seminar
SIGCOM Seminar: Time-Series Forecasting with Quantum Reservoir Computing
Seminar slides on quantum reservoir computing, time-series forecasting, and the practical steps from input encoding to classical readout.
On 24 July 2026, I presented Time-Series Forecasting with Quantum Reservoir Computing at the SIGCOM seminar at SnT, University of Luxembourg.
Open the interactive presentation
The talk starts with forecasting from a recent history window and follows the steps needed to use a quantum reservoir: encode the input, evolve a fixed quantum system, measure useful features, and train a classical readout. I also discuss how this differs from training a variational quantum circuit.
A six-atom teaching example makes the encoding, observables, and measurement process concrete. The slides distinguish exact emulator features from finite-shot measurement estimates and explain why a useful feature map alone does not establish quantum advantage. The final section considers how classical and quantum processors can share the work, with accuracy and computational cost evaluated against classical baselines.
The presentation includes animations and interactive explanations and can be viewed directly in your browser.
Related reading
- Fujii and Nakajima, Harnessing Disordered-Ensemble Quantum Dynamics for Machine Learning, 2017.
- Kornjača and colleagues, Large-scale quantum reservoir learning with an analog quantum computer, 2024.