The brain has long attracted the interest of researchers. Some tasks, such as pattern
recognition and optimization, have proven to be exceptionally difficult for conventional
computing systems to perform, but are executed by the brain almost effortlessly. Due to
the large number of neurons and interconnections, it has proven impossible...
The objective of this thesis is to present the architecture and
design of a neural network-based pattern classifier. The classifier
detects textual characters which have been translated, rotated, and
corrupted by noise. This form of pattern classifier differs
significantly from traditional pattern classifiers. The neural network
architecture used in implementing...
This note briefly discusses some of the classical results of McCulloch and Pitts. It then deals with some current research in neural nets. Several questions about neural nets are shown to be computationally difficult by showing that they are NP-Complete or worse. The size of neural nets necessary to compute...
We took the back-propagation algorithms of Werbos for recurrent and feed-forward neural networks and implemented them on machines with graphics processing units (GPU). The parallelism of these units gave our implementations a 10 to 100 fold increase in speed. For nets with less than 20 neurons the machine performed faster...
Analog computation in the form of neural networks is currently
receiving much attention. Existing algorithms cannot be easily
implemented in hardware because of the large number of neurons needed
and the number of connections necessary between them. These problems
have motivated development of alternatives to a conventional
implementation. Hence, a...
Generalized Radial Basis Functions were used to construct networks
that learn input-output mappings from given data. They are
developed out of a theoretical framework for approximation based
on regularization techniques and represent a class of three-layer
networks similar to backpropagation networks with one hidden
layer.
A network using Gaussian base...
Machine learning models for natural language processing have traditionally relied on large numbers of discrete features, built up from atomic categories such as word forms and part-of-speech labels, which are considered completely distinct from each other. Recently however, the advent of dense feature representations coupled with deep learning techniques has...
Deep learning has greatly improved visual recognition in recent years. However, recent research has shown that there exist many adversarial examples that can negatively impact the performance of such an architecture. Different from previous perspectives that focus on improving the classifiers to detect the adversarial examples, this work focuses on...
This thesis focuses on the problem of object tracking. Given a video, the general objective of tracking is to track the location over time of one or more targets in the image sequence. This is a very challenging task as algorithms need to deal with problems such as appearance variations,...
Sustainable product design is becoming an important component of the development of consumer products. Currently there are limited design resources to aid in the creation of environmentally sustainable products. The purpose of this research is to theorize a new method for integrating sustainable design knowledge into the early design phase...