检索结果-内蒙古大学图书馆

A neuroscience-inspired spiking neural network for EEG-based auditory spatial attention detection

NEURAL NETWORKS 2022年第0期152卷 555-565页

作者： Faghihi, Faramarz Cai, Siqi Moustafa, Ahmed A. Univ Med Ctr Utrecht Dept Med Physiol Div Heart & Lungs Utrecht Netherlands Univ Bremen Machine Listening Lab Bremen Germany Natl Univ Singapore Dept Elect & Comp Engn Singapore Singapore Bond Univ Fac Soc & Design Sch Psychol Gold Coast Qld Australia Univ Johannesburg Fac Hlth Sci Dept Human Anat & Physiol Johannesburg South Africa

Recent studies have shown that alpha oscillations (8-13 Hz) enable the decoding of auditory spatial attention. Inspired by sparse coding in cortical neurons, we propose a spiking neural network model for auditory spatial attention detection. The proposed model can extract the patterns of recorded EEG of leftward and rightward attention, independently, and uses them to train the network to detect auditory spatial attention. Specifically, our model is composed of three layers, two of which are Integrate and Fire spiking neurons. We formulate a new learning rule that is based on the firing rate of pre- and post-synaptic neurons in the first and second layers of spiking neurons. The third layer has 10 spiking neurons and the pattern of their firing rate is used in the test phase to decode the auditory spatial attention of a given test sample. Moreover, the effects of using low connectivity rates of the layers and specific range of learning parameters of the learning rule are investigated. The proposed model achieves an average accuracy of 90% with only 10% of EEG signals as training data. This study also provides new insights into the role of sparse coding in both cortical networks subserving cognitive tasks and brain-inspired machine learning.

关键词： EEG Spiking neurons Auditory attention Connectivity sparse coding

来源：评论

学校读者我要写书评

暂无评论

Accelerating MR Parameter Mapping Using Nonlinear Compressive Manifold Learning and Regularized Pre-Imaging

引用

IEEE TRANSACTIONS ON BIOMEDICAL ENGINEERING 2022年第10期69卷 2996-3007页

作者： Zhou, Yihang Wang, Haifeng Liu, Yuanyuan Liang, Dong Ying, Leslie Hong Kong Sanat & Hosp Res Dept Hong Kong Peoples R China Chinese Acad Sci Shenzhen Inst Adv Technol Paul C Lauterbur Res Ctr Biomed Imaging Shenzhen 518055 Peoples R China Key Lab Magnet Resonance & Multimodal Imaging Gua Shenzhen 518055 Peoples R China Shenzhen Key Lab Ultrasound Imaging & Therapy Shenzhen Peoples R China Chinese Acad Sci Shenzhen Inst Adv Technol Res Ctr Med AI Shenzhen 518055 Peoples R China SUNY Buffalo Dept Elect Engn Buffalo NY USA SUNY Buffalo Dept Biomed Engn Buffalo NY USA

In this study, we present a novel method to reconstruct the MR parametric maps from highly undersampled k-space data. Specifically, we utilize a nonlinear model to sparsely represent the unknown MR parameter-weighted images in high-dimensional feature space. Each image at a specific time point is assumed to belong to a low-dimensional manifold which is learned from training images created based on the parametric model. The final reconstruction is carried out by venturing the sparse representation of the images in the feature space back to the input space, using the pre-imaging technique. Particularly, among an infinite number of solutions that satisfy the data consistency, the one that is closest to the manifold is selected as the desired solution. The underlying optimization problem is solved using kernel trick, sparse coding, and split Bregman iteration algorithm. In addition, both spatial and temporal regularizations are utilized to further improve the reconstruction quality. The proposed method is validated on both phantom and in vivo human brain T2 mapping data. Results suggest that the proposed method is superior to the conventional linear model-based reconstruction methods, in terms of artifact removal and quantitative estimation accuracy. The proposed method could be potentially beneficial for quantitative MR applications.

关键词： Image reconstruction Kernel Manifolds Training Brain modeling Encoding sparse matrices Kernel trick MR parameter mapping nonlinear manifold regularized pre-image sparse coding

来源：评论

学校读者我要写书评

暂无评论

GraFT: Graph Filtered Temporal Dictionary Learning for Functional Neural Imaging

引用

IEEE TRANSACTIONS ON IMAGE PROCESSING 2022年 31卷 3509-3524页

作者： Charles, Adam S. Cermak, Nathan Affan, Rifqi O. Scott, Benjamin B. Schiller, Jackie Mishne, Gal Kavli Neurosci Discovery Inst Dept Biomed Engn Baltimore MD 21218 USA Johns Hopkins Univ Ctr Imaging Sci Baltimore MD 21287 USA Johns Hopkins Univ Math Inst Data Sci Baltimore MD 21287 USA Technion Israel Inst Technol IL-31096 Haifa Israel Boston Univ Dept Psychol & Brain Sci Boston MA 02215 USA Halicioglu Data Sci Inst La Jolla CA 92093 USA Univ Calif San Diego Neurosci Grad Program La Jolla CA 92093 USA

