Investigating the use of Semiconductor Nanowires for Neural Networks

Abstract

In this work, we intend to employ nanowires for the realisation of an artificial neuron which can be used to design a neural network that will guide the new generation of non-von Neumann architectures such as Neuromorphic Computing. Possible applications are strongly related to Neuromorphic architectures and artificial intelligence, such as high-performance computers, robotics hardware and autonomous drones.

Here, we aim to design a device that can perform as a node in an optically communicating neural network. We employ nanowires to form a single artificial neuron by coupling a receiver and a transmitter of optical signals, together with a transistor as an in-between control element. Based on this configuration, we find optimal optical properties for our desired neural network communication. The node-to-node communication relies on the strength of the connection known as weights. The connection strength can be tailored by the position and rotation of the individual nanowires making up the nodes and layers; however, this might set some correlations and limitations on the weights.

We frame our main research questions as a set of hypotheses that we test by numerical experiments using the finite-difference time-domain (FDTD) method. The FDTD method is one of many numerical methods that exist to tackle this type of electromagnetic simulation. Among the valid alternatives are the finite element method and the method of the moment. However, we decided on the FDTD method because it is the one that best suits our needs to perform electromagnetic simulation on nanowires and to study a node-to-node communication.

We split our study and perform separate simulations on the receiver and transmitter. During the simulation of the receiver, we sweep the incoming light angle to study the absorption difference between two distinct depletion regions. The difference in absorption between the two regions yields a potential difference that allows switching on and off the transistor that controls the current through the transmitter. The emitted light from the transmitter is assumed proportional to the current. To study the transmitter, a dipole source was placed in the middle of a nanowire to study and evaluate the ability of the emitter to emit light in a specific direction, namely directivity.

The results from the receiver and transmitter studies are combined into a weight function after connecting the two simulated nanowires that form an artificial neuron. The weight function is our ultimate result from our experiment, which can be employed to design a custom neural network based on already known weights.

Popular Abstract

Springtime of a physical Artificial Neuron

Can we take advantage of the latest semiconductor technologies to create the next generation of brain-inspired computers and replace ordinary computers? Brain-inspired computers can be integrated into standard computers and used as independent units to tackle specific tasks in Artificial Intelligence. Moreover, thanks to their reduced power consumption and limited footprint, applications such as self-fly drones and supercomputers may benefit from it.

During the last century, many brilliant minds have contributed to the arrival of television, computers, and the internet. Artificial Neurons are exciting objects that have risen to the media’s attention due to impact on Artificial Intelligence and the Internet of Things. In general, an abstract Neural Network involves mathematical algorithms and codes that are the brain of this "thinking machine". On the other hand, a set of physical neurons can be made of semiconductor nanostructure materials that most computers and chips are made of nowadays. Therefore, a physical Artificial Neuron is the key to designing novel brain-inspired computers.
In our current project, we have investigated how to simulate an Artificial Neuron able
to communicate optically in a Neural Network under certain circumstances. We need two main components to construct Artificial Neurons: Receiver and
Transmitter nanowires as illustrated in the figure. These components are built using semiconductor materials with unique characteristics to absorb and emit light, making them very appreciated in many fields of science and research. The ones we employ are known as nanowires. We used a Receiver nanowire during our simulations, which can be thought as a solar cell panel that can absorb light from various illumination angles. The second is a Transmitter nanowire that emits wave signals like a light bulb or light- emitting diode (LED) that functions as a light source.
In this work, we show that we can tune the strength of the neuron-to-neuron communication in a Neural Network by rotating either a Transmitter or a Receiver nanowire, as exemplified in the figure. In addition, our main achievement, the Weight function, describes the strength of the neuron-to-neuron communication which is the key-role component to construct a model that best mimic a neural network.

• author: Serafini, Alfredo
• supervisor: Anders Mikkelsen; David Winge
• organization: Synchrotron Radiation Research; Department of Physics
• course: FYSM30 20212
• year: 2022
• type: H2 - Master's Degree (Two Years)
• subject: Physics and Astronomy
• keywords: FDTD method and simulation, optical broadcasting neural network, optical communication of nanowires, sub-wavelength absorption.
• language: English
• id: 9102513
• date added to LUP: 2022-11-02 11:36:31
• date last changed: 2022-11-02 11:36:31