LGAICVNEMar 4, 2024

Encodings for Prediction-based Neural Architecture Search

arXiv:2403.02484v19 citationsh-index: 9Has CodeICML
AI Analysis

This work addresses the efficiency of NAS for machine learning practitioners by providing a systematic analysis and a cost-effective predictor, though it appears incremental as it builds on existing encoding methods.

The paper tackles the problem of improving predictor-based Neural Architecture Search (NAS) by investigating and categorizing neural encodings, and introduces FLAN, a predictor that reduces training costs by more than an order of magnitude through unified encodings and transfer learning.

Predictor-based methods have substantially enhanced Neural Architecture Search (NAS) optimization. The efficacy of these predictors is largely influenced by the method of encoding neural network architectures. While traditional encodings used an adjacency matrix describing the graph structure of a neural network, novel encodings embrace a variety of approaches from unsupervised pretraining of latent representations to vectors of zero-cost proxies. In this paper, we categorize and investigate neural encodings from three main types: structural, learned, and score-based. Furthermore, we extend these encodings and introduce \textit{unified encodings}, that extend NAS predictors to multiple search spaces. Our analysis draws from experiments conducted on over 1.5 million neural network architectures on NAS spaces such as NASBench-101 (NB101), NB201, NB301, Network Design Spaces (NDS), and TransNASBench-101. Building on our study, we present our predictor \textbf{FLAN}: \textbf{Fl}ow \textbf{A}ttention for \textbf{N}AS. FLAN integrates critical insights on predictor design, transfer learning, and \textit{unified encodings} to enable more than an order of magnitude cost reduction for training NAS accuracy predictors. Our implementation and encodings for all neural networks are open-sourced at \href{https://github.com/abdelfattah-lab/flan_nas}{https://github.com/abdelfattah-lab/flan\_nas}.

Code Implementations1 repo
Foundations

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

Your Notes