Optical Neural Networks

Optical neural networks (ONNs) have the potential to overcome scaling limitations of transistor-based systems due to their inherent low latency and large available bandwidth. However, encoding the information directly in the physical properties of light fields also imposes new computational constraints, for example the restriction to only positive intensity values for incoherent photonic processors. We aim to address emerging challenges and develop novel algorithms to overcome the physical limitations of ONNs. In addition, we are using machine learning methods to facilitate the fabrication of ONNs.

Reinforcement Learning for Nanophotonic Inverse Design
© David Lemli

Physical Chemical Activations Functions
© cvmls

Related Publications:

Becker, M., Butz, M., Lemli, D., Schuck, C., & Risse, B. (2024). Learning Proposal Distributions in Simulated Annealing via Template Networks: A Case Study in Nanophotonic Inverse Design. International Conference on Pattern Recognition (pp. 188–202)
Schulte, L., Butz, M., Becker, M., Risse, B., & Schuck, C. (2024). Accelerating Finite-Difference Frequency-Domain Simulations for Inverse Design Problems in Nanophotonics using Deep Learning. Journal of the Optical Society of America B, 41 (4), (pp. 1039–1046)
Brückerhoff-Plückelmann, F., Borras, H., Klein, B., Varri, A., Becker, M., Dijkstra, J., Brückerhoff, M., Wright, CD., Salinga, M., Bhaskaran, H., Risse, B., Fröning, H., & Pernice, W. (2024). Probabilistic photonic computing with chaotic light. Nature Communications, 15 (1), (pp. 10445–10445)
Wendland, D., Becker, M., Brückerhoff-Plückelmann, F., Bente, I., Busch, K., Risse, B., & Pernice, WH. (2023). Coherent dimension reduction with integrated photonic circuits exploiting tailored disorder. Journal of the Optical Society of America B, 40 (3), (pp. B35–B40)
Becker, M., Riegelmeyer, J., Seyfried, M., Ravoo, B.-J., Schuck, C., & Risse, B. (2023). Adaptive Photo-Chemical Nonlinearities for Optical Neural Networks. Advanced Intelligent Systems, ‎ 5 (12)
Brückerhoff-Plückelmann, F., Bente, I., Becker, M., Vollmar, N., Farmakidis, N., Lomonte, E., Lenzini, F., Wright, C. D., Bhaskaran, H., Salinga, M., Risse, B., & Pernice, W. HP. (2023). Event-driven adaptive optical neural network. Science advances, 9 (42)
Becker, M., Drees, D., Brückerhoff-Plückelmann, F., Schuck, C., Pernice, W., & Risse, B. (2023). Activation Functions in Non-Negative Neural Networks. Machine Learning and the Physical Sciences Workshop, ML4PS Workshop @NeurIPS