Disentangled Neural Architecture Search
Xinyue Zheng, Peng Wang, Qigang Wang, Zhongchao Shi

TL;DR
Disentangled Neural Architecture Search (DNAS) introduces an interpretable framework that disentangles architecture representations, enabling targeted, efficient search for high-performance neural networks under various constraints.
Contribution
DNAS is the first method to disentangle architecture representations, improving interpretability and search efficiency in neural architecture search.
Findings
Achieves state-of-the-art 94.21% on NASBench-101 with less than 1/13 computational cost
Disentangles architecture components like operations, skip connections, and layers
Flexible architecture search under different FLOPS restrictions
Abstract
Neural architecture search has shown its great potential in various areas recently. However, existing methods rely heavily on a black-box controller to search architectures, which suffers from the serious problem of lacking interpretability. In this paper, we propose disentangled neural architecture search (DNAS) which disentangles the hidden representation of the controller into semantically meaningful concepts, making the neural architecture search process interpretable. Based on systematical study, we discover the correlation between network architecture and its performance, and propose a dense-sampling strategy to conduct a targeted search in promising regions that may generate well-performing architectures. We show that: 1) DNAS successfully disentangles the architecture representations, including operation selection, skip connections, and number of layers. 2) Benefiting from…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsAdversarial Robustness in Machine Learning · Advanced Neural Network Applications · Explainable Artificial Intelligence (XAI)
MethodsGumbel Softmax · Differentiable Neural Architecture Search
