Neuromorphic Computing Based on Parametrically-Driven Oscillators and Frequency Combs

In a groundbreaking study published on arXiv (2604.21861), researchers Mahadev Sunil Kumar and Adarsh Ganesan have demonstrated that parametrically-driven oscillators can serve as a natural platform for neuromorphic computation. Their two-mode system, which exhibits 2:1 parametric resonance, operates as a reservoir computer capable of performing one step-ahead prediction of benchmark chaotic systems, including Mackey-Glass, Rossler, and Lorenz dynamics. The key innovation lies in encoding input signals into the drive amplitude and sampling both temporal and spectral responses to achieve high-dimensional transformation and memory.

The study reveals that optimal computational performance is achieved within the parametric resonance regime, where nonlinear interactions are activated while temporal coherence is preserved. In contrast, frequency-comb states, despite offering increased spectral dimensionality, show inconsistent performance across their existence band and degrade in chaotic comb regimes due to loss of phase coherence. The researchers mapped prediction error over parameter space, establishing a direct correspondence between computational capability and the underlying bifurcation structure, with low-error regions aligned with the parametric resonance boundary. These findings provide design principles for tuning physical systems toward optimal neuromorphic functionality.