The University of Sydney（悉尼大学）；Max Planck Institute for Mathematics in the Sciences（马克斯·普朗克科学数学研究所）；Shanghai Jiao Tong University（上海交通大学）；The University of New South Wales（新南威尔士大学）；University of Cambridge（剑桥大学）；Zhejiang Normal University（浙江师范大学）；University of California, Los Angeles（加州大学洛杉矶分校）
Xuebin Zheng, Bingxin Zhou, Junbin Gao, Yu Guang Wang, Pietro Lio, Ming Li, Guido Montufar
This paper presents a new approach for assembling graph neural networks based on framelet transforms. The latter provides a multi-scale representation for graph-structured data. We decompose an input graph into low-pass and high-pass frequencies coefficients for network training, which then defines a framelet-based graph convolution. The framelet decomposition naturally induces a graph pooling strategy by aggregating the graph feature into low-pass and high-pass spectra, which considers both the feature values and geometry of the graph data and conserves the total information. The graph neural networks with the proposed framelet convolution and pooling achieve state-of-the-art performance in many node and graph prediction tasks. Moreover, we propose shrinkage as a new activation for the framelet convolution, which thresholds high-frequency information at different scales. Compared to ReLU, shrinkage activation improves model performance on denoising and signal compression: noises in both node and structure can be significantly reduced by accurately cutting off the high-pass coefficients from framelet decomposition, and the signal can be compressed to less than half its original size with well-preserved prediction performance.
King Abdullah University of Science and Technology（阿卜杜拉国王科技大学）；Shanghai Jiao Tong University（上海交通大学）
Graph neural networks (GNNs) have achieved champion in wide applications. Neural message passing is a typical key module for feature propagation by aggregating neighboring features. In this work, we propose a new message passing based on multiscale framelet transforms, called Framelet Message Passing. Different from traditional spatial methods, it integrates framelet representation of neighbor nodes from multiple hops away in node message update. We also propose a continuous message passing using neural ODE solvers. It turns both discrete and continuous cases can provably achieve network stabi…
University of Sydney（悉尼大学）；Shanghai Jiao Tong University（上海交通大学）；Max Planck Institute for Mathematics in Sciences（马克斯·普朗克科学数学研究所）；University of New South Wales（新南威尔士大学）；City University of Hong Kong（香港城市大学）
Graph representation learning has many real-world applications, from super-resolution imaging, 3D computer vision to drug repurposing, protein classification, social networks analysis. An adequate representation of graph data is vital to the learning performance of a statistical or machine learning model for graph-structured data. In this paper, we propose a novel multiscale representation system for graph data, called decimated framelets, which form a localized tight frame on the graph. The decimated framelet system allows storage of the graph data representation on a coarse-grained chain and…
University of Sydney（悉尼大学）；Western Sydney University（西悉尼大学）
In this work, we provide a theoretical understanding of the framelet-based graph neural networks through the perspective of energy gradient flow. By viewing the framelet-based models as discretized gradient flows of some energy, we show it can induce both low-frequency and high-frequency-dominated dynamics, via the separate weight matrices for different frequency components. This substantiates its good empirical performance on both homophilic and heterophilic graphs. We then propose a generalized energy via framelet decomposition and show its gradient flow leads to a novel graph neural network…
The University of Sydney（悉尼大学）；Shanghai Jiao Tong University（上海交通大学）；Shanghai Artificial Intelligence Laboratory（上海人工智能实验室）；The University of New South Wales（新南威尔士大学）
As graph data collected from the real world is merely noise-free, a practical representation of graphs should be robust to noise. Existing research usually focuses on feature smoothing but leaves the geometric structure untouched. Furthermore, most work takes L2-norm that pursues a global smoothness, which limits the expressivity of graph neural networks. This paper tailors regularizers for graph data in terms of both feature and structure noises, where the objective function is efficiently solved with the alternating direction method of multipliers (ADMM). The proposed scheme allows to take m…
The University of Sydney Business School（悉尼大学商学院）；The University of Sydney（悉尼大学）
In this paper, we propose a simple yet effective graph neural network for directed graphs (digraph) based on the classic Singular Value Decomposition (SVD), named SVD-GCN. The new graph neural network is built upon the graph SVD-framelet to better decompose graph signals on the SVD ``frequency'' bands. Further the new framelet SVD-GCN is also scaled up for larger scale graphs via using Chebyshev polynomial approximation. Through empirical experiments conducted on several node classification datasets, we have found that SVD-GCN has remarkable improvements in a variety of graph node learning tas…
