Computing Informatics Research Seminar Series

Approximation of high-dimensional functions with neural networks

open book on table
Seminars

As part of the School of Science and Technology's Computing Informatics Research Seminar Series, Professor Peter Andras, Keele University presents: Approximation of high-dimensional functions with neural networks.

  • From: Wednesday 12 December 2018, 1 pm
  • To: Wednesday 12 December 2018, 2 pm
  • Location: 161, Erasmus Darwin, Nottingham Trent University, Clifton Campus, Clifton Lane, Nottingham, NG11 8NS

Event details

As part of the School of Science and Technology's Computing Informatics Research Seminar Series, Professor Peter Andras, Keele University presents: Approximation of high-dimensional functions with neural networks.

Abstract

High-dimensional function approximation is an increasingly common problem in the era of big data, e.g. multi-sensor data from engines, proteomic / genomic data, high resolution neuroscience data. Neural networks are useful tools for this purpose, however the volume of data required for robust training grows rapidly with the dimensionality of the data. Here we will discuss about approaches that combine mapping of high dimensional data into a lower dimensional space and neural network learning of the approximated function in the lower dimensional space. We will explore the relevant error bounds and features of the high dimension - low dimension mappings that facilitate faithful approximation of the original function.

All Welcome

For any enquires please contact  Mufti Mahmud and Omprakash Kaiwartya

Location details

Room/Building:

161, Erasmus Darwin

Address:

Nottingham Trent University
Clifton Campus
Clifton Lane
Nottingham
NG11 8NS

Still need help?

+44 (0)115 941 8418