Optical imaging of calcium signals in the brain has enabled researchers to observe the activity of hundreds-to-thousands of individual neurons simultaneously. Current methods predominantly use morphological information, typically focusing on expected shapes of cell bodies, to better identify neurons in the field-of-view. The explicit shape constraints limit the applicability of automated cell identification to other important imaging scales with more complex morphologies, e.g., dendritic or widefield imaging. Specifically, fluorescing components may be broken up, incompletely found, or merged in ways that do not accurately describe the underlying neural activity. Here we present Graph Filtered Temporal Dictionary (GraFT), a new approach that frames the problem of isolating independent fluorescing components as a dictionary learning problem. Specifically, we focus on the time-traces-the main quantity used in scientific discovery-and learn a time trace dictionary with the spatial maps acting as the presence coefficients encoding which pixels the time-traces are active in. Furthermore, we present a novel graph filtering model which redefines connectivity between pixels in terms of their shared temporal activity, rather than spatial proximity. This model greatly eases the ability of our method to handle data with complex non-local spatial structure. We demonstrate important properties of our method, such as robustness to morphology, simultaneously detecting different neuronal types, and implicitly inferring number of neurons, on both synthetic data and real data examples. Specifically, we demonstrate applications of our method to calcium imaging both at the dendritic, somatic, and widefield scales.

关键词： Imaging Calcium Dictionaries Machine learning Shape Neurons Optical imaging Dictionary learning sparse coding calcium imaging two-photon microscopy re-weighted l1

来源：评论

学校读者我要写书评

暂无评论

AN EFFICIENT AND INTERPRETABLE SPEECH ENHANCEMENT NETWORK VIA DEEP DICTIONARY LEARNING 49

AN EFFICIENT AND INTERPRETABLE SPEECH ENHANCEMENT NETWORK VI...

引用

2024 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2024

作者： Xu, Xinmeng Zhang, Yiqun Tu, Weiping Yang, Yuhong NERCMS School of Computer Science Wuhan University China Hubei Luojia Laboratory China Hubei Key Laboratory of Multimedia and Network Communication Engineering Wuhan University China

ISBN: (纸本)9798350344851

Speech enhancement is a vital and highly ill-posed problem for many speech downstream tasks. While currently existing deep learning based speech enhancement methods have held state-of-the-art results, they still possess apparent shortcomings in that most of the deep learning based models lack interpretability. This deficiency results in unsatisfied speech enhancement performance in many sophisticated scenarios. To tackle this problem, we integrate dictionary learning and sparse coding into deep learning networks for speech enhancement and present a deep dictionary learning based speech enhancement network (DicLSENet). Specifically, the proposed DicLSENet strictly follows the principle of dictionary learning, learns the priors for both representation coefficients and dictionaries, and adaptively adjusts the dictionary for each input. Experimental results show that the proposed model outperforms state-of-the-art fully deep learning based methods with attractive computational costs. © 2024 IEEE.

关键词： dictionary learning sparse coding Speech enhancement

来源：评论

学校读者我要写书评

暂无评论

Dictionary Learning of Symmetric Positive Definite Data Based on Riemannian Manifold Tangent Spaces and Local Homeomorphism

引用

IEEE ACCESS 2022年 10卷 46114-46127页

作者： He, Hui Liu, Shuyu Ma, Zhengming Feng, Weijia Sun Yat Sen Univ Sch Elect & Informat Technol Guangzhou 510006 Peoples R China Sun Yat Sen Univ Publ Expt Teaching Ctr Guangzhou 510006 Peoples R China China Telecom Cloud Technol Co Ltd Guangzhou 510000 Peoples R China

Symmetric Positive Definite (SPD) data are increasingly prevalent in dictionary learning recently. SPD data are the typical non-Euclidean data and cannot constitute a Euclidean space. Therefore, many dictionary learning algorithms cannot be directly adopted on SPD data. Reproducing Kernel Hilbert Spaces (RKHS) is now commonly used to deal with this problem. However, RKHS is an infinite-dimensional Hilbert space, rather than a Euclidean space, resulting in the inability of the dictionary learning to be directly used on SPD data. In this paper, we propose a novel dictionary learning algorithm for SPD data, which is based on the Riemannian Manifold Tangent Space (RMTS). Since RMTS is based on a finite-dimensional Hilbert space, i.e., Euclidean space, most machine learning algorithms developed on Euclidean space can be directly applied to RMTS. The introduction of RMTS provides a better mathematical platform for machine learning of non-Euclidean data. The proposed RMTS method first selects a point on a Riemannian manifold as an anchor point. It transforms the other points on the Riemannian manifold into tangent vectors of the geodesics between these points and the anchor point, with the tangent vectors at the anchor point. We set the length of the tangent vector equal to the length of the geodesic. As a result, such tangent vectors have an explicit geometric meaning, such as direction information, while the RKHS method may cause some geometric meaning to be lost in the original data during the mapping process. In addition, the proposed algorithm adds a regular term of local homomorphism between SPD data and its RMTS dictionary coding coefficients, so that the similarity of SPD data and its RMTS dictionary coding coefficients are as close as possible. Experimental results on four public datasets demonstrate that the proposed algorithm significantly outperforms five state-of-the-art algorithms. This work opens a new pathway towards SPD data dictionary learning methods.