In this paper, we develop a general theoretical framework for constructing Haar-type tight framelets on any compact set with a hierarchical partition. In particular, we construct a novel area-regular hierarchical partition on the 2-sphere and establish its corresponding spherical Haar tight framelets with directionality. We conclude by evaluating and illustrating the effectiveness of our area-regular spherical Haar tight framelets in several denoising experiments. Furthermore, we propose a convolutional neural network (CNN) model for spherical signal denoising which employs the fast framelet d…
Korea Advanced Institute of Science and Technology (KAIST)（韩国科学技术院）；Seoul National University Bundang Hospital（首尔国立大学盆唐医院）
Model based iterative reconstruction (MBIR) algorithms for low-dose X-ray CT are computationally expensive. To address this problem, we recently proposed a deep convolutional neural network (CNN) for low-dose X-ray CT and won the second place in 2016 AAPM Low-Dose CT Grand Challenge. However, some of the texture were not fully recovered. To address this problem, here we propose a novel framelet-based denoising algorithm using wavelet residual network which synergistically combines the expressive power of deep learning and the performance guarantee from the framelet-based denoising algorithms. …
Machine learning-based analysis of medical images often faces several hurdles, such as the lack of training data, the curse of dimensionality problem, and the generalization issues. One of the main difficulties is that there exists computational cost problem in dealing with input data of large size matrices which represent medical images. The purpose of this paper is to introduce a framelet-pooling aided deep learning method for mitigating computational bundle, caused by large dimensionality. By transforming high dimensional data into low dimensional components by filter banks with preserving …
Korea Advanced Institute of Science and Technology（韩国科学技术院）；Bio Imaging and Signal Processing Lab.（生物成像与信号处理实验室）；Department of Bio and Brain Engineering（生物与脑工程系）
Recently, deep learning approaches with various network architectures have achieved significant performance improvement over existing iterative reconstruction methods in various imaging problems. However, it is still unclear why these deep learning architectures work for specific inverse problems. To address these issues, here we show that the long-searched-for missing link is the convolution framelets for representing a signal by convolving local and non-local bases. The convolution framelets was originally developed to generalize the theory of low-rank Hankel matrix approaches for inverse pr…
An Empirical Study of Retrieval-enhanced Graph Neural Networks
University of Oxford（牛津大学）；Mila-HEC Montreal（蒙特利尔高等商学院米拉实验室）；University of Cambridge（剑桥大学）；Tencent AI Lab（腾讯人工智能实验室）；BioMap（百图生科）；MBZUAI（穆罕默德·本·扎耶德人工智能大学）；The University of Hong Kong（香港大学）
It is hard to directly implement Graph Neural Networks (GNNs) on large scaled graphs. Besides of existed neighbor sampling techniques, scalable methods decoupling graph convolutions and other learnable transformations into preprocessing and post classifier allow normal minibatch training. By replacing redundant concatenation operation with attention mechanism in SIGN, we propose Scalable and Adaptive Graph Neural Networks (SAGN). SAGN can adaptively gather neighborhood information among different hops. To further improve scalable models on semi-supervised learning tasks, we propose Self-Label-…
The University of Sydney（悉尼大学）；Shanghai Jiao Tong University（上海交通大学）；University of Cambridge（剑桥大学）；The University of New South Wales（新南威尔士大学）
Many real-world relational systems, such as social networks and biological systems, contain dynamic interactions. When learning dynamic graph representation, it is essential to employ sequential temporal information and geometric structure. Mainstream work achieves topological embedding via message passing networks (e.g., GCN, GAT). The temporal evolution, on the other hand, is conventionally expressed via memory units (e.g., LSTM or GRU) that possess convenient information filtration in a gate mechanism. Though, such a design prevents large-scale input sequence due to the over-complicated enc…
In this work, we formulate NEWRON: a generalization of the McCulloch-Pitts neuron structure. This new framework aims to explore additional desirable properties of artificial neurons. We show that some specializations of NEWRON allow the network to be interpretable with no change in their expressiveness. By just inspecting the models produced by our NEWRON-based networks, we can understand the rules governing the task. Extensive experiments show that the quality of the generated models is better than traditional interpretable models and in line or better than standard neural networks.
In the recent past, there has been a growing interest in Neural-Symbolic Integration frameworks, i.e., hybrid systems that integrate connectionist and symbolic approaches to obtain the best of both worlds. In a previous work, we proposed KENN (Knowledge Enhanced Neural Networks), a Neural-Symbolic architecture that injects prior logical knowledge into a neural network by adding a new final layer which modifies the initial predictions accordingly to the knowledge. Among the advantages of this strategy, there is the inclusion of clause weights, learnable parameters that represent the strength of…
Research Center for Information Technology Innovation, Academia Sinica（中央研究院资讯科技创新研究中心）；National Taiwan University（台湾大学）
Numerous studies have investigated the effectiveness of neural network quantization on pattern classification tasks. The present study, for the first time, investigated the performance of speech enhancement (a regression task in speech processing) using a novel exponent-only floating-point quantized neural network (EOFP-QNN). The proposed EOFP-QNN consists of two stages: mantissa-quantization and exponent-quantization. In the mantissa-quantization stage, EOFP-QNN learns how to quantize the mantissa bits of the model parameters while preserving the regression accuracy using the least mantissa p…
UiT the Arctic University of Norway（挪威北极大学）；Sapienza University of Rome（罗马第一大学）；NORCE Norwegian Research Centre（挪威NORCE研究中心）
Recently, subgraphs-enhanced Graph Neural Networks (SGNNs) have been introduced to enhance the expressive power of Graph Neural Networks (GNNs), which was proved to be not higher than the 1-dimensional Weisfeiler-Leman isomorphism test. The new paradigm suggests using subgraphs extracted from the input graph to improve the model's expressiveness, but the additional complexity exacerbates an already challenging problem in GNNs: explaining their predictions. In this work, we adapt PGExplainer, one of the most recent explainers for GNNs, to SGNNs. The proposed explainer accounts for the contribut…
Graph Neural Networks (GNNs) have shown promising results in various tasks, among which link prediction is an important one. GNN models usually follow a node-centric message passing procedure that aggregates the neighborhood information to the central node recursively. Following this paradigm, features of nodes are passed through edges without caring about where the nodes are located and which role they played. However, the neglected topological information is shown to be valuable for link prediction tasks. In this paper, we propose Structure Enhanced Graph neural network (SEG) for link predic…
Multiscale fusion enhanced spiking neural network for invasive BCI neural signal decoding
University of Chinese Academy of Sciences（中国科学院大学）；Center for Excellence in Brain Science and Intelligence Technology, Chinese Academy of Sciences（中国科学院脑科学与智能技术卓越创新中心）；Institute of Automation, Chinese Academy of Sciences（中国科学院自动化研究所）
Encoder-decoder networks using convolutional neural network (CNN) architecture have been extensively used in deep learning literatures thanks to its excellent performance for various inverse problems. However, it is still difficult to obtain coherent geometric view why such an architecture gives the desired performance. Inspired by recent theoretical understanding on generalizability, expressivity and optimization landscape of neural networks, as well as the theory of convolutional framelets, here we provide a unified theoretical framework that leads to a better understanding of geometry of en…
Institute of Computational Mathematics and Scientific/Engineering Computing（科学与工程计算研究所）
We propose in this work the gradient-enhanced deep neural networks (DNNs) approach for function approximations and uncertainty quantification. More precisely, the proposed approach adopts both the function evaluations and the associated gradient information to yield enhanced approximation accuracy. In particular, the gradient information is included as a regularization term in the gradient-enhanced DNNs approach, for which we present similar posterior estimates (by the two-layer neural networks) as those in the path-norm regularized DNNs approximations. We also discuss the application of this …
Trainable Adaptive Activation Function Structure (TAAFS) Enhances Neural Network Force Field Performance with Only Dozens of Additional Parameters
State Key Lab of Processors, Institute of Computing Technology, Chinese Academy of Sciences（中国科学院计算技术研究所 处理器国家重点实验室）；University of Chinese Academy of Sciences（中国科学院大学）
