Seminar by Dr. Reza Hosseini

Dear all,

Neural Computation Unit (Doya Unit) would like to invite you to the following seminar.

------------------------------------------

Date: Wednesday, November 13

Time: 11:00-12:00

Venue: D015, Lab1

------------------------------------------

Speaker:

Reza Hosseini, PhD Statistics

JSPS foreign researcher, Dep. of Mathematical Informatics, University of Tokyo.

Abstract of the talk:

In this talk we develop a framework to approximate functions when the data are sparse but slow-moving patterns allow for a useful fit. This is done by investigating the properties of Lipschitz functions and extending the concept to wiggly functions by allowing  a deviation. Deterministic bounds are found for the approximation error of various methods in terms of the Lipschitz constant and the deviation. Moreover we present an  optimal method which outperforms other methods such as nearest neighbor and linear interpolation. The developed methods can also use the extra assumption of periodicity to obtain better prediction error bounds. This work also finds optimal sampling times which minimize the approximation error.

=======================================================

We hope to see many of you at the seminar.

Sincerely,

Emiko Asato

Research Administrator

Doya Unit