Reservoir computing uses the changing states of a driven dynamical system to process information. Regular oscillator networks offer a framework for studying how coupling, delay and symmetry shape this approach to analogue computing.
Articles Tagged: reservoir computing
Photonic Reservoir Computing With Dual Feedback Loops
A dual-loop optoelectronic reservoir computing design adds a second nonlinear node and feedback delay to process temporal signals. Its simulated results across prediction, speech recognition, and channel equalization highlight both the promise of richer dynamics and the need to tune feedback parameters.
Spintronic Computing: A Path to Unconventional Systems
Spintronic devices offer nonvolatility, tunable dynamics and CMOS integration potential. These properties could help implement reservoir, probabilistic and memcomputing architectures for tasks that challenge conventional computing systems.
Photonic Reservoir Computing for Optical Signal Recovery
Photonic reservoir computing offers a hardware-based approach to recovering optical communication signals distorted by fibre dispersion and nonlinear effects. Experimental results show substantial bit-error-rate improvements, while also highlighting the latency and integration challenges that must be addressed before real-time telecom deployment.
Photonic Reservoir Computing Equalizers for Optical Links
A simulated silicon photonics reservoir computing equalizer uses delayed optical signal combinations and a trained readout to address link impairments. Results for metro and short-reach links show its potential as a nonlinear alternative to conventional linear equalization.