This work proposes a method for underwater image enhancement using the principle of histogram equalization. Since underwater images have a global strong dominant colour, their colourfulness and contrast are often degraded. Before applying the histogram equalisation technique on the image, the image is converted from coloured image to a gray scale image for further operations. Histogram equalization is a technique for adjusting image intensities to enhance contrast. The colours of the image are retained using a convolutional neural network model which is trained by the datasets of underwater im…
Deep neural networks have gained tremendous importance in many computer vision tasks. However, their power comes at the cost of large amounts of annotated data required for supervised training. In this work we review and compare different techniques available in the literature to improve training results without acquiring additional annotated real-world data. This goal is mostly achieved by applying annotation-preserving transformations to existing data or by synthetically creating more data.
In hearing aids, the presence of babble noise degrades hearing intelligibility of human speech greatly. However, removing the babble without creating artifacts in human speech is a challenging task in a low SNR environment. Here, we sought to solve the problem by finding a `mapping' between noisy speech spectra and clean speech spectra via supervised learning. Specifically, we propose using fully Convolutional Neural Networks, which consist of lesser number of parameters than fully connected networks. The proposed network, Redundant Convolutional Encoder Decoder (R-CED), demonstrates that a co…
This work evaluates six state-of-the-art deep neural network (DNN) architectures applied to the problem of enhancing camera-captured document images. The results from each network were evaluated both qualitatively and quantitatively using Image Quality Assessment (IQA) metrics, and also compared with an existing approach based on traditional computer vision techniques. The best performing architectures generally produced good enhancement compared to the existing algorithm, showing that it is possible to use DNNs for document image enhancement. Furthermore, the best performing architectures cou…
This paper aims to propose a novel deep learning-integrated framework for deriving reliable simulation input models through incorporating multi-source information. The framework sources and extracts multisource data generated from construction operations, which provides rich information for input modeling. The framework implements Bayesian deep neural networks to facilitate the purpose of incorporating richer information in input modeling. A case study on road paving operation is performed to test the feasibility and applicability of the proposed framework. Overall, this research enhances inpu…
Rice University（莱斯大学）；Prudence Medical Technologies Ltd（审慎医疗技术有限公司）
We train an enhanced deep convolutional neural network in order to identify eight cardiac abnormalities from the standard 12-lead electrocardiograms (ECGs) using the dataset of 14000 ECGs. Instead of straightforwardly applying an end-to-end deep learning approach, we find that deep convolutional neural networks enhanced with sophisticated hand crafted features show advantages in reducing generalization errors. Additionally, data preprocessing and augmentation are essential since the distribution of eight cardiac abnormalities are highly biased in the given dataset. Our approach achieves promis…
In this work, a novel multifidelity machine learning (ML) model, the gradient-enhanced multifidelity neural networks (GEMFNNs), is proposed. This model is a multifidelity version of gradient-enhanced neural networks (GENNs) as it uses both function and gradient information available at multiple levels of fidelity to make function approximations. Its construction is similar to multifidelity neural networks (MFNNs). This model is tested on three analytical function, a one, two, and a 20 variable function. It is also compared to neural networks (NNs), GENNs, and MFNNs, and the number of samples r…
Progress in making neural networks more robust against adversarial attacks is mostly marginal, despite the great efforts of the research community. Moreover, the robustness evaluation is often imprecise, making it difficult to identify promising approaches. We analyze the classification decisions of 19 different state-of-the-art neural networks trained to be robust against adversarial attacks. Our findings suggest that current untargeted adversarial attacks induce misclassification towards only a limited amount of different classes. Additionally, we observe that both over- and under-confidence…
