Surprisingly Strong Performance Prediction with Neural Graph Features

arXiv:2404.16551v1 Announce Type: new
Abstract: Performance prediction has been a key part of the neural architecture search (NAS) process, allowing to speed up NAS algorithms by avoiding resource-consuming network training. Although many performance predictors correlate well with ground truth performance, they require training data in the form of trained networks. Recently, zero-cost proxies have been proposed as an efficient method to estimate network performance without any training. However, they are still poorly understood, exhibit biases with network properties, and their …

abstract algorithms architecture arxiv cost cs.lg data features form graph key nas network networks network training neural architecture search part performance prediction process search speed training training data truth type