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.

Computationally Efficient Low Power Neuron Model for Digital Brain

A Power Efficient Hardware Implementation of the IF Neuron Model.

Ultra-Low Power Silicon Neuron Circuit for Extreme-Edge Neuromorphic Intelligence

Ultra low power of artificial cognitive memory for brain-like computation

Low-Power Neuromorphic Hardware for Signal Processing Applications

A Brain-Inspired In-Memory Computing System for Neuronal Communication Via Memristive Circuits.

Subthreshold CMOS Implementation of the Izhikevich Neuron Model

Design and Performance Analysis of an Ultra-Low Power Integrate-and-Fire Neuron Circuit Using Nanoscale Side-contacted Field Effect Diode Technology

A Versatile Neuromorphic System Based on Simple Neuron Model

A Low-Power Vertical Dual-Gate Neurotransistor with Short-Term Memory for High Energy-Efficient Neuromorphic Computing

Implementation of a 12-Million Hodgkin-Huxley Neuron Network on a Single Chip

μBrain: An Event-Driven and Fully Synthesizable Architecture for Spiking Neural Networks

55 Nm CMOS Mixed-Signal Neuromorphic Circuits for Constructing Energy-Efficient Reconfigurable SNNs

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

Band-to-Band Tunneling based Ultra-Energy Efficient Silicon Neuron

Efficient Neural Networks Algorithm Based on Connectivity Properties of Biological Brain Networks

AGN: A Power-Adaptive and Growth-Capable Neuron Model and Circuit Implementation

Compact Digital-Controlled Neuromorphic Circuit with Low Power Consumption

Toward Human-Scale Brain Computing Using 3D Wafer Scale Integration

eBrainII: a 3 kW Realtime Custom 3D DRAM Integrated ASIC Implementation of a Biologically Plausible Model of a Human Scale Cortex

Emerging Devices for Biologically Accurate Neuron