GrMoNAS: A granularity-based multi-objective NAS framework for efficient medical diagnosis

Xin Liu, Jie Tian, Peiyong Duan, Qian Yu, Gaige Wang, Yingjie Wang

Abstract: Highlights•The framework refines architecture from coarse to fine for efficiency.•GrMoNAS optimizes both accuracy and efficiency using multi-objective algorithms.•Through the experimental results, the framework outperforms the other methods.•This framework was validated in various medical scenarios like COVID-19.•Suits hospitals with limited resources, reducing traditional NAS method costs.