QScheduler: Adaptive Gradient Sampling for Zeroth-Order On-Device Training on INT8 NPUs

Victor Felipe Domingues Do Amaral, Pierre Demaj, Erwan Libessart +3

Zeroth-Order (ZO) optimization enables On-Device Learning (ODL) on NPU-equipped microcontrollers by estimating gradients through forward passes alone, bypassing the need for backpr…