Nonlinear Schr\"odinger Kernel for hardware acceleration in machine learning inference

TL;DR

This paper introduces a novel optical nonlinear kernel method leveraging femtosecond optical pulses and nonlinear dynamics to significantly improve data classification accuracy and reduce latency in machine learning tasks, especially with limited training data.

Contribution

The paper presents a new optical nonlinear kernel approach that enhances classification performance and speed, validated across diverse datasets and demonstrated with single-shot operation.

Findings

Increased classification accuracy with optical nonlinear dynamics

Reduced latency by several orders of magnitude

Effective with small training datasets

Abstract

Alternative machine learning approaches that are computationally light with low latency and can work with only a small training dataset are needed for applications where the insatiable demand of deep learning methods for computing power and large training data cannot be met. We show that spectral mapping of data onto femtosecond optical pulses and a projection into an implicit, higher dimensional space using nonlinear optical dynamics increases the accuracy and reduces the latency in data classification by several orders of magnitude. The approach is validated by the classification of various datasets, including brain intracranial pressure, cancer cell imaging, spoken digit recognition, and the classic exclusive OR benchmark for nonlinear classification. The concept is demonstrated by seeding the nonlinear dynamics that are responsible for many fascinating natural phenomena, such as optical rogue waves, with the data before processing the output with a light classifier. A quantitative comparison with a well-known numerical technique is used to provide insight into this physical technique. Single-shot operation is demonstrated using time stretch data acquisition.