HILAL: Hessian-Informed Layer Allocation for Heterogeneous Analog–Digital Inference

Heterogeneous AI accelerators that combine high-precision digital cores with energy-efficient analog in-memory computing (AIMC) units offer a promising path to overcome the energy and scalability limits of deep learning. A key challenge, however, is to determine which neural network layers can be executed on noisy analog units without compromising accuracy. Existing mapping strategies rely largely on ad-hoc heuristics and lack principled noise-sensitivity estimation. We propose HILAL (Hessian-Informed Layer Allocation), a framework that systematically quantifies layer robustness to analog noise using two complementary metrics: noise-aware Expected Loss Increase and spectral concentration. Layers are partitioned into robust and sensitive groups via clustering, enabling threshold-free mapping to analog or digital units. To further mitigate accuracy loss, we gradually offload layers to AIMC while retraining with noise-injection. Experiments on convolutional networks (AlexNet, VGG-16, ResNet-8, ResNet-50) and transformers (ViT, MobileBERT) across CIFAR-10, CIFAR-100, and SQuAD show that HILAL is on average 3.09x faster in search and mapping runtime than state-of-the-art methods while achieving less accuracy degradation and maximizing analog utilization.