Identifying the dynamics of physical systems requires a machine learning model that can assimilate observational data, but also incorporate the laws of physics. Neural Networks based on physical principles such as the Hamiltonian or Lagrangian NNs have recently shown promising results in generating extrapolative predictions and accurately representing the system's dynamics. We show that by additionally considering the actual energy level as a regularization term during training and thus using physical information as inductive bias, the results can be further improved. Especially in the case wh…
The Graduate Center, CUNY（纽约市立大学研究生院）；Brooklyn College（布鲁克林学院）
Traditional speech enhancement systems produce speech with compromised quality. Here we propose to use the high quality speech generation capability of neural vocoders for better quality speech enhancement. We term this parametric resynthesis (PR). In previous work, we showed that PR systems generate high quality speech for a single speaker using two neural vocoders, WaveNet and WaveGlow. Both these vocoders are traditionally speaker dependent. Here we first show that when trained on data from enough speakers, these vocoders can generate speech from unseen speakers, both male and female, with …
Institute of Automation, Chinese Academy of Sciences（中国科学院自动化研究所）；CAS Center for Excellence in Brain Science and Intelligence Technology（中国科学院脑科学与智能技术卓越创新中心）；University of Chinese Academy of Sciences（中国科学院大学）
Recurrent neural networks (RNNs) have shown significant improvements in recent years for speech enhancement. However, the model complexity and inference time cost of RNNs are much higher than deep feed-forward neural networks (DNNs). Therefore, these limit the applications of speech enhancement. This paper proposes a deep time delay neural network (TDNN) for speech enhancement with full data learning. The TDNN has excellent potential for capturing long range temporal contexts, which utilizes a modular and incremental design. Besides, the TDNN preserves the feed-forward structure so that its in…
Deep neural networks (DNNs) have shown exceptional performances in a wide range of tasks and have become the go-to method for problems requiring high-level predictive power. There has been extensive research on how DNNs arrive at their decisions, however, the inherently uninterpretable networks remain up to this day mostly unobservable "black boxes". In recent years, the field has seen a push towards interpretable neural networks, such as the visually interpretable Neural Additive Models (NAMs). We propose a further step into the direction of intelligibility beyond the mere visualization of fe…
Fourier-Enhanced Recurrent Neural Networks for Electrical Load Time Series Downscaling
Department of Statistics University of Chicago Chicago, IL, USA（统计学系 芝加哥大学）；Computer Science Division Argonne National Laboratory（计算机科学 division 阿贡国家实验室）
Rutgers, the State University of New Jersey（罗格斯新泽西州立大学）
Most current speech enhancement models use spectrogram features that require an expensive transformation and result in phase information loss. Previous work has overcome these issues by using convolutional networks to learn long-range temporal correlations across high-resolution waveforms. These models, however, are limited by memory-intensive dilated convolution and aliasing artifacts from upsampling. We introduce an end-to-end fully-recurrent hourglass-shaped neural network architecture with residual connections for waveform-based single-channel speech enhancement. Our model can efficiently …
Rice University（莱斯大学）；Facebook AI Research（Facebook人工智能研究院）
Kernels derived from deep neural networks (DNNs) in the infinite-width regime provide not only high performance in a range of machine learning tasks but also new theoretical insights into DNN training dynamics and generalization. In this paper, we extend the family of kernels associated with recurrent neural networks (RNNs), which were previously derived only for simple RNNs, to more complex architectures including bidirectional RNNs and RNNs with average pooling. We also develop a fast GPU implementation to exploit the full practical potential of the kernels. Though RNNs are typically only ap…
We apply a fast kernel method for mask-based single-channel speech enhancement. Specifically, our method solves a kernel regression problem associated to a non-smooth kernel function (exponential power kernel) with a highly efficient iterative method (EigenPro). Due to the simplicity of this method, its hyper-parameters such as kernel bandwidth can be automatically and efficiently selected using line search with subsamples of training data. We observe an empirical correlation between the regression loss (mean square error) and regular metrics for speech enhancement. This observation justifies …
