Skip to content

IDEA / PHOTONIC INFORMATION

Dynamics that compute

Can optical nonlinearity, memory and coupling do useful computation?

Editorial synthesis by the site author across the reading notes below. It is an interpretation, not a finding reported by any single paper. Each note keeps the authors’ findings separate from commentary and links the original source.

Working summary

VCSEL spikes and programmable semiconductor nonlinearities provide physical operations for information processing. Device dynamics and the encoding, readout and training workflow jointly determine performance.

Connections

Spin-flip and coupled-mode models explain the dynamical resources; hardware experiments test whether those resources survive real noise and control constraints.

Draws on the reading notes for: A laser that predicts a time series, Programmable optical activation, Anisotropy changes the laser dynamics, Two lasers as an oscillator.

Questions and limits

An optical operation time does not establish end-to-end speed or energy efficiency. Include electronic control, memory, conversion and training in system comparisons.

SOURCES

Referenced articles

Ordered by publication year. A shared theme does not imply direct citation or historical influence.

2023 / Theory & simulation

Spin vertical-cavity surface-emitting lasers with linear gain anisotropy: Prediction of exceptional points and nontrivial dynamical regimes

M. Drong, J. Peřina Jr., T. Fördös, H. Y. Jaffrès, K. Postava & H.-J. Drouhin

Extends the spin-flip model and studies polarization dynamics near exceptional points.

Summary & commentary : Spin vertical-cavity surface-emitting lasers with linear gain anisotropy: Prediction of exceptional points and nontrivial dynamical regimes
2025 / Experiment

Photonic spiking neural network built with a single VCSEL for high-speed time series prediction

Dafydd Owen-Newns, Lina Jaurigue, Joshua Robertson, Andrew Adair, Jonnel Anthony Jaurigue, Kathy Lüdge & Antonio Hurtado

Uses a single-VCSEL spiking-network architecture with temporally separated inputs that provide memory.

Summary & commentary : Photonic spiking neural network built with a single VCSEL for high-speed time series prediction

RELATED NOTES

Related notes