III: Medium: Non-Convex Methods for Discovering High-Dimensional Structures in Big and Corrupted Data

III:媒介:在大数据和损坏数据中发现高维结构的非凸方法

基本信息

  • 批准号:
    1704458
  • 负责人:
  • 金额:
    $ 115万
  • 依托单位:
  • 依托单位国家:
    美国
  • 项目类别:
    Standard Grant
  • 财政年份:
    2017
  • 资助国家:
    美国
  • 起止时间:
    2017-08-01 至 2023-09-30
  • 项目状态:
    已结题

项目摘要

Discovering structure in high-dimensional data, such as images, videos and 3D point clouds, has become an essential part of scientific discovery in many disciplines, including machine learning, computer vision, pattern recognition, and signal processing. This has motivated extraordinary advances in the past decade, including various sparse and low-rank modeling methods based on convex optimization with provable theoretical guarantees of correct recovery. However, existing theory and algorithms rely on the assumption that high-dimensional data can be well approximated by low-dimensional structures. While this assumption is adequate for some datasets, e.g., images of faces under varying illumination, it may be violated in many emerging datasets, e.g., 3D point clouds. The goal of this project is to develop a mathematical modeling framework and associated non-convex optimization tools for discovering high-dimensional structures in big and corrupted data.This project will develop provably correct and scalable optimization algorithms for learning a union of high-dimensional subspaces from big and corrupted data. The proposed algorithms will be based on a novel framework called Dual Principal Component Pursuit that, instead of learning a basis for each subspace, seeks to learn a basis for their orthogonal complements. In sharp contrast with existing sparse and low-rank methods, which require both the dimensions of the subspaces and the percentage of outliers to be sufficiently small, the proposed framework will lead to results where even subspaces of the highest possible dimension (i.e., hyperplanes) can be correctly recovered from highly corrupted data. This will be achieved by solving a family of non-convex sparse representation problems whose analysis will require the development of novel theoretical results to guarantee the correct recovery of the subspaces from corrupted data. The project will also develop scalable algorithms for solving these non-convex optimization problems and study conditions for their convergence to the global optimum. These algorithms will be evaluated in two major applications in computer vision: segmentation of point clouds and clustering of image categorization datasets.
在图像、视频和3D点云等高维数据中发现结构已成为许多学科科学发现的重要组成部分,包括机器学习、计算机视觉、模式识别和信号处理。这在过去的十年中推动了非凡的进步,包括基于凸优化的各种稀疏和低秩建模方法,这些方法具有正确恢复的可证明理论保证。然而,现有的理论和算法依赖于这样的假设,即高维数据可以很好地近似低维结构。虽然这种假设对于某些数据集是足够的,例如,在变化的照明下的面部图像,它可能在许多新兴的数据集中被违反,例如,三维点云。该项目的目标是开发一个数学建模框架和相关的非凸优化工具,用于发现大数据和损坏数据中的高维结构。该项目将开发可证明正确和可扩展的优化算法,用于从大数据和损坏数据中学习高维子空间的并集。所提出的算法将基于一种称为双主成分追踪的新框架,该框架不是学习每个子空间的基础,而是寻求学习其正交互补的基础。与现有的稀疏和低秩方法形成鲜明对比的是,现有的稀疏和低秩方法要求子空间的维度和离群值的百分比都足够小,所提出的框架将导致这样的结果,即使是最高可能维度的子空间(即,超平面)可以从高度损坏的数据中正确地恢复。这将通过解决一系列非凸稀疏表示问题来实现,这些问题的分析将需要开发新的理论结果,以保证从损坏的数据中正确恢复子空间。该项目还将开发解决这些非凸优化问题的可扩展算法,并研究其收敛到全局最优值的条件。这些算法将在计算机视觉中的两个主要应用中进行评估:点云的分割和图像分类数据集的聚类。

