Event

ANNOUNCEMENT: Ph.D. Dissertation Defense

Name: Haoying Dai

 

Committee
Professor Yanne Chembo, Chair
Professor Thomus Murphy
Professor Kevin Daniels
Professor Gong Cheng

Professor Daniel Lathrop, Dean’s Representative

Date/Time: Tuesday, May 27, 2025 at 11:00 am

Location: ERF 1207,  Energy Research Facility 

Title:  PHOTONIC RESERVOIR COMPUTING BASED ON OPTOELECTRONIC OSCILLATORS

Abstract:

The Optoelectronic Oscillator is a versatile device that is widely in modern
communication systems - featuring as a distinctive component that is able to handle
both optical and RF signals with wide applications, including but not limited to
generating ultra-low phase noise signals, chaos communication, and sensing
technologies. In recent years, OEO-based reservoir computing has been proven to be a
simple and efficient ML platform with ultra-low latency and are suitable for applications
in communication networks.

 
In this defense, we will discuss two new types of OEO-based reservoir computing,
namely the (i) narrowband OEO-based RC and (ii) the broadband OEO-based RC. 
We will show that the narrowband OEO-based RC is suitable for processing the I-Q
modulated signals and the theory behind it.  We numerically simulate this narrowband
OEO-based RC and train it on IQ-modulation classification tasks, which reaches state-
of-the-art performance with a reduced training set size.

Then, we further extend the narrowband OEO-based RC to the field of RF fingerprinting
- a technology that is widely used to identify RF transmitters. In this chapter, we
thoroughly evaluate the performance of the narrowband OEO-based RC across a wide
range of benchmark datasets. Our simulation results demonstrate the suitability of the
narrowband OEO-based RC for RF fingerprinting. Moreover, we once again show that
the narrowband OEO not only requires significantly less training data for the IQ
modulation classification task but also maintains excellent performance under limited-
resource conditions, extending its effectiveness to RF fingerprinting.  We further push
the narrowband OEO-based RC to an extreme by limiting its computational source and
training data size. We experimentally test the narrowband OEO-based RC’s ability in
such scenarios. Meanwhile, we propose a new metric named NET to measure ML
platforms’ performances across different algorithms that takes into account the
complexity of the algorithm, the performance, and the data size required for training.
In the following, we introduce the concept of the directly laser-modulated OEOs for RCs
(DL-OEO based RC). In our proposed scheme, we cancel the external modulator and
replace it by directly modulating the laser, which not only significantly reduces the
system size, but also the total budget for implementation an OEO-based RCs. This
implementation makes our DL-OEO based RC one of the simplest OEO-based RCs
ever known. 

 
Lastly, we show the relationship between the number of cavity modes supported by
OEO and the total number of virtual nodes required by OEO-based RCs. We conduct
extensive simulations of both types of OEO-based RCs (including narrowband and
wideband configurations) across a diverse set of datasets. Our simulation shows that
the total number of virtual nodes required for sufficiently good performance of OEO-
based RC can be as less as the total number of cavity modes .