ATD: Landscape Networks and Nonlinear Diffusions for Anomaly Detection and Active Learning
ATD:用于异常检测和主动学习的景观网络和非线性扩散
基本信息
- 批准号:1924513
- 负责人:
- 金额:$ 15.79万
- 依托单位:
- 依托单位国家:美国
- 项目类别:Standard Grant
- 财政年份:2019
- 资助国家:美国
- 起止时间:2019-07-15 至 2023-06-30
- 项目状态:已结题
- 来源:
- 关键词:
项目摘要
Statistical and machine learning are revolutionizing scientific fields ranging from computer vision, to medicine, to natural language processing, and inference of natural physical laws. Despite these rapid and impressive empirical advances, machine learning remains only partially understood mathematically. In particular, unsupervised anomaly detection in which algorithms must distinguish background from anomaly with no labeled data and active learning in which only a very small but carefully selected number of points may be queried for labels are ripe for transformational advances. As sensors generate ever increasing datasets, the sheer volume of data overwhelms human capacity for generating the kinds of large training sets necessary for traditional supervised learning algorithms. The future of machine learning relies on developing new mathematical approaches to unsupervised and active learning, where no or little training data is required. Innovations in this direction have potential to transform fields as diverse as computational medicine, network security, and image processing. This project will support 1 graduate student in the second and third years of the grant.This research project develops new algorithms for anomaly detection and active learning in spatiotemporal data. The emphasis is on the analysis of high-dimensional, time-evolving data sets in a manner that is robust to nonlinear geometries, variable sampling rates, and large quantities of noise and outliers. The PI proposes two distinct but related lines of research. First, to devise multitemporal anomaly detection algorithms using landscape cluster networks. This approach handles temporally varying distributions and labels clusters and anomalies at different levels of granularity, providing confidence estimates and uncertainty quantifications. Second, diffusion geometric active learning algorithms for spatiotemporal data will be developed to allow a human analyst to label a small number of queries from the algorithm. These queries are carefully chosen, and the labels provided by the human analyst can radically improve cluster and anomaly detection at minimal computational burden. The proposed methods are robust to complicated data geometries, temporal sampling rates, noise and outliers, and ambient dimensionality of the data. Beyond the topics of machine learning, this project makes broader contributions to probability theory, harmonic analysis, spectral graph theory, high-dimensional statistics, and computational linear algebra. Mathematical and algorithmic contributions will be developed in parallel with scientific collaborations analyzing large spatiotemporal datasets. This project focuses on anomaly detection and active learning in three distinct spatiotemporal data settings: large-scale commuting networks, hyperspectral image analysis, and high energy particle physics. The proposed methods allow for real-time anomaly and threat detection, are scalable, and mitigate the need for large training sets.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.
统计学和机器学习正在给科学领域带来革命性的变化,从计算机视觉到医学,再到自然语言处理和自然物理定律的推理。尽管取得了这些快速而令人印象深刻的经验进展,但机器学习在数学上仍然只有部分理解。特别是,算法必须在没有标记数据的情况下区分背景和异常的无监督异常检测,以及其中仅可以查询非常少量但仔细选择的点来寻找标记的主动学习,对于变换进展是成熟的。随着传感器产生越来越多的数据集,庞大的数据量超过了人类生成传统监督学习算法所需的大型训练集的能力。机器学习的未来依赖于开发新的无监督和主动学习的数学方法,在这种方法中,不需要或只需要很少的训练数据。这一方向的创新具有改变计算医学、网络安全和图像处理等多个领域的潜力。这个项目将支持一名研究生在第二年和第三年的资助。这个研究项目开发了时空数据中异常检测和主动学习的新算法。重点是以一种对非线性几何、可变采样率以及大量噪声和离群值具有健壮性的方式分析高维、时间演变的数据集。PI提出了两条截然不同但相互关联的研究路线。首先,设计基于景观集群网络的多时相异常检测算法。这种方法处理随时间变化的分布,并在不同的粒度级别标记集群和异常,提供置信度估计和不确定性量化。其次,将开发用于时空数据的扩散几何主动学习算法,以允许人类分析师从该算法中标记少量查询。这些查询经过仔细的选择,由人类分析师提供的标签可以在最小的计算负担下从根本上改进集群和异常检测。所提出的方法对复杂的数据几何形状、时间采样率、噪声和离群值以及数据的环境维度具有很强的鲁棒性。除了机器学习的主题,这个项目对概率论、调和分析、谱图理论、高维统计和计算线性代数做出了更广泛的贡献。数学和算法方面的贡献将与分析大型时空数据集的科学合作并行发展。该项目专注于在三个不同的时空数据环境中进行异常检测和主动学习:大规模通勤网络、高光谱图像分析和高能粒子物理。建议的方法允许实时异常和威胁检测,可扩展,并减少了对大型训练集的需求。该奖项反映了NSF的法定使命,并通过使用基金会的智力优势和更广泛的影响审查标准进行评估,被认为值得支持。
项目成果
期刊论文数量(24)
专著数量(0)
科研奖励数量(0)
会议论文数量(0)
专利数量(0)
GLIDE: combining local methods and diffusion state embeddings to predict missing interactions in biological networks
- DOI:10.1093/bioinformatics/btaa459
- 发表时间:2020-07-01
- 期刊:
- 影响因子:5.8
- 作者:Devkota, Kapil;Murphy, James M.;Cowen, Lenore J.
