Two recent papers in Chaos journal by Loughborough University researchers highlight exciting advances in both photonic and quantum reservoir computing
- 7 minutes ago
- 1 min read
🚀 Two recent papers in Chaos journal by Loughborough University researchers highlight exciting advances in both photonic and quantum reservoir computing.
📄 Unwrapping photonic reservoirs: Enhanced expressivity via random Fourier encoding over stretched domains demonstrates that deliberately “over-wrapping” optical phase information beyond its conventional 2π range can improve computational performance. By exploiting nonlinear interference and synthetic frequencies, the authors show that photonic reservoirs can tackle regression and classification problems previously considered out of reach, opening new possibilities for designing more powerful optical computing systems.
📄 Minimal quantum reservoirs with Hamiltonian encoding introduces an elegant and experimentally practical quantum reservoir computing architecture in which input data are encoded directly into Hamiltonian parameters rather than quantum states. The approach avoids many of the usual quantum hardware challenges - such as complex state preparation, feedback, and state tomography - yet still achieves nonlinear regression and prediction tasks using only a minimal number of qubits and delay embeddings.
Researchers involved were Luana Olivieri, Girish Tripathy, Giulia Marcucci, Juan Sebastian Totero Gongora and Gerard McCaul.
Together, these studies reveal a common theme: smarter encoding strategies can dramatically enhance the computational power of physical reservoir computing systems while simplifying their implementation. 🌟
📚 Chaos 35 (2025), Articles 093129 and 093135.