Optimisation of Massively Parallel Neural Networks Book

Most current artificial neural networks exist only within softwaresimulators running on conventional computers. Simulators can provide great flexibility, but require immensely powerful and costly hardware for even very small networks.

An artificial neural network implemented as a custom integrated circuit could operate many thousands of times faster than any simulator as each neuron can operate simultaneously. A significant problem with implementing neural networks in hardware is that larger networks require a great deal of silicon area, making them too costly to design and produce.

In this book, I test the effectiveness of a number of algorithms that reduce the size of a trained neural network while maintaining accuracy.

Author Bio: Michael Oldroyd is a software development veteran who started progamming professionally back in 1992. He is now development manager at AES Data Systems.

He has worked as a consultant and software developer for a number of international organisations including Mobil Oil, The European Commission, Deutsche Bank, Compaq Computer, and the Cabinet Office. He has developed several bespoke AI trading and decision support tools used on trading floors in the currency, stock and energy markets. He is a professional member of the IEEE and the Computational Intelligence Society.