2000 character limit reached
Lower Bounds for Sparse Oblivious Subspace Embeddings
Published 21 Dec 2021 in cs.DS, cs.CG, and cs.DM | (2112.10987v1)
Abstract: An oblivious subspace embedding (OSE), characterized by parameters , is a random matrix such that for any -dimensional subspace , . For and at most a small constant, we show that any OSE with one nonzero entry in each column must satisfy that , establishing the optimality of the classical Count-Sketch matrix. When an OSE has nonzero entries in each column, we show it must hold that , improving on the previous lower bound due to Nelson and Nguyen (ICALP 2014).
Paper Prompts
Sign up for free to create and run prompts on this paper.