Interpretable Neural Architecture Search via Bayesian Optimisation with Weisfeiler-Lehman Kernels
This addresses the need for interpretable NAS to guide architecture improvements, offering a data-efficient method that is scalable and provides insights, though it builds incrementally on existing BO and kernel techniques.
The paper tackles the problem of neural architecture search (NAS) lacking interpretability by proposing a Bayesian optimisation approach with Weisfeiler-Lehman kernels, which achieves state-of-the-art performance on closed- and open-domain search spaces while providing insights into useful network features.
Current neural architecture search (NAS) strategies focus only on finding a single, good, architecture. They offer little insight into why a specific network is performing well, or how we should modify the architecture if we want further improvements. We propose a Bayesian optimisation (BO) approach for NAS that combines the Weisfeiler-Lehman graph kernel with a Gaussian process surrogate. Our method optimises the architecture in a highly data-efficient manner: it is capable of capturing the topological structures of the architectures and is scalable to large graphs, thus making the high-dimensional and graph-like search spaces amenable to BO. More importantly, our method affords interpretability by discovering useful network features and their corresponding impact on the network performance. Indeed, we demonstrate empirically that our surrogate model is capable of identifying useful motifs which can guide the generation of new architectures. We finally show that our method outperforms existing NAS approaches to achieve the state of the art on both closed- and open-domain search spaces.