Researchers from Alibaba Group and Peking University have open-sourced Kernel Neural Architecture Search (KNAS), an efficient automated machine learning (AutoML) algorithm that can evaluate proposed architectures without training. KNAS uses a gradient kernel as a proxy for model quality, and uses an order of magnitude less compute power than baseline methods, told InfoQ.

KNAS is very simple and only requires gradient vectors to get MGM scores, was told in a Github’s publication.

The algorithm and a set of experiments were described in a paper published in Proceedings of Machine Learning Research. The purpose of KNAS is to predict which model architectures will perform well without actually training them and potentially to save many hours of compute time, added InfoQ.