Optimizing for In-Memory Deep Learning With Emerging Memory Technology

Zhehui Wang

doi:10.1109/TNNLS.2023.3285488

Optimizing for In-Memory Deep Learning With Emerging Memory Technology

2024-11-01

1

OA

AI

DOI：10.1109/TNNLS.2023.3285488

Original

Original request for help

Save

Abstract

En 中文

In-memory deep learning executes neural network models where they are stored, thus avoiding long-distance communication between memory and computation units, resulting in considerable savings in energy and time. In-memory deep learning has already demonstrated orders of magnitude higher performance density and energy efficiency. The use of emerging memory technology (EMT) promises to increase density, energy, and performance even further. However, EMT is intrinsically unstable, resulting in random data read fluctuations. This can translate to nonnegligible accuracy loss, potentially nullifying the gains. In this article, we propose three optimization techniques that can mathematically overcome the instability problem of EMT. They can improve the accuracy of the in-memory deep learning model while maximizing its energy efficiency. Experiments show that our solution can fully recover most models' state-of-the-art (SOTA) accuracy and achieves at least an order of magnitude higher energy efficiency than the SOTA.

Keywords:

Fluctuations

Deep learning

Computational modeling

Optimization

Energy efficiency

In-memory computing

Phase change random access memory

Deep learning

emerging memory technology (EMT)

in-memory computing

optimization

Optimizing for In-Memory Deep Learning With Emerging Memory Technology

Abstract

Journal

Organization