项目成果

期刊论文数量(6)
专著数量(0)
科研奖励数量(0)
会议论文数量(0)
专利数量(0)
A Critique of Self-Expressive Deep Subspace Clustering
  • DOI:
  • 发表时间:
    2020-10
  • 期刊:
  • 影响因子:
    0
  • 作者:
    B. Haeffele;Chong You;R. Vidal
  • 通讯作者:
    B. Haeffele;Chong You;R. Vidal
A nonconvex formulation for low rank subspace clustering: algorithms and convergence analysis
  • DOI:
    10.1007/s10589-018-0002-6
  • 发表时间:
    2018-03
  • 期刊:
  • 影响因子:
    2.2
  • 作者:
    Hao Jiang;Daniel P. Robinson;R. Vidal;Chong You
  • 通讯作者:
    Hao Jiang;Daniel P. Robinson;R. Vidal;Chong You
A novel variational form of the Schatten-p quasi-norm
  • DOI:
  • 发表时间:
    2020-10
  • 期刊:
  • 影响因子:
    0
  • 作者:
    Paris V. Giampouras;R. Vidal;A. Rontogiannis;B. Haeffele
  • 通讯作者:
    Paris V. Giampouras;R. Vidal;A. Rontogiannis;B. Haeffele
Dual Principal Component Pursuit for Learning a Union of Hyperplanes: Theory and Algorithms
  • DOI:
  • 发表时间:
    2021
  • 期刊:
  • 影响因子:
    0
  • 作者:
    Tianyu Ding;Zhihui Zhu;M. Tsakiris;R. Vidal;Daniel P. Robinson
  • 通讯作者:
    Tianyu Ding;Zhihui Zhu;M. Tsakiris;R. Vidal;Daniel P. Robinson
The fastest L1,oo prox in the west
西方最快的L1,oo prox
{{ 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 }}

Rene Vidal其他文献

Rene Vidal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('Rene Vidal', 18)}}的其他基金

Collaborative Research: SCH: Multimodal Algorithms for Motor Imitation Assessment in Children with Autism
合作研究:SCH:自闭症儿童运动模仿评估的多模式算法
  • 批准号:
    2124277
  • 财政年份:
    2021
  • 资助金额:
    $ 115万
  • 项目类别:
    Standard Grant
Collaborative Research: Transferable, Hierarchical, Expressive, Optimal, Robust, Interpretable Networks
协作研究:可转移、分层、富有表现力、最优、稳健、可解释的网络
  • 批准号:
    2031985
  • 财政年份:
    2020
  • 资助金额:
    $ 115万
  • 项目类别:
    Continuing Grant
HDR TRIPODS: Institute for the Foundations of Graph and Deep Learning
HDR TRIPODS:图形和深度学习基础研究所
  • 批准号:
    1934979
  • 财政年份:
    2019
  • 资助金额:
    $ 115万
  • 项目类别:
    Continuing Grant
RI: Small: An Optimization Framework for Understanding Deep Networks
RI:小型:理解深度网络的优化框架
  • 批准号:
    1618485
  • 财政年份:
    2016
  • 资助金额:
    $ 115万
  • 项目类别:
    Standard Grant
CIF: Small: Collaborative Research: Sparse and Low Rank Methods for Imbalanced and Heterogeneous Data
CIF:小型:协作研究:针对不平衡和异构数据的稀疏和低秩方法
  • 批准号:
    1618637
  • 财政年份:
    2016
  • 资助金额:
    $ 115万
  • 项目类别:
    Standard Grant
RI: Small: Object Detection, Pose Estimation, and Semantic Segmentation Using 3D Wireframe Models
RI:小:使用 3D 线框模型进行物体检测、姿势估计和语义分割
  • 批准号:
    1527340
  • 财政年份:
    2015
  • 资助金额:
    $ 115万
  • 项目类别:
    Continuing Grant
BIGDATA: F: DKA: Learning a Union of Subspaces from Big and Corrupted Data
BIGDATA:F:DKA:从大数据和损坏数据中学习子空间并集
  • 批准号:
    1447822
  • 财政年份:
    2014
  • 资助金额:
    $ 115万
  • 项目类别:
    Standard Grant
Geometry and Statistics on Spaces of Dynamical Systems for Pattern Recognition in High-Dimensional Time Series
用于高维时间序列模式识别的动力系统空间的几何和统计
  • 批准号:
    1335035
  • 财政年份:
    2013
  • 资助金额:
    $ 115万
  • 项目类别:
    Standard Grant
RI: Small: Structured Sparse Conditional Random Fields Models for Joint Categorization and Segmentation of Objects.
RI:小型:用于对象联合分类和分割的结构化稀疏条件随机场模型。
  • 批准号:
    1218709
  • 财政年份:
    2012
  • 资助金额:
    $ 115万
  • 项目类别:
    Standard Grant
CDI-Type I: Collaborative Research: A Bio-Inspired Approach to Recognition of Human Movements and Movement Styles
CDI-I 型:协作研究:识别人类运动和运动风格的仿生方法
  • 批准号:
    0941463
  • 财政年份:
    2010
  • 资助金额:
    $ 115万
  • 项目类别:
    Standard Grant

