Neuromorphic Simplified

Intel has been working on morphic chips and announced its offering Loihi just last month. Qualcomm is working on the Neural Processing Unit Zeroth, IBM is making broader investments in the neural intelligence tech space. IBM is developing TrueNorth; the Technion is working on one with Memristor-based on phase change memory. General Vision has NeuroMem. When you build these chips, they get managed by the processor which is the core of the model.
The military is using neuromorphic hardware for drone footage. IBM with TrueNorth came from DARPA the project derived from the military’s broader Synapse brain-chip project. The market is over 2 Billion Dollars. With autonomous cars, and medical will grow to 20 billion by 2020.
Most of the is a neural process performed is without utilizing programs. The processor uses an object-based language which abstracts the synapses and neurons. The model developed is defined by the neurons in the model with the object language. That way it is based on context. The CPU uses spiking neurons, and the results can be Statistically Deterministic to create the model. Each neuron is in a grid having the only event based on spikes to create events for analysis.
What will happen is that when you drive an autonomous car, it will learn a context based on multimodal inputs such as ice on the road or fog using vision and controls from the car. The behavior in the model will be entered based on the random spikes by exposing the vehicle to the environment. When you work on genomics, the model gets based on neural networks and the artificial intelligence which has expert knowledge understanding the events deployed in its ecosystem