About the Course
The course explores the concept of optical computing — a technology that uses light for transmitting and processing information. Unlike the classical von Neumann architecture, optical systems enable parallel data processing at the speed of light with low power consumption.
The course is built upon the SVETlANNa library for designing diffractive neural networks.
Syllabus
01
The basics of electrodynamics
- Maxwell's equations
- Harmonic waves
- Electromagnetic medium
- Wave equation
02
Fourier-optics
- Green’s function
- Weyl representation of spherical waves
- Longitudinal component of the field
03
Angular Spectrum Method
- Plane wave decomposition
04
Diffractive Optical Elements
- Thin transparency approach
- Thin lens
- Diffractive layer
- Spatial Light Modulator (SLM)
05
Backpropagation algorithm
- Common schema
- Forward pass
- Backward pass
06
Optimization methods
- Problem formulation
- Gradient descent
- Adam
- SGD
07
Optical computers
- Linear Diffractive Neural Network
- Diffractive Recurrent Neural Network
- Convolutional Diffractive Network
For Whom
Students
who are interested in neural networks, computational physics and machine learning
Researchers
who are interested in optical computing, diffractive neural networks and modern photonic technologies
Engineers
who are working on optical systems and AI accelerators
Start training
Access course materials or explore library documentation to start your journey into optical computing