- 通讯作者:Cowen, Lenore J.
Multiscale Clustering of Hyperspectral Images Through Spectral-Spatial Diffusion Geometry
- DOI:10.1109/igarss47720.2021.9554397
- 发表时间:2021-03
- 期刊:
- 影响因子:0
- 作者:Sam L. Polk;James M. Murphy
- 通讯作者:Sam L. Polk;James M. Murphy
Diffusion State Distances: Multitemporal Analysis, Fast Algorithms, and Applications to Biological Networks
- DOI:10.1137/20m1324089
- 发表时间:2021-01-01
- 期刊:
- 影响因子:3.6
- 作者:Cowen, Lenore;Devkota, Kapil;Wu, Kaiyi
- 通讯作者:Wu, Kaiyi
Patch-Based Diffusion Learning for Hyperspectral Image Clustering
基于补丁的高光谱图像聚类扩散学习
- DOI:10.1109/igarss39084.2020.9323091
- 发表时间:2020
- 期刊:
- 影响因子:0
- 作者:Murphy, James M.
- 通讯作者:Murphy, James M.
Spectral–Spatial Diffusion Geometry for Hyperspectral Image Clustering
- DOI:10.1109/lgrs.2019.2943001
- 发表时间:2019-02
- 期刊:
- 影响因子:4.8
- 作者:James M. Murphy;M. Maggioni
- 通讯作者:James M. Murphy;M. Maggioni
{{
item.title }}
{{ item.translation_title }}
- DOI:
{{ item.doi }} - 发表时间:
{{ item.publish_year }} - 期刊:
- 影响因子:{{ item.factor }}
- 作者:
{{ item.authors }} - 通讯作者:
{{ item.author }}
数据更新时间:{{ journalArticles.updateTime }}
{{ item.title }}
- 作者:
{{ item.author }}
数据更新时间:{{ monograph.updateTime }}
{{ item.title }}
- 作者:
{{ item.author }}
数据更新时间:{{ sciAawards.updateTime }}
{{ item.title }}
- 作者:
{{ item.author }}
数据更新时间:{{ conferencePapers.updateTime }}
{{ item.title }}
- 作者:
{{ item.author }}
数据更新时间:{{ patent.updateTime }}
James Murphy其他文献
Dental, Oral, and Maxillofacial Diseases and Conditions and Their Treatment
牙科、口腔和颌面疾病和病症及其治疗
- DOI:
- 发表时间:
2018 - 期刊:
- 影响因子:0
- 作者:
B. Cornwall;K. Marti;C. Skouteris;James Murphy;B. Ward;I. Makovey;S. Edwards - 通讯作者:
S. Edwards
CARDINAL AND ORDINAL NUMBERS
- DOI:
10.1007/978-0-387-22767-2_1 - 发表时间:
2009 - 期刊:
- 影响因子:0
- 作者:
James Murphy - 通讯作者:
James Murphy
"I'd be watching him contour till 10 o'clock at night": Understanding Tensions between Teaching Methods and Learning Needs in Healthcare Apprenticeship
“我会看着他的轮廓直到晚上 10 点”:理解医疗学徒培训中教学方法和学习需求之间的紧张关系
- DOI:
- 发表时间:
2024 - 期刊:
- 影响因子:0
- 作者:
M. Yarmand;Chen Chen;Kexin Cheng;James Murphy;Nadir Weibel - 通讯作者:
Nadir Weibel
Ocular hypertension following 40 mg sub-Tenon triamcinolone versus 0.7 mg dexamethasone implant versus 2 mg intravitreal triamcinolone.
