Dynamic neural network with memristive CIM and CAM for 2D and 3D vision

TL;DR

This paper introduces a dynamic neural network leveraging memristor-based CIM and CAM circuits for efficient 2D and 3D vision tasks, mimicking brain-like associative memory with significant energy and computation savings.

Contribution

It presents a hardware-software co-design of a semantic memory-based DNN using memristors, enabling associative learning and efficient processing for vision applications.

Findings

Achieves accuracy comparable to software models on MNIST and ModelNet datasets.

Reduces computational budget by up to 48.1%.

Reduces energy consumption by up to 93.3%.

Abstract

The brain is dynamic, associative and efficient. It reconfigures by associating the inputs with past experiences, with fused memory and processing. In contrast, AI models are static, unable to associate inputs with past experiences, and run on digital computers with physically separated memory and processing. We propose a hardware-software co-design, a semantic memory-based dynamic neural network (DNN) using memristor. The network associates incoming data with the past experience stored as semantic vectors. The network and the semantic memory are physically implemented on noise-robust ternary memristor-based Computing-In-Memory (CIM) and Content-Addressable Memory (CAM) circuits, respectively. We validate our co-designs, using a 40nm memristor macro, on ResNet and PointNet++ for classifying images and 3D points from the MNIST and ModelNet datasets, which not only achieves accuracy on par with software but also a 48.1% and 15.9% reduction in computational budget. Moreover, it delivers a 77.6% and 93.3% reduction in energy consumption.