CVLGMar 26, 2020

Are Labels Necessary for Neural Architecture Search?

arXiv:2003.12056v280 citations
AI Analysis

This addresses the problem of reducing annotation costs in neural architecture search for computer vision researchers, though it is incremental as it builds on existing NAS methods.

The paper investigates whether high-quality neural architectures can be discovered using only images without human-annotated labels, finding that architecture rankings with and without labels are highly correlated and unsupervised searches yield competitive results.

Existing neural network architectures in computer vision -- whether designed by humans or by machines -- were typically found using both images and their associated labels. In this paper, we ask the question: can we find high-quality neural architectures using only images, but no human-annotated labels? To answer this question, we first define a new setup called Unsupervised Neural Architecture Search (UnNAS). We then conduct two sets of experiments. In sample-based experiments, we train a large number (500) of diverse architectures with either supervised or unsupervised objectives, and find that the architecture rankings produced with and without labels are highly correlated. In search-based experiments, we run a well-established NAS algorithm (DARTS) using various unsupervised objectives, and report that the architectures searched without labels can be competitive to their counterparts searched with labels. Together, these results reveal the potentially surprising finding that labels are not necessary, and the image statistics alone may be sufficient to identify good neural architectures.

Code Implementations2 repos
Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes