Approximate Message Passing Algorithms and Networks
近似消息传递算法和网络
基本信息
- 批准号:1716388
- 负责人:
- 金额:$ 49.96万
- 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:2017-07-01 至 2021-06-30
- 项目状态:已结题
- 来源:
- 关键词:
项目摘要
A problem of paramount importance in engineering, science, and medicine is that of recovering information signals from high-dimensional measurements. This problem manifests in many forms, e.g., reconstructing a high-quality image from a few noisy Fourier projections, determining which features in patient data are most likely associated with a given disease, or classifying which objects are present within an image. Until recently, the dominant approach to signal recovery was algorithmic. But nowadays, algorithms are increasingly being replaced by deep neural networks (DNNs), which can learn optimal inference strategies directly from the data. This project researches algorithmic as well as deep-neural-network (DNN) approaches to high-dimensional signal recovery, leveraging connections between them to make advances in both.On the algorithmic front, this project investigates the vector approximate message passing (VAMP) algorithm. Like the original AMP algorithm of Donoho, Maleki, and Montanari, the VAMP algorithm enjoys low complexity and a scalar state-evolution that rigorously and concisely characterizes its behavior. However, VAMP is applicable to a much larger class of problems than AMP. On the DNN front, this project investigates DNNs whose architecture is inspired by the processing steps within VAMP. The resulting DNNs are highly interpretable and, for some simple applications, statistical optimal. This project aims to develop this VAMP-based DNN design framework to work with more complex applications.
工程、科学和医学中最重要的问题是从高维测量中恢复信息信号的问题。这个问题表现在许多形式上,例如,从几个噪声傅立叶投影重建高质量的图像,确定患者数据中的哪些特征最有可能与给定的疾病相关联,或者分类图像中存在哪些对象。直到最近,信号恢复的主导方法还是算法。但如今,算法正越来越多地被深度神经网络(DNN)所取代,深度神经网络可以直接从数据中学习最优推理策略。本课题研究了算法和深度神经网络(DNN)在高维信号恢复中的应用,充分利用两者之间的联系,在算法方面研究了向量近似消息传递(VAMP)算法。与Donoho、Maleki和Montanari的原始AMP算法一样,VAMP算法具有低复杂度和标量状态演化的特点,严格而简洁地表征了其行为。然而,VAMP适用于比AMP大得多的一类问题。在DNN方面,本项目研究其架构灵感来自VAMP中的处理步骤的DNN。得到的DNN是高度可解释的,并且对于一些简单的应用来说,是统计最优的。该项目旨在开发基于VAMP的DNN设计框架,以处理更复杂的应用程序。
项目成果
期刊论文数量(21)
专著数量(0)
科研奖励数量(0)
会议论文数量(0)
专利数量(0)
prDeep: Robust Phase Retrieval with a Flexible Deep Network
- DOI:
- 发表时间:2018-03
- 期刊:
- 影响因子:0
- 作者:Christopher A. Metzler;Philip Schniter;A. Veeraraghavan;Richard Baraniuk
- 通讯作者:Christopher A. Metzler;Philip Schniter;A. Veeraraghavan;Richard Baraniuk
Plug in estimation in high dimensional linear inverse problems a rigorous analysis
- DOI:10.1088/1742-5468/ab321a
- 发表时间:2018-06
- 期刊:
- 影响因子:0
- 作者:A. Fletcher;S. Rangan;Subrata Sarkar;P. Schniter
- 通讯作者:A. Fletcher;S. Rangan;Subrata Sarkar;P. Schniter
On the Convergence of Approximate Message Passing With Arbitrary Matrices
- DOI:10.1109/tit.2019.2913109
- 发表时间:2014-02
- 期刊:
- 影响因子:2.5
- 作者:S. Rangan;P. Schniter;A. Fletcher;Subrata Sarkar
- 通讯作者:S. Rangan;P. Schniter;A. Fletcher;Subrata Sarkar
Autotuning Plug-and-Play Algorithms for MRI
- DOI:10.1109/ieeeconf51394.2020.9443493
- 发表时间:2020-11
- 期刊:
- 影响因子:0
- 作者:S. K. Shastri;R. Ahmad;P. Schniter
- 通讯作者:S. K. Shastri;R. Ahmad;P. Schniter
Non-Coherent Multi-User Detection Based on Expectation Propagation
基于期望传播的非相干多用户检测
- DOI:10.1109/ieeeconf44664.2019.9049073
- 发表时间:2019
- 期刊:
- 影响因子:0
- 作者:Ngo, Khac-Hoang;Guillaud, Maxime;Decurninge, Alexis;Yang, Sheng;Sarkar, Subrata;Schniter, Philip
- 通讯作者:Schniter, Philip
{{
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 }}
Philip Schniter其他文献
6 Equalization of Time-Varying Channels
6 时变通道的均衡
- DOI:
- 发表时间:
2010 - 期刊:
- 影响因子:0
- 作者:
Philip Schniter;S. Hwang;Sibasish Das;A. P. Kannu - 通讯作者:
A. P. Kannu
