CVLGMay 11, 2024

GRASP-GCN: Graph-Shape Prioritization for Neural Architecture Search under Distribution Shifts

arXiv:2405.06994v11 citationsh-index: 22024 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)
Originality Incremental advance
AI Analysis

This work addresses the computational inefficiency and poor generalization of NAS methods when applied to multiple datasets, offering a domain-specific solution for machine learning practitioners dealing with distribution shifts.

The paper tackles the problem of neural architecture search (NAS) under data distribution shifts by proposing GRASP-GCN, a ranking graph convolutional network that incorporates layer shape information and is trained with not-at-convergence accuracies, resulting in a 3.3% improvement on Cifar-10 and enhanced generalization abilities.

Neural Architecture Search (NAS) methods have shown to output networks that largely outperform human-designed networks. However, conventional NAS methods have mostly tackled the single dataset scenario, incuring in a large computational cost as the procedure has to be run from scratch for every new dataset. In this work, we focus on predictor-based algorithms and propose a simple and efficient way of improving their prediction performance when dealing with data distribution shifts. We exploit the Kronecker-product on the randomly wired search-space and create a small NAS benchmark composed of networks trained over four different datasets. To improve the generalization abilities, we propose GRASP-GCN, a ranking Graph Convolutional Network that takes as additional input the shape of the layers of the neural networks. GRASP-GCN is trained with the not-at-convergence accuracies, and improves the state-of-the-art of 3.3 % for Cifar-10 and increasing moreover the generalization abilities under data distribution shift.

Foundations

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

Your Notes