Designing Neural Nets Faster

Automating the design of nets has been a long time goal. 

MIT researchers have developed an efficient algorithm that could provide a “push-button” solution for automatically designing fast-running neural networks on specific hardware.

Kicking neural network design automation into high gear Algorithm designs optimized machine-learning models up to 200 times faster than traditional methods.    By Rob Matheson | MIT News Office

A new area in artificial intelligence involves using algorithms to automatically design machine-learning systems known as neural networks, which are more accurate and efficient than those developed by human engineers. But this so-called neural architecture search (NAS) technique is computationally expensive....."