nyvind

Hardware meets AI!

NYV is a neuromorphic computing architecture for neural networks. We are building hardware that emulates neurons for building brain like circuits.

Our approach

The memristor

It was in 1971 that Leon Chua reported his prediction of a device that behaves like a nonlinear resistor with memory, which he then termed a memristor (a contraction of memory and resistor)2. Thirty-seven years later, the team from Hewlett Packard Labs, which was led by R. Stanley Williams, connected their experimental observations to Chua’s theoretical prediction and the fourth fundamental circuit element was found.

“Over the last five decades we have processed digital signals and have computed using digital techniques, which has taken us very far,” said Southampton researcher Dr Alexantrou Serb. “However, if we are to truly compute at the limits of energy efficiency, that the laws of physics allow, it would seem imperative that we need to move towards analogue computation techniques whilst being much savvier about how to mix analogue and digital signals for maximum effect.”

We are experimenting!

Memristors are non-volatile electrically-adjustable resistors. Made from metal oxides, they scale well and are proposed as memory elements in nanometre-scale logic.

“Over the last five decades we have processed digital signals and have computed using digital techniques, which has taken us very far,” said Southampton researcher Dr Alexantrou Serb. “However, if we are to truly compute at the limits of energy efficiency, that the laws of physics allow, it would seem imperative that we need to move towards analogue computation techniques whilst being much savvier about how to mix analogue and digital signals for maximum effect.”

Behaviour while executing Neural Networks

Earlier this year, the University demonstrated 128 distinct analogue levels stored on a single memristor – equivalent to seven data bits.

“A direct impact of this research on modern technologies could be the creation of ultra-efficient AI hardware,” said the University. “AI by nature lends itself to analogue implementation of computation much more readily than to the current digital-based techniques.”

Our technology

We designed a PCB to use our 8 neuron chip, here you can find everything you need to build your own board and start experimenting! We go from the technical design to building the PCB and sending you the complete kit with our chip.

Schematic

PCB design

Architecture of neurons

I-V characterization

Repository