Loading...

Optimal lower bounds for universal relation, and for samplers and finding duplicates in streams

Kapralov, M ; Sharif University of Technology

263 Viewed
  1. Type of Document: Article
  2. DOI: 10.1109/FOCS.2017.50
  3. Abstract:
  4. In the communication problem UR (universal relation), Alice and Bob respectively receive x, y {0,1}^n with the promise that xy. The last player to receive a message must output an index i such that x-iy-i. We prove that the randomized one-way communication complexity of this problem in the public coin model is exactly Θ(min{n,log(1/Δ)log^2(rac n{log(1/Δ)})}) for failure probability Δ. Our lower bound holds even if promised mathop{support}(y)mathop{support}(x). As a corollary, we obtain optimal lower bounds for -p-sampling in strict turnstile streams for 0le p streams for 0 ≤ p © 2017 IEEE
  5. Keywords:
  6. Lower bounds ; Streaming ; P-sampling
  7. Source: Annual Symposium on Foundations of Computer Science - Proceedings, 15 October 2017 through 17 October 2017 ; Volume 2017-October , 2017 , Pages 475-486 ; 02725428 (ISSN) ; 9781538634646 (ISBN)
  8. URL: https://ieeexplore.ieee.org/document/8104082