HDR TRIPODS: Collaborative Research: Institute for Data, Econometrics, Algorithms and Learning

HDR TRIPODS：协作研究：数据、计量经济学、算法和学习研究所

• 批准号: 1934813
• 负责人: Chao Gao
• 金额: $ 15.46万
• 依托单位: University of Chicago
• 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: 2019-09-15 至 2023-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1934813&HistoricalAwards=false
• 关键词: HDR; TRIPODS; Collaborative; Research; Institute

The Institute for Data, Econometrics, Algorithms, and Learning (IDEAL) is a multi-discipline (computer science, statistics, economics, electrical engineering, and operations research) and multi-institution (Northwestern University, Toyota Technological Institute at Chicago, and University of Chicago) collaborative institute that focuses on key aspects of the theoretical foundations of data science. The institute will support the study of foundational problems related to machine learning, high-dimensional data analysis and optimization in both strategic and non-strategic environments. The primary activity of the institute will be thematically focused quarters which will coordinate graduate course work with workshops and external visitors. The institute will facilitate collaboration between Chicago-area institutions through a number of initiatives, and across multiple disciplines. Several components of the research agenda have direct applications areas, and the PIs will involve practitioners in development economics, online markets, public policy, as well as data scientists. The research areas supported by the institute focus on three broad themes: (1) High dimensional data analysis, to address algorithmic and statistical challenges in dealing with high dimensional data, and investigate topics like metric embeddings, sketching, and problems in unsupervised learning; (2) Data Science in Strategic Environments, to address computational and information theoretic challenges in econometric models of strategic behavior like inference on high-dimensional structural parameter spaces, dealing with unobserved heterogeneity, partial identification, and machine learning in econometrics; and (3) Machine learning and optimization, to address foundational questions in both continuous and discrete optimization and its use in machine learning including topics like representation learning, robustness in learning, and provable bounds for non-convex optimization. Initially, six research topics will be selected that tie interests across the institutions: inference and data science on networks; theory of deep learning; incentives in shared data infrastructure; robustness in high-dimensional statistics; high-dimensional data analysis; and algorithms for partially identified models. There will be special quarters (fall and spring) where the Institute will bring together investigators, postdocs, and Ph.D. students to focus on one of the topics. In the following quarter (winter and summer) teams will continue research that advance the proposal topics.This project is part of the National Science Foundation's Harnessing the Data Revolution (HDR) Big Idea activity.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.

数据，计量经济学，算法和学习研究所（理想）是多学科（计算机科学，统计，经济学，电气工程和运营研究）和多机构（西北大学，芝加哥的丰田技术学院和芝加哥大学的丰田技术学院，以及芝加哥大学）的合作，重点介绍数据科学的关键方面。 该研究所将支持与机器学习，高维数据分析以及战略和非战略环境中优化相关的基本问题的研究。 该研究所的主要活动将是主题集中的宿舍，将与研讨会和外部访问者协调研究生课程。 该研究所将通过多个倡议以及多个学科促进芝加哥地区机构之间的合作。研究议程的几个组成部分具有直接的应用领域，PIS将使从业者参与发展经济学，在线市场，公共政策以及数据科学家。 研究所支持的研究领域重点介绍了三个广泛的主题：（1）高维数据分析，以应对处理高维数据时的算法和统计挑战，并研究诸如公制嵌入，草图和无培养学习的问题之类的主题； （2）战略环境中的数据科学，以解决战略行为的计算和信息理论挑战，例如对高维结构参数空间的推断，处理未观察到的异质性，部分识别和计量经济学的机器学习； （3）机器的学习和优化，以连续和离散优化的基础问题及其在机器学习中的使用，包括代表性学习，鲁棒性和可证明的范围，以进行非convex优化。 最初，将选出六个研究主题，以将整个机构的利益联系起来：网络上的推论和数据科学；深度学习理论；共享数据基础架构的激励措施；高维统计数据的鲁棒性；高维数据分析；和部分识别模型的算法。 该研究所将有特殊的季度（秋季和春季），将调查人员，博士后和博士学位汇集在一起​​。学生专注于其中一个主题。 在接下来的季度（冬季和夏季）中，团队将继续研究提出提案主题。该项目是国家科学基金会利用数据革命（HDR）大创意活动的一部分。该奖项反映了NSF的法定任务，并被认为是值得通过基金会的知识分子优点和更广泛影响的评估审查标准来通过评估来支持的。

项目成果

Partial recovery for top-k ranking: Optimality of MLE and SubOptimality of the spectral method

top-k 排序的部分恢复：MLE 的最优性和谱方法的次最优性

• DOI: 10.1214/21-aos2166
• 发表时间: 2022
• 期刊: The Annals of Statistics
• 作者: Chen, Pinhan;Gao, Chao;Zhang, Anderson Y.
• 通讯作者: Zhang, Anderson Y.

Controlling Epidemic Spread: Reducing Economic Losses with Targeted Closures

• DOI: 10.1287/mnsc.2022.4318
• 发表时间: 2022-05-01
• 期刊: MANAGEMENT SCIENCE
• 影响因子: 5.4
• 作者: Birge, John R.;Candogan, Ozan;Feng, Yiding
• 通讯作者: Feng, Yiding

Dynamic Regret Minimization for Control of Non-stationary Linear Dynamical Systems

非平稳线性动力系统控制的动态遗憾最小化

• DOI: 10.1145/3508029
• 发表时间: 2022
• 期刊: Proceedings of the ACM on Measurement and Analysis of Computing Systems
• 作者: Luo, Yuwei;Gupta, Varun;Kolar, Mladen
• 通讯作者: Kolar, Mladen

Minimax rates for sparse signal detection under correlation

相关下稀疏信号检测的极小极大率

• DOI: 10.1093/imaiai/iaad044
• 发表时间: 2023
• 期刊: Information and Inference: A Journal of the IMA
• 作者: Kotekal, Subhodh;Gao, Chao
• 通讯作者: Gao, Chao

Optimal full ranking from pairwise comparisons

成对比较的最佳完整排名

• DOI: 10.1214/22-aos2175
• 发表时间: 2022
• 期刊: The Annals of Statistics
• 作者: Chen, Pinhan;Gao, Chao;Zhang, Anderson Y.
• 通讯作者: Zhang, Anderson Y.