Abstract:The high demand for machine intelligence of doubling every three months is driving novel hardware solutions beyond charging of electrical wires given a resurrection to application specific integrated circuit (ASIC)-based accelerators. These innovations include photonic ASICs (P-ASIC) due to prospects of performing optical linear (and also nonlinear) operations, such as multiply-accumulate for vector matrix multiplications or convolutions, without iterative architectures. Such photonic linear algebra enables picosecond delay when photonic integrated circuits are utilized, via on-the-fly mathematics. However, the neurons full function includes providing a nonlinear activation function, knowns as thresholding, to enable decision making on inferred data. Many P-ASIC solutions performing this nonlinearity in the electronic domain, which brings challenges in terms of data throughput and delay, thus breaking the optical link and introducing increased system complexity via domain crossings. This work follows the notion of utilizing enhanced light-matter-interactions to provide efficient, compact, and engineerable electro-optic neuron nonlinearity. Here, we introduce and demonstrate a novel electro-optic device to engineer the shape of this optical nonlinearity to resemble a rectifying linear unit (ReLU) - the most-commonly used nonlinear activation function in neural networks. We combine the counter-directional transfer functions from heterostructures made out of two electro-optic materials to design a diode-like nonlinear response of the device. Integrating this nonlinearity into a photonic neural network, we show how the electrostatics of this thresholders gating junction improves machine learning inference accuracy and the energy efficiency of the neural network.

Graphene/Silicon Heterojunction Enabling Reconfigurable Phase-Relevant Optical Activation Function

Graphene/silicon heterojunction for reconfigurable phase-relevant activation function in coherent optical neural networks

Dual-function Optical Modulation and Detection in Microring Resonators Integrated Graphene/mote2 Heterojunction

Electro-optical Logic Gates Based on Graphene–silicon Waveguides

A Multifunctional Device Based on Graphene/Silicon Heterojunction

Design and Optimization of Optical Modulators Based on Graphene-on-silicon Nitride Microring Resonators

Reconfigurable All-Optical Integrated Nonlinear Activator with Switchable Response Functions for Photonic Neural Networks

Enhancing Device Efficiency Using a Reconfigurable Terahertz Wavefront Modulator Based on Fermi Level of Graphene

Multifunctional graphene optical modulator and photodetector integrated on silicon waveguides

An ITO Graphene hybrid integrated absorption modulator on Si-photonics for neuromorphic nonlinear activation

Characteristics of Electro-Refractive Modulating Based on Graphene-Oxide-Silicon Waveguide

All-Optical Nonlinear Activators Using Silicon-Based Waveguides and Microring Resonators for Photonic Neural Networks

Device Architectures for Low Voltage and Ultrafast Graphene Integrated Phase Modulators

All Optical Nonlinear Activation Function Based on Ge/Si Hybrid Micro-Ring Resonator

High-Speed Electro-Optic Modulator Integrated with Graphene-Boron Nitride Heterostructure and Photonic Crystal Nanocavity

Optical bistability and free carrier dynamics in graphene–silicon photonic crystal cavities

High-performance Fiber-Integrated Multifunctional Graphene-Optoelectronic Device with Photoelectric Detection and Optic-Phase Modulation

High Efficiency Electro-Optic Modulation in a Graphene Silicon Hybrid Tapered Microring Resonator

Regenerative oscillation and four-wave mixing in graphene optoelectronics

Active tuning of near-infrared electromagnetic responses in the graphene/silicon hybrid nanohole arrays by electrical control