Hardware implementation of Bayesian network building blocks with stochastic spintronic devices

TL;DR

This paper demonstrates a hardware implementation of Bayesian network building blocks using stochastic spintronic devices, offering a promising approach to efficiently model complex probabilistic systems.

Contribution

It introduces an experimental spintronic device-based hardware platform for Bayesian networks, enabling stochastic behavior and scalable implementation of probabilistic models.

Findings

Successfully built a two-node stochastic network with controllable correlations

Mapped conditional probability tables to the spintronic hardware

Simulated a four-node Bayesian network using the proposed devices

Abstract

Bayesian networks are powerful statistical models to understand causal relationships in real-world probabilistic problems such as diagnosis, forecasting, computer vision, etc. For systems that involve complex causal dependencies among many variables, the complexity of the associated Bayesian networks become computationally intractable. As a result, direct hardware implementation of these networks is one promising approach to reducing power consumption and execution time. However, the few hardware implementations of Bayesian networks presented in literature rely on deterministic CMOS devices that are not efficient in representing the inherently stochastic variables in a Bayesian network. This work presents an experimental demonstration of a Bayesian network building block implemented with naturally stochastic spintronic devices. These devices are based on nanomagnets with perpendicular magnetic anisotropy, initialized to their hard axes by the spin orbit torque from a heavy metal under-layer utilizing the giant spin Hall effect, enabling stochastic behavior. We construct an electrically interconnected network of two stochastic devices and manipulate the correlations between their states by changing connection weights and biases. By mapping given conditional probability tables to the circuit hardware, we demonstrate that any two node Bayesian networks can be implemented by our stochastic network. We then present the stochastic simulation of an example case of a four node Bayesian network using our proposed device, with parameters taken from the experiment. We view this work as a first step towards the large scale hardware implementation of Bayesian networks.