FURI | Summer 2024, Needs Review

Automated Neural Architecture Search for Resource-Constrained Environments

This research aims to develop an efficient Neural Architecture Search (NAS) method for resource-constrained environments like embedded systems and mobile devices. The study focuses on overcoming limitations in memory, computational power, latency, bandwidth, energy consumption, and hardware compatibility. By creating a streamlined NAS approach tailored for these environments, it enhances the deployment of advanced AI models, making AI more accessible and practical for various applications. The study builds on existing NAS techniques and evaluates the proposed approach using benchmarks like CIFAR-10 and CIFAR-100. Future work will explore further optimizations and broader applicability across diverse resource-constrained scenarios.