关键词： Manifolds Machine learning Dictionaries Measurement Hilbert space Machine learning algorithms Encoding Symmetric positive definite matrices Riemannian manifolds dictionary learning sparse coding

来源：评论

学校读者我要写书评

暂无评论

Evolutionary Variational Optimization of Generative Models

引用

JOURNAL OF MACHINE LEARNING RESEARCH 2022年第1期23卷 1-51页

作者： Drefs, Jakob Guiraud, Enrico Luecke, Joerg Carl von Ossietzky Univ Oldenburg Machine Learning Lab D-26129 Oldenburg Germany CERN CH-1211 Geneva Switzerland

We combine two popular optimization approaches to derive learning algorithms for generative models: variational optimization and evolutionary algorithms. The combination is realized for generative models with discrete latents by using truncated posteriors as the family of variational distributions. The variational parameters of truncated posteriors are sets of latent states. By interpreting these states as genomes of individuals and by using the variational lower bound to define a fitness, we can apply evolutionary algorithms to realize the variational loop. The used variational distributions are very flexible and we show that evolutionary algorithms can effectively and efficiently optimize the variational bound. Furthermore, the variational loop is generally applicable ("black box") with no analytical derivations required. To show general applicability, we apply the approach to three generative models (we use Noisy-OR Bayes Nets, Binary sparse coding, and Spike-and-Slab sparse coding). To demonstrate effectiveness and efficiency of the novel variational approach, we use the standard competitive benchmarks of image denoising and inpainting. The benchmarks allow quantitative comparisons to a wide range of methods including probabilistic approaches, deep deterministic and generative networks, and non-local image processing methods. In the category of "zero-shot" learning (when only the corrupted image is used for training), we observed the evolutionary variational algorithm to significantly improve the state-of-the-art in many benchmark settings. For one well-known inpainting benchmark, we also observed state-of-the-art performance across all categories of algorithms although we only train on the corrupted image. In general, our investigations highlight the importance of research on optimization methods for generative models to achieve performance improvements.

关键词： Expectation Maximization Variational Methods Evolutionary Algorithms sparse coding Denoising Inpainting

来源：评论

学校读者我要写书评

暂无评论

Learnable Subspace Clustering

引用

IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS 2022年第3期33卷 1119-1133页

作者： Li, Jun Liu, Hongfu Tao, Zhiqiang Zhao, Handong Fu, Yun Northeastern Univ Coll Engn Dept Elect & Comp Engn Boston MA 02115 USA Nanjing Univ Sci & Technol Sch Comp Sci & Engn Nanjing 210094 Peoples R China Brandeis Univ Michtom Sch Comp Sci Waltham MA 02453 USA Santa Clara Univ Dept Comp Sci & Engn Santa Clara CA 95053 USA Adobe Res San Jose CA 95113 USA Northeastern Univ Khoury Coll Comp Sci Boston MA 02115 USA

This article studies the large-scale subspace clustering ((LSC)-C-2) problem with millions of data points. Many popular subspace clustering methods cannot directly handle the (LSC)-C-2 problem although they have been considered to be state-of-the-art methods for small-scale data points. A simple reason is that these methods often choose all data points as a large dictionary to build huge coding models, which results in high time and space complexity. In this article, we develop a learnable subspace clustering paradigm to efficiently solve the (LSC)-C-2 problem. The key concept is to learn a parametric function to partition the high-dimensional subspaces into their underlying low-dimensional subspaces instead of the computationally demanding classical coding models. Moreover, we propose a unified, robust, predictive coding machine (RPCM) to learn the parametric function, which can be solved by an alternating minimization algorithm. Besides, we provide a bounded contraction analysis of the parametric function. To the best of our knowledge, this article is the first work to efficiently cluster millions of data points among the subspace clustering methods. Experiments on million-scale data sets verify that our paradigm outperforms the related state-of-the-art methods in both efficiency and effectiveness.