Iterative Frequency-Domain Channel Estimation and Equalization for Single-Carrier Transmissions without Cyclic-Pre x
无循环预置的单载波传输的迭代频域信道估计和均衡
- DOI:
- 发表时间:
2009 - 期刊:
- 影响因子:0
- 作者:
Hong Liu;Philip Schniter - 通讯作者:
Philip Schniter
Beamforming and Combining Strategies for MIMO-OFDM over Doubly Selective Channels
双选信道 MIMO-OFDM 的波束成形和组合策略
- DOI:
10.1109/acssc.2006.354860 - 发表时间:
2006 - 期刊:
- 影响因子:0
- 作者:
Sibasish Das;Philip Schniter - 通讯作者:
Philip Schniter
On the design of non-(bi)orthogonal pulse-shaped FDM for doubly-dispersive channels
双色散通道非(双)正交脉冲整形FDM设计
- DOI:
10.1109/icassp.2004.1326670 - 发表时间:
2004 - 期刊:
- 影响因子:0
- 作者:
Philip Schniter - 通讯作者:
Philip Schniter
Exploiting structured sparsity in Bayesian experimental design
在贝叶斯实验设计中利用结构化稀疏性
- DOI:
- 发表时间:
2011 - 期刊:
- 影响因子:0
- 作者:
Philip Schniter - 通讯作者:
Philip Schniter
Philip Schniter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('Philip Schniter', 18)}}的其他基金
Collaborative Research: CIF: Medium: Learning and Inference in High-Dimensional Models: Rigorous Analysis and Applications
合作研究:CIF:中:高维模型中的学习和推理:严谨的分析和应用
- 批准号:
1955587 - 财政年份:2020
- 资助金额:
$ 49.96万 - 项目类别:
Continuing Grant
CIF: Small: Collaborative Research: Next Generation Communications with Low-Resolution ADCs: Fundamentals and Practical Design
CIF:小型:协作研究:采用低分辨率 ADC 的下一代通信:基础知识和实用设计
- 批准号:
1527162 - 财政年份:2015
- 资助金额:
$ 49.96万 - 项目类别:
Standard Grant
Message-Passing Strategies for High-Dimensional Inference
高维推理的消息传递策略
- 批准号:
1218754 - 财政年份:2012
- 资助金额:
$ 49.96万 - 项目类别:
Standard Grant
CIF: Small: Soft Inference under Structured Sparsity
CIF:小:结构化稀疏下的软推理
- 批准号:
1018368 - 财政年份:2010
- 资助金额:
$ 49.96万 - 项目类别:
Standard Grant
CAREER: Signal Processing for Practical Data Communication over the Doubly-Selective Wireless Channel
职业:双选无线信道上实际数据通信的信号处理
- 批准号:
0237037 - 财政年份:2003
- 资助金额:
$ 49.96万 - 项目类别:
Continuing Grant
相似海外基金
EAGER: Scalable Climate Modeling using Message-Passing Recurrent Neural Networks
EAGER:使用消息传递循环神经网络进行可扩展的气候建模
- 批准号:
2335773 - 财政年份:2023
- 资助金额:
$ 49.96万 - 项目类别:
Standard Grant
On Principles of Distributed Computing for Message-Passing, Shared-Memory, and Hybrid Systems
消息传递、共享内存和混合系统的分布式计算原理
- 批准号:
RGPIN-2022-03304 - 财政年份:2022
- 资助金额:
$ 49.96万 - 项目类别:
Discovery Grants Program - Individual
Implementation of a message passing concurrent categorical programming language
消息传递并发分类编程语言的实现
- 批准号:
563918-2021 - 财政年份:2021
- 资助金额:
$ 49.96万 - 项目类别:
University Undergraduate Student Research Awards
A Formal Concurrent Programming Language for Message Passing
一种用于消息传递的正式并发编程语言
- 批准号:
551545-2020 - 财政年份:2020
- 资助金额:
$ 49.96万 - 项目类别:
University Undergraduate Student Research Awards
Message Passing-Based Coding for Unsourced Random Access
用于无源随机访问的基于消息传递的编码
- 批准号:
550801-2020 - 财政年份:2020
- 资助金额:
$ 49.96万 - 项目类别:
University Undergraduate Student Research Awards
CRII: CIF: Approximate Message Passing Algorithms for High-Dimensional Estimation
CRII:CIF:高维估计的近似消息传递算法
- 批准号:
1849883 - 财政年份:2019
- 资助金额:
$ 49.96万 - 项目类别:
Standard Grant
Theoretical performance limits for message passing algorithms
消息传递算法的理论性能限制
- 批准号:
2104975 - 财政年份:2018
- 资助金额:
$ 49.96万 - 项目类别:
Studentship
Iterative Signal Recovery Algorithms --- A Unified View of Turbo and Message-Passing Approaches
迭代信号恢复算法——Turbo 和消息传递方法的统一视图
- 批准号:
404179757 - 财政年份:2018
- 资助金额:
$ 49.96万 - 项目类别:
Priority Programmes
SHF: Medium: Collaborative Research: Next-Generation Message Passing for Parallel Programming: Resiliency, Time-to-Solution, Performance-Portability, Scalability, and QoS
SHF:中:协作研究:并行编程的下一代消息传递:弹性、解决时间、性能可移植性、可扩展性和 QoS
- 批准号:
1822191 - 财政年份:2017
- 资助金额:
$ 49.96万 - 项目类别:
Continuing Grant
CAREER: Hardware Accelerated Bayesian Inference via Approximate Message Passing: A Bottom-Up Approach
职业:通过近似消息传递进行硬件加速贝叶斯推理:自下而上的方法
- 批准号:
1652065 - 财政年份:2017
- 资助金额:
$ 49.96万 - 项目类别:
Continuing Grant