Learning poisson binomial distributions
Name
Daskalakis-Learning Poisson.pdf
Size
330.79 KB
Format
Adobe PDF
Checksum (MD5)
da056e6c2330107ca6177cf7dff434bc
Author(s)
Daskalakis, Constantinos
Diakonikolas, Ilias
Servedio, Rocco A.
Date Issued
May 2012
Journal
Proceedings of the 44th symposium on Theory of Computing (STOC '12)
Publisher
Association for Computing Machinery (ACM)
Citation
Constantinos Daskalakis, Ilias Diakonikolas, and Rocco A. Servedio. 2012. Learning poisson binomial distributions. In Proceedings of the 44th symposium on Theory of Computing (STOC '12). ACM, New York, NY, USA, 709-728.
Version
Author's final manuscript
Abstract
We consider a basic problem in unsupervised learning: learning an unknown Poisson Binomial Distribution. A Poisson Binomial Distribution (PBD) over {0,1,...,n} is the distribution of a sum of n independent Bernoulli random variables which may have arbitrary, potentially non-equal, expectations. These distributions were first studied by S. Poisson in 1837 and are a natural n-parameter generalization of the familiar Binomial Distribution. Surprisingly, prior to our work this basic learning problem was poorly understood, and known results for it were far from optimal.
We essentially settle the complexity of the learning problem for this basic class of distributions. As our main result we give a highly efficient algorithm which learns to ε-accuracy using O(1/ε[superscript 3]) samples independent of n. The running time of the algorithm is quasilinear in the size of its input data, i.e. ~O(log(n)/ε[superscript 3) bit-operations (observe that each draw from the distribution is a log(n)-bit string). This is nearly optimal since any algorithm must use Ω(1/ε[superscript 2) samples. We also give positive and negative results for some extensions of this learning problem.
MIT Department
Massachusetts Institute of Technology. Computer Science and Artificial Intelligence Laboratory
Terms of Use
Creative Commons Attribution-Noncommercial-Share Alike 3.0
Persistent DSpace Link
DOI of Published Version
https://doi.org/10.1145/2213977.2214042
