Continuing Education Course / IEEE Baltimore Section

Title: Feedforward Neural Networks: Theory & Applications

Speaker: Frederick W. Chen, Ph.D.

Date: Feb 2, 2013 Time: 10 am - 2 pm Click for pictures from this talk

Location:

National Electronics Museum (NEM)

1745 W. Nursery Road, Linthicum, MD 21090

https://www.nationalelectronicsmuseum.org

Abstract: This tutorial will provide an introduction to neural networks and describe an example of their application. Neural networks are a computational intelligence method inspired by biological neural networks. They were developed to learn variations in data in a way similar to how biological organisms learn from their environments. The ability of neural networks to learn nonlinear variations in data without any assumptions can be very useful in problems where variations can be extremely difficult to express in closed form. In this tutorial, applications to satellite-based estimation of atmospheric temperature and water vapor profile and precipitation rate will be described.

This first part of the tutorial will describe how to use neural networks and explain the mathematics behind them. Important issues include data for training and evaluation and topology selection. Common mistakes will also be covered. While neural networks can be powerful, they are also frequently misused due to a widespread lack of understanding of their capabilities. Another important issue is preprocessing and postprocessing. Measurements (e.g. from satellites) are rarely in a form that neural networks can learn with, and frequently preprocessing is needed to make the measurements and auxiliary data most relevant to a particular problem. Specific examples of preprocessing will include methods based on principal component analysis (PCA) and methods that consider the topology of relevant variables. Preprocessing can also be used to reduce the dimensionality of the data set which can improve computational efficiency in both training and simulation. Postprocessing can also be useful in some situations, e.g. one in which the output varies over a wide range of scales.

An application in earth science remote sensing will be presented. Neural networks have been used to develop computationally efficient atmospheric retrieval algorithms. Blackwell (IEEE TGRS, 11/2005) has developed an algorithm for estimating temperature and water vapor profile that uses projection principal components to preprocess hyspectral infrared data from the Atmospheric Infrared Sounder (AIRS), and Chen and Staelin (IEEE TGRS, 2/2003) have developed a method for estimating precipitation rate using data from the Advanced Microwave Sounding Unit (AMSU). Applications in other areas are also possible.

1. Introduction

1. Inspiration from biological systems

2. Brief descriptions of NN (including self-organizing maps, feedback NN, ...)

3. In-depth description of feedforward NN

4. Training NN (training algorithms; training, validation, and testing sets)

5. Choosing appropriate topology

6. Simple examples to provide intuition

2. Preprocessing & postprocessing

1. PCA & variations (NAPC, PPC, constrained PCA)

2. Topological preprocessing (circular data)

3. Preprocessing based on a priori knowledge of system

4. Postprocessing

3. Application example: atmospheric remote sensing

1. AMSU precipitation

2. SCC/NN

Speaker’s bio: Frederick W. Chen is a Senior Engineer at the Signal Systems Corporation (SSC) in Severna Park, MD. He received the S.B., M.Eng., and Ph.D. degrees in electrical engineering from the Massachusetts Institute of Technology (MIT) in 1998, 1998, and 2004, respectively. From 2004 to 2007, he was a technical staff member at the MIT Lincoln Laboratory (Lexington, MA) where he worked on signal and image processing problems in satellite-based atmospheric remote sensing of temperature profile and precipitation using microwave and infrared data in support of the National Polar-Orbiting Operational Environmental Satellite System (NPOESS) program. Since 2009, he has been involved in acoustic signal processing work at SSC. His technical interests also include signal separation, transform coding, data compression, neural networks, and other areas in signal processing, computational intelligence, and information theory. He has co-authored a book on neural networks (Neural Networks in Atmospheric Remote Sensing, Artech House, 2009). He is a Senior Member of the IEEE.

Date and place: Feb 2, 2013, 10 am – 2 pm at the National Electronics Museum (former Historic Electronics Museum): 1745 W. Nursery Rd., Linthicum, MD

Registration: Send email to Boris Gramatikov (bgramat-at-jhmi.edu) indicating your IEEE status and IEEE member #, work affiliation, and whether you intend to apply for CEU credit. Those who would like to receive the credit and a certificate, should bring to the course a check for $18, issued to “IEEE, Baltimore Section”.