7 papers
Moderate Dimension Reduction for -Center Clustering
Shaofeng H. -C. Jiang, Robert Krauthgamer, Shay Sapir
The Johnson-Lindenstrauss (JL) Lemma introduced the concept of dimension reduction via a random linear map, which has become a fundamental technique in many computational settings.…
Is Randomness Necessary for Adaptive Data Analysis?
Edith Cohen, Haim Kaplan, Yishay Mansour +2
The Adaptive Data Analysis (ADA) problem formalizes the challenge of preventing false discovery and overfitting when a dataset is repeatedly reused. Formally, our input is a datase…
Load Balancing under Adaptive Bin Deletions
Haim Kaplan, Shay Sapir, Uri Stemmer
We analyze a balls-and-bins game against an adaptive adversary that sequentially deletes bins. Starting with balls distributed across bins, the adversary deletes a bin in e…
On the Adversarial Robustness of Online Importance Sampling
Yotam Kenneth-Mordoch, Shay Sapir
This paper studies the adversarial-robustness of importance-sampling (aka sensitivity sampling); a useful algorithmic technique that samples elements with probabilities proportiona…
Dimension Reduction for Clustering: The Curious Case of Discrete Centers
Shaofeng H. -C. Jiang, Robert Krauthgamer, Shay Sapir +2
The Johnson-Lindenstrauss transform is a fundamental method for dimension reduction in Euclidean spaces, that can map any dataset of points into dimension with low…
Randomized Dimensionality Reduction for Euclidean Maximization and Diversity Measures
Jie Gao, Rajesh Jayaram, Benedikt Kolbe +4
Randomized dimensionality reduction is a widely-used algorithmic technique for speeding up large-scale Euclidean optimization problems. In this paper, we study dimension reduction…