40 mg sub-Tenon 曲安西龙对比 0.7 mg 地塞米松植入物对比 2 mg 玻璃体内曲安西龙后出现高眼压。
- DOI:
10.1016/j.jcjo.2020.06.021 - 发表时间:
2020 - 期刊:
- 影响因子:0
- 作者:
Brandon Kuley;Philip P Storey;Maitri Pancholy;Nicholas Bello;James Murphy;J. Goodman;T. Wibbelsman;Anthony Obeid;A. Chiang;C. Regillo;Sunir J. Garg - 通讯作者:
Sunir J. Garg
Gluteal fold flaps for perineal reconstruction.
用于会阴重建的臀皱皮瓣。
- DOI:
- 发表时间:
2013 - 期刊:
- 影响因子:0
- 作者:
R. Winterton;G. Lambe;C. Ekwobi;D. Oudit;D. Mowatt;James Murphy;G. Ross - 通讯作者:
G. Ross
James Murphy的其他文献
{{
item.title }}
{{ item.translation_title }}
- DOI:
{{ item.doi }} - 发表时间:
{{ item.publish_year }} - 期刊:
- 影响因子:{{ item.factor }}
- 作者:
{{ item.authors }} - 通讯作者:
{{ item.author }}
{{ truncateString('James Murphy', 18)}}的其他基金
Doctoral Dissertation Research: Medium-scale farming systems and agricultural entrepreneuership
博士论文研究:中等规模农业系统与农业创业
- 批准号:
2233591 - 财政年份:2023
- 资助金额:
$ 15.79万 - 项目类别:
Standard Grant
ATD: Diffusion and Transport on Graphs: Active Learning, Low-Dimensional Representations, and Anomaly Detection
ATD:图上的扩散和传输:主动学习、低维表示和异常检测
- 批准号:
2318894 - 财政年份:2023
- 资助金额:
$ 15.79万 - 项目类别:
Standard Grant
Towards Harmonic Analysis in Wasserstein Space: Low-Dimensional Structures, Learning, and Algorithms
Wasserstein 空间中的调和分析:低维结构、学习和算法
- 批准号:
2309519 - 财政年份:2023
- 资助金额:
$ 15.79万 - 项目类别:
Continuing Grant
Collaborative Research: Data-driven Path Metrics for Machine Learning
协作研究:机器学习的数据驱动路径度量
- 批准号:
1912737 - 财政年份:2019
- 资助金额:
$ 15.79万 - 项目类别:
Standard Grant
Doctoral Dissertation Research: Assembling Community Economies
博士论文研究:整合社区经济
- 批准号:
1655094 - 财政年份:2017
- 资助金额:
$ 15.79万 - 项目类别:
Standard Grant
Doctoral Dissertation Research: National Integration or Regional Competition? Industrial Policy Debates in a Rising Power.
博士论文研究:国家一体化还是区域竞争?
- 批准号:
1234594 - 财政年份:2012
- 资助金额:
$ 15.79万 - 项目类别:
Standard Grant
Doctoral Dissertation Research: Electronic Waste Recycling in South Africa: Transition Management in Practice?
博士论文研究:南非的电子废物回收:实践中的转型管理?
