Cogniscience
Cogniscience is developing the research work of Professor Steve Furber's group on hardware support for large-scale neural nets, with a view to commercial implementation.
Neural nets self adapt to learn from information. They provide powerful models for processing inexact data and supporting complex decision making. However, neural nets have traditionally been deployed in software on a small scale.
Professor Furber's group have developed new approaches to large scale deployments. The group is working towards the ultimate research goal of building a hardware system capable of supporting a model of a billion neuron system that can operate in real time.
Related information:
SpiNNaker - A Universal Spiking Neural Network Architecture.
Professor Steve Furber's page.
