nucl-ex↗
arXiv2021

Semi-supervised Multi-task Learning for Semantics and Depth

Multi-Task Learning (MTL) aims to enhance the model generalization by sharing representations between related tasks for better performance. Typical MTL methods are jointly trained with the complete multitude of ground-truths for all tasks simultaneously. However, one single dataset may not contain the annotations for each task of interest. To address this issue, we propose the Semi-supervised Multi-Task Learning (SemiMTL) method to leverage the available supervisory signals from different datasets, particularly for semantic segmentation and depth estimation tasks. To this end, we design an adversarial learning scheme in our semi-supervised training by leveraging unlabeled data to optimize all the task branches simultaneously and accomplish all tasks across datasets with partial annotations. We further present a domain-aware discriminator structure with various alignment formulations to mitigate the domain discrepancy issue among datasets. Finally, we demonstrate the effectiveness of the proposed method to learn across different datasets on challenging street view and remote sensing benchmarks.

cs.CV↗
arXiv2021

Properties of the fast fission and the coincident emissions of light charged particles in $^{40}$Ar + $^{197}$Au reactions at 30 MeV/u

The experiment of Ar+Au reactions at 30 MeV/u have been performed using the Compact Spectrometer for Heavy IoN Experiments (CSHINE) in phase I. The light-charged particles are measured by the silicon stripe telescopes in coincidence with the fission fragments recorded by the parallel plate avalanche counters. The distribution properties of the azimuth difference $Δϕ$ and the time-of-flight difference $ΔTOF$ of the fission fragments are presented varying the folding angles which represents the linear momentum transfer from the projectile to the reaction system. The relative abundance of the light charged particles in the fission events to the inclusive events is compared as a function of the laboratory angle $θ_{\rm lab}$ ranging from $18^\circ$ to $60^\circ$ in various folding angle windows. The angular evolution of the yield ratios of p/d and t/d in coincidence with fission fragments is investigated. In a relative comparison, tritons are more abundantly emitted at small angles, while protons are more abundant at large angles. The angular evolution of the neutron richness of the light-charged particles is consistent with the results obtained in previous inclusive experiments.

nucl-ex↗
arXiv2021

PC-GAIN: Pseudo-label Conditional Generative Adversarial Imputation Networks for Incomplete Data

Datasets with missing values are very common in real world applications. GAIN, a recently proposed deep generative model for missing data imputation, has been proved to outperform many state-of-the-art methods. But GAIN only uses a reconstruction loss in the generator to minimize the imputation error of the non-missing part, ignoring the potential category information which can reflect the relationship between samples. In this paper, we propose a novel unsupervised missing data imputation method named PC-GAIN, which utilizes potential category information to further enhance the imputation power. Specifically, we first propose a pre-training procedure to learn potential category information contained in a subset of low-missing-rate data. Then an auxiliary classifier is determined using the synthetic pseudo-labels. Further, this classifier is incorporated into the generative adversarial framework to help the generator to yield higher quality imputation results. The proposed method can improve the imputation quality of GAIN significantly. Experimental results on various benchmark datasets show that our method is also superior to other baseline approaches. Our code is available at \url{https://github.com/WYu-Feng/pc-gain}.

cs.LG↗
arXiv2019

Long-time asymptotic behavior for an extended modified Korteweg-de Vries equation

We investigate an integrable extended modified Korteweg-de Vries equation on the line with the initial value belonging to the Schwartz space. By performing the nonlinear steepest descent analysis of an associated matrix Riemann--Hilbert problem, we obtain the explicit leading-order asymptotics of the solution of this initial value problem as time $t$ goes to infinity. For a special case $α=0$, we present the asymptotic formula of the solution to the extended modified Korteweg-de Vries equation in region $\mathcal{P}=\{(x,t)\in\bfR^2|0<x\leq Mt^{\frac{1}{5}},t\geq3\}$ in terms of the solution of a fourth order Painlevé II equation.

math.AP↗
arXiv2011

Increased Coupling in the Saliency Network is the main cause/effect of Attention Deficit Hyperactivity Disorder

To uncover the underlying mechanisms of mental disorders such as attention deficit hyperactivity disorder (ADHD) for improving both early diagnosis and therapy, it is increasingly recognized that we need a better understanding of how the brain's functional connections are altered. A new brain wide association study (BWAS) has been developed and used to investigate functional connectivity changes in the brains of patients suffering from ADHD using resting state fMRI data. To reliably find out the most significantly altered functional connectivity links and associate them with ADHD, a meta-analysis on a cohort of ever reported largest population comprising 249 patients and 253 healthy controls is carried out. The greatest change in ADHD patients was the increased coupling of the saliency network involving the anterior cingulate gyrus and anterior insula. A voxel-based morphometry analysis was also carried out but this revealed no evidence in the ADHD patients for altered grey matter volumes in the regions showing altered functional connectivity. This is the first evidence for the involvement of the saliency network in ADHD and it suggests that this may reflect increased sensitivity over the integration of the incoming sensory information and his/her own thoughts and the network as a switch is bias towards to the central executive network.

physics.bio-ph↗