- 批准号:
0927837 - 财政年份:2009
- 资助金额:
$ 15.79万 - 项目类别:
Standard Grant
The Role of Information-Communication Technologies in Enterprise Development and Industrial Change in Africa: Evidence from South Africa and Tanzania
信息通信技术在非洲企业发展和产业变革中的作用:来自南非和坦桑尼亚的证据
- 批准号:
0925151 - 财政年份:2009
- 资助金额:
$ 15.79万 - 项目类别:
Standard Grant
The Socio-Spatial Dimensions of Industrial Change in Bolivia: Manufacturers, Regions, and the Prospects for Global Value Chain Integration
玻利维亚产业变革的社会空间维度:制造商、地区和全球价值链一体化的前景
- 批准号:
0616030 - 财政年份:2006
- 资助金额:
$ 15.79万 - 项目类别:
Standard Grant
NSF/AFOSR Astronomy: Spatial and Temporal Variations in the Atmospheric Aerosol Content of Mars, Jupiter, and Saturn
NSF/AFOSR 天文学:火星、木星和土星大气气溶胶含量的时空变化
- 批准号:
0335665 - 财政年份:2003
- 资助金额:
$ 15.79万 - 项目类别:
Standard Grant
相似海外基金
Collaborative Research: Dynamic connectivity of river networks as a framework for identifying controls on flux propagation and assessing landscape vulnerability to change
合作研究:河流网络的动态连通性作为识别通量传播控制和评估景观变化脆弱性的框架
- 批准号:
2342936 - 财政年份:2024
- 资助金额:
$ 15.79万 - 项目类别:
Continuing Grant
Collaborative Research: Dynamic connectivity of river networks as a framework for identifying controls on flux propagation and assessing landscape vulnerability to change
合作研究:河流网络的动态连通性作为识别通量传播控制和评估景观变化脆弱性的框架
- 批准号:
2342937 - 财政年份:2024
- 资助金额:
$ 15.79万 - 项目类别:
Continuing Grant
EAGER: Quantifying the error landscape of deep neural networks
EAGER:量化深度神经网络的错误情况
- 批准号:
2226387 - 财政年份:2022
- 资助金额:
$ 15.79万 - 项目类别:
Standard Grant
EAGER: Quantifying the error landscape of deep neural networks
EAGER:量化深度神经网络的错误情况
- 批准号:
2132995 - 财政年份:2021
- 资助金额:
$ 15.79万 - 项目类别:
Standard Grant
CAREER: Optimization Landscape for Non-convex Functions - Towards Provable Algorithms for Neural Networks
职业:非凸函数的优化景观 - 走向可证明的神经网络算法
- 批准号:
1845171 - 财政年份:2019
- 资助金额:
$ 15.79万 - 项目类别:
Continuing Grant
Learning to recognize landslides and catastrophic landscape change with deep neural networks
学习利用深度神经网络识别山体滑坡和灾难性景观变化
- 批准号:
2292269 - 财政年份:2019
- 资助金额:
$ 15.79万 - 项目类别:
Studentship
Olympia (Greece) as a multi-dimensional space – a landscape archaeology approach to structure, interdependency and change of spatial networks
奥林匹亚(希腊)作为一个多维空间——一种研究空间网络结构、相互依赖性和变化的景观考古学方法
- 批准号:
270113181 - 财政年份:2015
- 资助金额:
$ 15.79万 - 项目类别:
Research Grants
Ubiquitous multimedia service provisioning over heterogeneous wireless networks in a diverse mobility landscape
在多样化的移动环境中通过异构无线网络提供无处不在的多媒体服务
- 批准号:
386118-2010 - 财政年份:2014
- 资助金额:
$ 15.79万 - 项目类别:
Discovery Grants Program - Individual
RUI: Collaborative Research: Pattern Emergence and Resilience of Rill Networks and Their Relation to Soil Loss, Landscape Degradation, and Erosion Prediction Technology
RUI:合作研究:细沟网络的模式出现和恢复力及其与土壤流失、景观退化和侵蚀预测技术的关系
- 批准号:
1359852 - 财政年份:2014
- 资助金额:
$ 15.79万 - 项目类别:
Continuing Grant
Collaborative Research: Pattern Emergence and Resilience of Rill Networks and Their Relation to Soil Loss, Landscape Degradation, and Erosion Prediction Technology
合作研究:细沟网络的模式出现和恢复力及其与土壤流失、景观退化和侵蚀预测技术的关系
- 批准号:
1359904 - 财政年份:2014
- 资助金额:
$ 15.79万 - 项目类别:
Continuing Grant