# Femto-joule threshold reconfigurable all-optical nonlinear activators for picosecond pulsed optical neural networks

## TL;DR

Researchers created an efficient, ultra-fast optical computing component that could enable powerful optical neural networks.

## Contribution

A reconfigurable all-optical nonlinear activator with femto-joule thresholds and picosecond response times for optical neural networks.

## Key findings

- The activator achieved a saturable absorption energy threshold of 4 fJ and a response time of 1.05 ps.
- The device has a reconfigurable nonlinear activation threshold of 30 fJ and a response time of 4 ps.
- The device size is 15 μm × 10 μm, enabling 10⁶ TOPs/W/mm² optical computing.

## Abstract

Achieving optical computing with thousands of tera-operations per second per watt per square millimeter (TOPs/W/mm2) is the key to surpassing electrical computing. This realization requires a breakthrough in the design of a new optical computing architecture and nonlinear activation functions. By leveraging the Kerr effect of silicon and the saturable absorption of graphene, we designed an all-optical nonlinear activator based on a graphene-silicon integrated photonic crystal cavity. The ultralow-threshold, high-speed, compact, and reconfigurable all-optical nonlinear activator could achieve a saturable absorption energy threshold of 4 fJ and a response time of 1.05 ps, a reconfigurable nonlinear activation threshold of 30 fJ and a response time of 4 ps, and an ultrasmall size of 15 μm × 10 μm. This device provides foundation blocks for the picosecond pulsed optical neural network chip to achieve 106 TOPs/W/mm2 level optical computing.

We developed a graphene-silicon integrated photonic crystal cavity-based all-optical activator (ultralow threshold, fast, compact, reconfigurable) enabling picosecond optical neural network chips with high-efficiency computing.