CIF:Medium:Collaborative Research: Geometric Network Information Theory

CIF：中：协作研究：几何网络信息论

• 批准号: 1704624
• 负责人: Ayfer Ozgur
• 金额: $ 45.13万
• 依托单位: Stanford University
• 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 2021-07-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1704624&HistoricalAwards=false
• 关键词: CIF; Medium; Collaborative; Research; Geometric

The proliferation of communication networks means that their efficiency and security now affects almost all aspects of our daily lives. At the societal level, reliable and efficient networks are of central importance to technological development, national economic growth and national defense. In light of this, the problem of understanding fundamental information limits of networked devices is more acute now than ever. Unfortunately, it has been a longstanding challenge in information theory to systematically extend its success beyond classical point-to-point exchanges of information. Despite significant research effort since the inception of the field in 1948, the information limits for most network settings, even for networks as fundamental as a three node relay channel, have remained open to date.This project aims to change the way many network problems are analyzed today in information theory. Traditional information-theoretic approaches for quantifying fundamental performance limits of networks often fail because the current toolset of information theory is not sufficiently refined to characterize tensions that emerge in networks due to competing objectives. Looking beyond traditional methods, this project will study the role that inequalities from geometric and functional analysis can play in characterizing tensions between information measures. The research project will be complemented with a strong inter-university education program, which includes joint and collaborative student and post-doctoral researcher advising.

通信网络的扩散意味着它们的效率和安全现在几乎影响了我们日常生活的所有方面。在社会层面，可靠和高效的网络对于技术发展，国民经济增长和国防至关重要。鉴于此，了解网络设备的基本信息限制的问题现在比以往任何时候都更加敏锐。不幸的是，在信息理论中，系统地将其成功扩展到经典的点对点信息交换之外，这是一个长期的挑战。尽管自1948年该领域成立以来的研究工作巨大，但大多数网络设置的信息限制，即使是像三个节点接力渠道这样基本的网络，也一直保持开放。该项目旨在改变今天信息理论中今天在信息理论中分析许多网络问题的方式。量化网络基本绩效限制的传统信息理论方法通常会失败，因为信息理论的当前工具集不能充分完善以表征由于竞争目标而导致网络中出现的紧张局势。 除了传统方法外，该项目将研究几何和功能分析中不平等的作用，可以在表征信息测量之间的紧张局势方面发挥作用。该研究项目将与强大的大学间教育计划相辅相成，该计划包括联合和协作学生和博士后研究人员建议。

项目成果

Understanding Entropic Regularization in GANs

了解 GAN 中的熵正则化

• DOI: 10.1109/isit45174.2021.9517791
• 发表时间: 2021
• 期刊: 2021 IEEE International Symposium on Information Theory (ISIT
• 作者: Reshetova, Daria;Bai, Yikun;Wu, Xiugang;Ozgur, Ayfer
• 通讯作者: Ozgur, Ayfer

The Courtade-Kumar Most Informative Boolean Function Conjecture and a Symmetrized Li-Médard Conjecture are Equivalent

Courtade-Kumar 信息量最大的布尔函数猜想和对称 Li-Médard 猜想是等价的

• DOI: 10.1109/isit44484.2020.9174063
• 发表时间: 2020
• 期刊: 2020 IEEE International Symposium on Information Theory (ISIT
• 作者: Barnes, Leighton Pate;Ozgur, Ayfer
• 通讯作者: Ozgur, Ayfer

Fisher Information and Mutual Information Constraints

Fisher 信息和互信息约束

• DOI: 10.1109/isit45174.2021.9518061
• 发表时间: 2021
• 期刊: 2020 IEEE International Symposium on Information Theory (ISIT
• 作者: Barnes, Leighton Pate;Ozgur, Ayfer
• 通讯作者: Ozgur, Ayfer

A Geometric Characterization of Fisher Information from Quantized Samples with Applications to Distributed Statistical Estimation

• DOI: 10.1109/allerton.2018.8635899
• 发表时间: 2018-01-01
• 期刊: 2018 56TH ANNUAL ALLERTON CONFERENCE ON COMMUNICATION, CONTROL, AND COMPUTING (ALLERTON)
• 作者: Barnes, Leighton Pate;Han, Yanjun;Ozgur, Ayfer
• 通讯作者: Ozgur, Ayfer

Information Constrained Optimal Transport: From Talagrand, to Marton, to Cover

信息约束最优运输：从塔拉格兰，到马顿，到 Cover

• DOI: 10.1109/tit.2023.3237073
• 发表时间: 2023
• 期刊: IEEE Transactions on Information Theory
• 影响因子: 2.5
• 作者: Bai, Yikun;Wu, Xiugang;Özgür, Ayfer
• 通讯作者: Özgür, Ayfer