RESEARCH SPOTLIGHT
Hi, I am
Aycan Deniz Vit
HOST INSTITUTE
Ghent University
PROJECT TITLE
Photonic Reservoir Computing for Telecom Applications
KEY WORDS
Photonic reservoir computing; optical communications; signal equalization; integrated photonics; machine learning
RESEARCH AIM
My research aims to develop compact photonic processors that recover distorted optical signals so that future communication networks can transmit data more reliably and efficiently.
RESEARCH OVERVIEW
My research explores how light can help recover information distorted during transmission through optical fibres. Tiny networks of optical paths mix the incoming signal with its recent history, giving the system a short memory. By learning how to combine these signals, the system can estimate the original data. I use simulations and planned experiments to investigate how these optical processors behave and how well they recover information under different conditions. The aim is to guide the development of compact photonic technologies that could make future communication networks more efficient and reliable.
RESEARCH CHALLENGE
Optical signals become distorted as they travel through fibre, making the original data harder to recover. Correcting this distortion becomes increasingly demanding as communication speeds rise. My research asks how compact photonic reservoirs can help recover the transmitted information while keeping the processing practical. A central challenge is balancing signal recovery, robustness and system complexity, and understanding how performance in simulation translates to an experimental setup.
RESEARCH INNOVATION
The approach uses the physical behaviour of light to help process information. Propagation and interference in a photonic reservoir create a memory of the incoming signal, and a trained readout combines those optical states to recover the data. This connects device architecture, learning and communication-system performance in one study. Combining numerical benchmarks with experiments on reservoir architectures and coherent detection will help assess how these ideas could work in practice.
RESEARCH IMPACT
What excites me most is the idea that the physics of light can become part of the computation. A network of optical paths can create useful memory through propagation and interference, and learning determines how to use that information. I especially enjoy the connection between a physical system, an optimization algorithm and a concrete outcome: recovering the transmitted data more accurately.
RESEARCH SUMMARY
My research investigates integrated photonic reservoir computing for signal equalization in optical telecommunications. Passive networks of waveguides mix an incoming optical signal with its recent history, creating temporal features that a trained readout combines to recover transmitted information. I explore model-free optimization methods, including covariance matrix adaptation evolution strategies (CMA-ES) and parameter-exploring policy gradients (PEPG), to train readout coefficients using performance feedback without requiring derivatives of the photonic system. A particular focus is optimizing communication performance directly through bit-error-based objectives and comparing these approaches with conventional regression methods. Using reproducible simulations of intensity-modulated and coherent links, I evaluate signal recovery, convergence and robustness across operating conditions. The goal is to identify practical training strategies for compact photonic equalizers and inform their future implementation in communication systems.
RESEARCH VISUAL
RESEARCH JOURNEY
RESEARCH SECONDMENT
VLC Photonics - Spain - 2 months. I will carry out experimental work on new photonic reservoir architectures. I expect to develop practical skills in photonic device testing, characterisation and the experimental evaluation of reservoir performance.
Aston University - United Kingdom - 3 months. I will carry out coherent detection experiments on photonic reservoirs. I expect to strengthen my skills in coherent optical receiver experiments, signal acquisition and processing, and the evaluation of reservoir equalization performance.
LEARNING GOALS
At VLC Photonics, I am looking forward to gaining practical experience with new reservoir architectures and learning how their behaviour is characterised experimentally. At Aston University, I look forward to developing my skills in coherent detection experiments and analysing the resulting signals. Together, these placements will help me connect numerical studies with the practical realities of optical hardware and receiver systems.
RESEARCH BENEFITS
The VLC Photonics secondment will allow me to explore new reservoir architectures experimentally, while the Aston University secondment will provide experience with coherent detection experiments on reservoirs. These complementary activities will help me test assumptions from simulation, understand experimental constraints and assess the relevance of my training methods to physical systems. They will also strengthen collaboration around the project.
CAREER DEVELOPMENT
They would help me build experience across research environments, communicate across disciplines and develop an international professional network. Exposure to academic and industrial perspectives would also help me understand where my skills in photonics, optimization and signal processing can make the strongest contribution after the PhD.
MORE THAN PHD
Why did you decide to join POSTDIGITAL+ ?
POSTDIGITAL+ brings together the areas that my project depends on: emerging computing hardware, learning methods and practical applications. The combination of doctoral training, international collaboration and links with industry offers an opportunity to develop both a strong technical foundation and a broader view of how research can become useful technology.
What do you think is the biggest advantage of being part of the International Doctoral Network?
The biggest advantage is the opportunity to connect complementary expertise across universities, research institutes and industry. For a project combining photonics, machine learning and telecommunications, this brings different perspectives on the same challenge and helps link fundamental research with practical needs. Shared training, secondments and exchanges with other doctoral researchers also create a supportive community and opportunities for lasting international collaboration.
What skill have you developed the most so far?
Turning a broad research question into a reproducible numerical experiment. This combines scientific programming, signal processing and optimization with careful choices about training, validation and benchmarking. It also means learning to investigate unexpected results and explain clearly what the evidence does and does not show.
What are your career ambitions after completing your doctorate?
What inspired you to pursue this research Area?
I am drawn to the intersection of physics and computation: using the behaviour of a physical system to process information. Photonic reservoir computing brings that idea together with machine learning and a concrete engineering challenge in optical communications. The possibility of turning propagation and interference into useful computation makes this a particularly engaging research area for me.
What surprised you most since starting your PhD?
How much a useful result depends on asking the right question and designing a fair comparison. In interdisciplinary research, understanding the assumptions behind a method can be just as important as implementing it. The network also highlights how differently researchers can approach similar problems, and how valuable it is to exchange those perspectives.