Abstract:The brain performs intelligent tasks with extremely low energy consumption. This work takes its inspiration from two strategies used by the brain to achieve this energy efficiency: the absence of separation between computing and memory functions and reliance on low-precision computation. The emergence of resistive memory technologies indeed provides an opportunity to tightly co-integrate logic and memory in hardware. In parallel, the recently proposed concept of a Binarized Neural Network, where multiplications are replaced by exclusive NOR (XNOR) logic gates, offers a way to implement artificial intelligence using very low precision computation. In this work, we therefore propose a strategy for implementing low-energy Binarized Neural Networks that employs brain-inspired concepts while retaining the energy benefits of digital electronics. We design, fabricate, and test a memory array, including periphery and sensing circuits, that is optimized for this in-memory computing scheme. Our circuit employs hafnium oxide resistive memory integrated in the back end of line of a 130-nm CMOS process, in a two-transistor, two-resistor cell, which allows the exclusive NOR operations of the neural network to be performed directly within the sense amplifiers. We show, based on extensive electrical measurements, that our design allows a reduction in the number of bit errors on the synaptic weights without the use of formal error-correcting codes. We design a whole system using this memory array. We show on standard machine learning tasks (MNIST, CIFAR-10, ImageNet, and an ECG task) that the system has inherent resilience to bit errors. We evidence that its energy consumption is attractive compared to more standard approaches and that it can use memory devices in regimes where they exhibit particularly low programming energy and high endurance. We conclude the work by discussing how it associates biologically plausible ideas with more traditional digital electronics concepts.

Dot-product computation and logistic regression with 2D hexagonal-Boron Nitride (h-BN) memristor arrays

Hexagonal boron nitride (h-BN) memristor arrays for analog-based machine learning hardware

Unsupervised learning in hexagonal boron nitride memristor-based spiking neural networks

Memristor‐Based Analog Computation and Neural Network Classification with a Dot Product Engine

A Leaky Integrate-and-Fire Neuron Based on Hexagonal Boron Nitride (h-BN) Monocrystalline Memristor

Hybrid architecture based on two-dimensional memristor crossbar array and CMOS integrated circuit for edge computing

Implementation of high thermal conductivity and synaptic metaplasticity in vertically-aligned hexagonal boron nitride-based memristor

Highly Textured CMOS-compatible Hexagonal Boron Nitride-Based Neuristor for Reservoir Computing

Wafer-Scale 2D Hafnium Diselenide Based Memristor Crossbar Array for Energy-Efficient Neural Network Hardware

Hybrid 2D/CMOS microchips for memristive applications

In-Memory Computing With Memristor Arrays

Digital Biologically Plausible Implementation of Binarized Neural Networks With Differential Hafnium Oxide Resistive Memory Arrays

Advancing Memristive Analog Neuromorphic Networks: Increasing Complexity, and Coping with Imperfect Hardware Components

Introduction of Defects in Hexagonal Boron Nitride for Vacancy-Based 2D Memristors.

A Fully Hardware-Based Memristive Multilayer Neural Network

Programmable Multilevel Memtransistors Based on van der Waals Heterostructures

Two-dimensional Hexagonal Boron Nitride Based Memristor

High-uniformity Memristor Arrays Based on Two-dimensional MoTe<sub>2</sub> for Neuromorphic Computing

Analog Ion-Slicing LiNbO<sub>3</sub> Memristor Based on Hopping Transport for Neuromorphic Computing

Crystallinity‐Controlled Hexagonal Boron Nitride‐Based Memristors for Fully Integrated Reservoir Computing Processors

Wafer-scale integration of two-dimensional materials in high-density memristive crossbar arrays for artificial neural networks