English

Hardware-aware training for large-scale and diverse deep learning inference workloads using in-memory computing-based accelerators

Machine Learning 2023-02-17 v1 Emerging Technologies

Abstract

Analog in-memory computing (AIMC) -- a promising approach for energy-efficient acceleration of deep learning workloads -- computes matrix-vector multiplications (MVMs) but only approximately, due to nonidealities that often are non-deterministic or nonlinear. This can adversely impact the achievable deep neural network (DNN) inference accuracy as compared to a conventional floating point (FP) implementation. While retraining has previously been suggested to improve robustness, prior work has explored only a few DNN topologies, using disparate and overly simplified AIMC hardware models. Here, we use hardware-aware (HWA) training to systematically examine the accuracy of AIMC for multiple common artificial intelligence (AI) workloads across multiple DNN topologies, and investigate sensitivity and robustness to a broad set of nonidealities. By introducing a new and highly realistic AIMC crossbar-model, we improve significantly on earlier retraining approaches. We show that many large-scale DNNs of various topologies, including convolutional neural networks (CNNs), recurrent neural networks (RNNs), and transformers, can in fact be successfully retrained to show iso-accuracy on AIMC. Our results further suggest that AIMC nonidealities that add noise to the inputs or outputs, not the weights, have the largest impact on DNN accuracy, and that RNNs are particularly robust to all nonidealities.

Keywords

Cite

@article{arxiv.2302.08469,
  title  = {Hardware-aware training for large-scale and diverse deep learning inference workloads using in-memory computing-based accelerators},
  author = {Malte J. Rasch and Charles Mackin and Manuel Le Gallo and An Chen and Andrea Fasoli and Frederic Odermatt and Ning Li and S. R. Nandakumar and Pritish Narayanan and Hsinyu Tsai and Geoffrey W. Burr and Abu Sebastian and Vijay Narayanan},
  journal= {arXiv preprint arXiv:2302.08469},
  year   = {2023}
}

Comments

35 pages, 7 figures, 5 tables

R2 v1 2026-06-28T08:42:07.565Z