Transfer-Once-For-All: AI Model Optimization for Edge

• Published in: 2023 IEEE International Conference on Edge Computing and Communications (EDGE) pp. 26 - 35
• Main Authors: Kundu, Achintya, Wynter, Laura, Lee, Rhui Dih, Angel Bathen, Luis
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.07.2023
• Subjects: Computational modeling; Computer architecture; Data models; inference at the edge; Network topology; neural architecture search; Neural networks; semi-supervised supernet training; Topology; Training; weight-sharing NAS
• ISSN: 2767-9918
• DOI: 10.1109/EDGE60047.2023.00017

Weight-sharing neural architecture search aims to optimize a configurable neural network model (supernet) for a variety of deployment scenarios across many devices with different resource constraints. Existing approaches use evolutionary search to extract models of different sizes from a supernet trained on a very large data set, and then fine-tune the extracted models on the typically small, real-world data set of interest. The computational cost of training thus grows linearly with the number of different model deployment scenarios. Hence, we propose Transfer-Once-For-All (TOFA) for supernet-style training on small data sets with constant computational training cost over any number of edge deployment scenarios. Given a task, TOFA obtains custom neural networks, both the topology and the weights, optimized for any number of edge deployment scenarios. To overcome the challenges arising from small data, TOFA utilizes a unified semi-supervised training loss to simultaneously train all subnets within the supernet, coupled with on-the-fly architecture selection at deployment time.