Neural-network training can be slow and energy intensive, owing to the need to transfer the weight data for the network between conventional digital memory chips and processor chips. Analogue non-volatile memory can accelerate the neural-network training algorithm known as backpropagation by performing parallelized multiply-accumulate operations in the analogue domain at the location of the weight data. However, the classification accuracies of such in situ training using non-volatile-memory hardware have generally been less than those of software-based training, owing to insufficient dynamic range and excessive weight-update asymmetry. Here we demonstrate mixed hardware-software neural-network implementations that involve up to 204,900 synapses and that combine long-term storage in phase-change memory, near-linear updates of volatile capacitors and weight-data transfer with 'polarity inversion' to cancel out inherent device-to-device variations. We achieve generalization accuracies (on previously unseen data) equivalent to those of software-based training on various commonly used machine-learning test datasets (MNIST, MNIST-backrand, CIFAR-10 and CIFAR-100). The computational energy efficiency of 28,065 billion operations per second per watt and throughput per area of 3.6 trillion operations per second per square millimetre that we calculate for our implementation exceed those of today's graphical processing units by two orders of magnitude. This work provides a path towards hardware accelerators that are both fast and energy efficient, particularly on fully connected neural-network layers.

Custom low power hardware for real-time network security and anomaly detection are in great demand, as these would allow for efficient security in battery-powered network devices. This paper presents a memristor based system for real-time intrusion detection, as well as an anomaly detection based on autoencoders. Intrusion detection is based on a single autoencoder, and the overall detection accuracy of this system is 92.91% with a malicious packet detection accuracy of 98.89%. The system described in this paper is also capable of using two autoencoders to perform anomaly detection using real-time online learning. Using this system, we show that anomalous data is flagged by the system, but over time the system stops flagging a particular datatype if its presence is abundant. Utilizing memristors in these designs allows us to present extreme low power systems for intrusion and anomaly detection, while sacrificing little accuracy. CCS CONCEPTSCCS → Hardware → Emerging technologies → Analysis and design of emerging devices and systems → Emerging architectures

No abstract