The concept of neuromorphic computing has been around for decades, yet has only recently seen success in realizing the machine’s capabilities to match and even surpass that of the human brain. Neuromorphic computing is a method of processing data that allows machines to imitate the learning process of the brain, by using a combination of analog resistors, nonlinear activation functions, and an array of machine learning algorithms such as back-propagation. Scientists are continually developing various models of spiking neural networks that replicate the thoughts and learning abilities of the human brain.
This presentation by Linnainma (1970) and first implemented by Werbos (1974), looks into the potential of frame-based neuromorphic computing using an array of analog resistors. It will further discuss the tight interplay between materials, algorithms, architecture, and application, in order to successfully materialize this technology. Due to the advancement in computing hardware and the availability of large datasets, machine learning is making successful leaps ahead of traditional models like McCulloch & Pitts’s (1943) simple mathematical neuron model.
The success of Krizhevsky, I. Sutskever and G. Hinton (2012) and others has further highlighted the actual power of machine learning and how the combination of powerful algorithms and advanced computation hardware can push scientific boundaries. Numerous applications of machine learning are now making life better for business owners, with potential implications to every facet of the industry.
Regarding the company, Krizhevsky, I. Sutskever and G. Hinton (2012) are one of the companies leading the charge into machine learning and neuromorphic computing with their advances in powerful algorithms, backed up by some of the top entertainers in the industry. Their emphasis on research and development has opened paths for the advancement of these technologies.
The individual mentioned in this article is Linnainma (1970), who first implemented the algorithm of back-propagation which still remains the key ingredient of many machine learning algorithms today. Linnainma’s technological advancements has enabled machines to learn quicker and more efficiently, setting the standard of the performance in neuromorphic computing. It was in his works that machine learning set off on its global take-off, with the major propagation in large datasets and advances in computing hardware.