相似海外基金

Welsh-Medium Education: Exploring the attitudes, beliefs and experiences of parents and pupils from non-Welsh speaking households.
威尔士语教育:探索非威尔士语家庭的家长和学生的态度、信仰和经历。
  • 批准号:
    2863855
  • 财政年份:
    2023
  • 资助金额:
    $ 115万
  • 项目类别:
    Studentship
Detection of microRNAs in the conditioned culture medium as a non-invasive "miR-print" of human preimplantation embryo competence
检测条件培养基中的 microRNA,作为人类植入前胚胎能力的非侵入性“miR-print”
  • 批准号:
    488408
  • 财政年份:
    2023
  • 资助金额:
    $ 115万
  • 项目类别:
    Operating Grants
Coupling of Modified Equation of State and Percolation Theory to Study Static and Dynamic Non-Equilibrium Phase Behavior of Heavy Oil in the Presence of Porous Medium
修正状态方程与渗流理论耦合研究多孔介质中稠油静态和动态非平衡相行为
  • 批准号:
    RGPIN-2019-06103
  • 财政年份:
    2022
  • 资助金额:
    $ 115万
  • 项目类别:
    Discovery Grants Program - Individual
Collaborative Research: HCC: Medium: Intelligent support for non-experts to navigate large information spaces
协作研究:HCC:中:为非专家导航大型信息空间提供智能支持
  • 批准号:
    2106882
  • 财政年份:
    2021
  • 资助金额:
    $ 115万
  • 项目类别:
    Standard Grant
Collaborative Research: HCC: Medium: RUI: Intelligent support for non-experts to navigate large information spaces
协作研究:HCC:中:RUI:为非专家导航大型信息空间提供智能支持
  • 批准号:
    2106896
  • 财政年份:
    2021
  • 资助金额:
    $ 115万
  • 项目类别:
    Standard Grant
Collaborative Research: CNS Core: Medium: Understanding and Strengthening Memory Security for Non-Volatile Memory
合作研究:CNS 核心:中:理解和加强非易失性内存的内存安全性
  • 批准号:
    2106629
  • 财政年份:
    2021
  • 资助金额:
    $ 115万
  • 项目类别:
    Continuing Grant
Collaborative Research: HCC: Medium: Intelligent support for non-experts to navigate large information spaces
协作研究:HCC:中:为非专家导航大型信息空间提供智能支持
  • 批准号:
    2107334
  • 财政年份:
    2021
  • 资助金额:
    $ 115万
  • 项目类别:
    Standard Grant
SHF: Medium: Cross-Cutting Effort to Make Non-Volatile Memories Truly Usable
SHF:中:使非易失性存储器真正可用的跨领域努力
  • 批准号:
    2107470
  • 财政年份:
    2021
  • 资助金额:
    $ 115万
  • 项目类别:
    Continuing Grant
Collaborative Research: CNS Core: Medium: Understanding and Strengthening Memory Security for Non-Volatile Memory
合作研究:CNS 核心:中:理解和加强非易失性内存的内存安全性
  • 批准号:
    2106893
  • 财政年份:
    2021
  • 资助金额:
    $ 115万
  • 项目类别:
    Continuing Grant
Coupling of Modified Equation of State and Percolation Theory to Study Static and Dynamic Non-Equilibrium Phase Behavior of Heavy Oil in the Presence of Porous Medium
修正状态方程与渗流理论耦合研究多孔介质中稠油静态和动态非平衡相行为
  • 批准号:
    RGPIN-2019-06103
  • 财政年份:
    2021
  • 资助金额:
    $ 115万
  • 项目类别:
    Discovery Grants Program - Individual
{{ showInfoDetail.title }}

作者:{{ showInfoDetail.author }}

知道了