Permanents and modeling probability distributions
August 31 to September 4, 2009
American Institute of Mathematics,
San Jose, California
and Krishnamurthy Viswanathan
This workshop will study
the problem of estimating a probability distribution from a small data sample
it generates. The workshop will investigate consolidating a
theoretical and algorithmic framework for this topic.
Aspects addressed will include distribution, probability, and
population estimation, prediction, and classification. Emphasis
will be on recent methods related to Good-Turing estimation
and patterns, which cast the problem in a combinatorial and
machine-learning perspective and relate it to integer and set
partitions, symmetric polynomials over many variables,
computation of matrix permanents, Markov chain Monte Carlo
techniques, universal compression, and geometric
Material from the workshop
A list of participants.
The workshop schedule.
A report on the workshop activities.