关键词： Encoding Data models Computational modeling Dictionaries Phase change materials Big Data Image coding Elastic net regression large-scale spectral clustering (LsSpC) large-scale subspace clustering ((LSC)-C-2) low-rank representation (LRR) neural networks sparse coding

来源：评论

学校读者我要写书评

暂无评论

Surgical Procedure Understanding, Evaluation, and Interpretation: A Dictionary Factorization Approach

IEEE TRANSACTIONS ON MEDICAL ROBOTICS AND BIONICS

引用

IEEE TRANSACTIONS ON MEDICAL ROBOTICS AND BIONICS 2022年第2期4卷 423-435页

作者： Soleymani, Abed Li, Xingyu Tavakoli, Mahdi Univ Alberta Elect & Comp Engn Dept Edmonton AB T6G 2R3 Canada

In this study, we present a novel machine learning-based technique to help surgical mentors assess surgical motion trajectories and corresponding surgical skills levels in surgical training programs. The proposed method is a variation of sparse coding and dictionary learning that is straightforward to optimize and produces approximate trajectory decomposition for structured tasks. Our approach is superior to existing stochastic or deep learning-based methods in terms of transparency of the model and interpretability of the results. We introduce a dual-sparse coding algorithm which encourages the elimination of redundant and unnecessary atoms and targets to reach the most informative dictionary, representing the most important temporal variations within a given surgical trajectory. Since surgical tool trajectories are time series signals, we further incorporate the idea of floating atoms along the temporal axis in trajectory analysis, which improves the model's accuracy and prevents information loss in downstream tasks. Using JIGSAWS data set, we present preliminary results showing the feasibility of the proposed method for clustering and interpreting surgical trajectories in terms of user's skills-related behaviors.

关键词： Machine learning dictionary learning sparse coding surgical trajectory decomposition surgical skills assessment

来源：评论

学校读者我要写书评

暂无评论

Revisiting the Problem of Uniqueness in sparse Reconstruction 24

Revisiting the Problem of Uniqueness in Sparse Reconstructio...

引用

7th International Conference on Mathematics and Statistics, ICoMS 2024

作者： Domokos, Zoltan University of Szeged Szeged Hungary

ISBN: (纸本)9798400707223

Finding sparse representations of signals in general dictionaries is a key research topic, with many applications in signal processing and machine learning. In the sparse recovery literature, k-sparsity is a widely accepted assumption. In this work, we thrive to reach beyond the standard uniqueness conditions by considering a novel constraint that can be viewed as a generalization of k-sparsity. The proposed formalism enables the selection of a set of indices. When it is known that none of the ground truth vectors can have a support included in this set, guarantees of exact recovery can be extended. In this work, we define the rules for constructing the set mentioned above. Furthermore, the guarantees of an accurate reconstruction will be presented, establishing compatibility with our novel constraints. © 2024 Copyright held by the owner/author(s).

关键词： sparse coding

来源：评论

学校读者我要写书评

暂无评论

Learning-based high-efficiency compression framework for light field videos

引用

MULTIMEDIA TOOLS AND APPLICATIONS 2022年第6期81卷 7527-7560页

作者： Wang, Bing Xiang, Wei Wang, Eric Peng, Qiang Gao, Pan Wu, Xiao James Cook Univ Coll Sci & Engn Cairns Qld 4878 Australia Southwest Jiaotong Univ Sch Comp & Artificial Intelligence Chengdu 611756 Peoples R China La Trobe Univ Sch Engn & Math Sci Melbourne Vic 3686 Australia Nanjing Univ Aeronaut & Astronaut Coll Comp Sci & Technol Nanjing 211106 Peoples R China

The massive amount of data usage for light field (LF) information poses grand challenges for efficient compression designs. There have been several LF video compression methods focusing on exploring efficient prediction structures reported in the literature. However, the number of possible prediction structures is infinite, and these methods fail to fully exploit the intrinsic geometry between views of an LF video. In this paper, we propose a deep learning-based high-efficiency LF video compression framework by exploiting the inherent geometrical structure of LF videos. The proposed framework is composed of several crucial components, namely sparse coding based on a universal view sampling method (UVSM) and a CNN-based LF view synthesis algorithm (LF-CNN), a high-efficiency adaptive prediction structure (APS), and a synthesized candidate reference (SCR)-based inter-frame prediction strategy. Specifically, instead of encoding all the views in an LF video, only parts of views are compressed while the remaining views are reconstructed from the encoded views with LF-CNN. The prediction structure of the selected views is able to adapt itself to the similarity between views. Inspired by the effectiveness of view synthesis algorithms, synthesized results are served as additional candidate references to further reduce inter-frame redundancies. Experimental results show that the proposed LF video compression framework can achieve an average of over 34% bitrate savings against state-of-the-art LF video compression methods over multiple LF video datasets.

关键词： Light field video compression Prediction structure sparse coding View synthesis

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：