Search for: All records

In this paper, we present AnalogVNN, a simulation framework built on PyTorch that can simulate the effects of optoelectronic noise, limited precision, and signal normalization present in photonic neural network accelerators. We use this framework to train and optimize linear and convolutional neural networks with up to nine layers and ∼1.7 × 106 parameters, while gaining insights into how normalization, activation function, reduced precision, and noise influence accuracy in analog photonic neural networks. By following the same layer structure design present in PyTorch, the AnalogVNN framework allows users to convert most digital neural network models to their analog counterparts with just a few lines of code, taking full advantage of the open-source optimization, deep learning, and GPU acceleration libraries available through PyTorch. 

Optical phase-change materials have enabled nonvolatile programmability in integrated photonic circuits by leveraging a reversible phase transition between amorphous and crystalline states. To control these materials in a scalable manner on-chip, heating the waveguide itself via electrical currents is an attractive option which has been recently explored using various approaches. Here, we compare the heating efficiency, fabrication variability, and endurance of two promising heater designs which can be easily integrated into silicon waveguides—a resistive microheater using n-doped silicon and one using a silicon p-type/intrinsic/n-type (PIN) junction. Raman thermometry is used to characterize the heating efficiencies of these microheaters, showing that both devices can achieve similar peak temperatures but revealing damage in the PIN devices. Subsequent endurance testing and characterization of both device types provide further insights into the reliability and potential damage mechanisms that can arise in electrically programmable phase-change photonic devices.

 

Abstract The exponential growth of information stored in data centers and computational power required for various data-intensive applications, such as deep learning and AI, call for new strategies to improve or move beyond the traditional von Neumann architecture. Recent achievements in information storage and computation in the optical domain, enabling energy-efficient, fast, and high-bandwidth data processing, show great potential for photonics to overcome the von Neumann bottleneck and reduce the energy wasted to Joule heating. Optically readable memories are fundamental in this process, and while light-based storage has traditionally (and commercially) employed free-space optics, recent developments in photonic integrated circuits (PICs) and optical nano-materials have opened the doors to new opportunities on-chip. Photonic memories have yet to rival their electronic digital counterparts in storage density; however, their inherent analog nature and ultrahigh bandwidth make them ideal for unconventional computing strategies. Here, we review emerging nanophotonic devices that possess memory capabilities by elaborating on their tunable mechanisms and evaluating them in terms of scalability and device performance. Moreover, we discuss the progress on large-scale architectures for photonic memory arrays and optical computing primarily based on memory performance.