Topic Description

The field of robust statistics addresses the challenge of devising estimators that demonstrate reliable performance in situations where the data deviates substantially from the assumed idealized models. Over the past decade, a line of work in computer science has led to significant advances in the algorithmic aspects of this field. This workshop will concentrate on exploring the upcoming research directions in algorithmic robust statistics, emphasizing connections with different privacy and associated concepts of algorithmic stability.

List of Participants

Ainesh Bakshi (MIT)
Gavin Brown (U. Washington)
Sitan Chen (Harvard)
Yu Cheng (Brown)
Ilias Diakonikolas (UW Madison)
Chao Gao (U. Chicago)
Rong Ge (Duke)
Sam Hopkins (MIT)
Ioannis Iakovidis (UW Madison)
He Jia (Northwestern)
Gautam Kamath (Waterloo)
Daniel Kane (UCSD)
Sushrut Karmalkar (UW Madison)
Jasper Lee (UW Madison)
Jerry Li (MSR)
Allen Liu (MIT)
Sihan Liu (UCSD)
Mingchen Ma (UW Madison)
Mahbod Majid (CMU)
Stas Minsker (USC)
Argyris Mouzakis (Waterloo)
Shuo Pang (University of Copenhagen)
Ankit Pensia (IBM)
Thanasis Pittas (UW Madison)
Aaron Potechin (U. Chicago)
Lisheng Ren (UW Madison)
Mahdi Soltanokotabi (USC)
Adam Smith (BU)
Yuxin Sun (UW Madison)
Kevin Tian (UT Austin)
Stefan Tiegel (ETH)
Santosh Vempala (GaTech)
Aravindan Vijayraghavan (Northwestern)
Lydia Zakynthinou (Berkeley)
Nikos Zarifis (UW Madison)

Schedule

Wednesday (June 12)

Time	Talk
9:00-9:20	Continental Breakfast
9:20-9:30	Address by Avrim Blum
9:30-10:15	He Jia (Northwestern/IDEAL) Robustly Learning Mixtures of Arbitrary Gaussians [pdf]
10:15-11:00	Allen Liu (MIT) Clustering Mixtures with Almost Optimal Separation in Polynomial Time via Implicit Moment Estimation [pdf]
11:00-11:15	Coffee Break
11:15- 12:00	Jasper Lee (UW Madison) Clustering Mixtures of Bounded Covariance Distributions Under Optimal Separation [pdf]
12:00-12:45	Ainesh Bakshi (MIT) Clustering Mixture Models: Beyond Gaussians [pdf]
1:00-2:00	Lunch at TTIC
2:00-3:00	Yu Cheng (Brown) Bridging High-Dimensional Robust Statistics and Non-Convex Optimization [pdf]
3:00-4:00	Kevin Tian (UT Austin) Matrix Multiplicative Weights and Nearly-Linear Time Robust Statistics [pdf]
4:00-4:15	Coffee Break
4:15 - 5:15	Ankit Pensia (IBM) Robust Sparse Estimation [pdf]

Thursday (June 13)

Time	Talk
8:45-9:15	Continental Breakfast
9:15-10:15	Sam Hopkins (MIT) Privacy to Robustness and Back Again [pdf]
10:15-11:00	Lydia Zakynthinou (Berkeley) A Black-Box Transformation from Robust to Private Estimators [pdf]
11:00-11:15	Coffee Break
11:15 - 12:00	Mahbod Majid (CMU) Private and Efficient Algorithm Design from Robust Estimators [pdf]
12:00-12:45	Gavin Brown (UW) Stable Estimators for Fast Private Statistics [pdf]
1:00-2:00	Lunch at TTIC
2:00-2:30	Chao Gao (U. Chicago) Is Adaptive Robust Confidence Interval Possible? [pdf]
2:30-3:00	Sushrut Karmalkar (UW-Madison) Regression in the Presence of Additive Oblivious Corruptions [pdf]
3:00-3:20	Coffee Break
3:20-3:50	Jerry Li (MSR) Avoiding Generative Overfitting in Algorithm Design via Semi-random Models
3:50-4:20	Stas Minsker (USC) The median of means estimator: old and new [pdf]
4:30-5:30	Open Problems Session

Friday (June 14)

Time	Talk
9:00-9:30	Continental Breakfast
9:30-10:20	Ilias Diakonikolas (UW Madison) Information-Computation Tradeoffs via Non-Gaussian Component Analysis [pdf]
10:30-11:30	Aaron Potechin (U Chicago) Sum of Squares Lower Bounds Versus Low-Degree Polynomial Lower Bounds [pdf]
11:30-11:45	Coffee Break
11:45-12:45	Shuo Pang (U. Copenhagen) Sum-of-Squares Lower Bounds for Non-Gaussian Component Analysis: Challenges and New Techniques [pdf]
1:00-2:00	Lunch at TTIC
2:00-3:00	Spotlight Talks
3:30-4:30	Working Groups