Search papers, labs, and topics across Lattice.
This paper introduces Dynamical Mode Pruning (DMP), a novel method for pruning Echo State Networks (ESNs) by evaluating neurons based on their contributions to dominant transition modes derived from the trajectory-averaged Jacobian Gramian. By focusing on dynamic influences rather than static connectivity, DMP effectively reduces over-parameterization in ESNs while maintaining or enhancing forecasting accuracy on both chaotic and real-world time-series data. The findings indicate that incorporating dynamical perspectives into reservoir refinement can lead to more efficient and effective temporal prediction models.
Pruning Echo State Networks dynamically can enhance forecasting accuracy while significantly reducing model complexity.
Echo State Networks (ESNs) offer an efficient framework for temporal prediction, but their randomly initialized reservoirs are often over-parameterized and dynamically redundant. Existing pruning methods largely rely on static connectivity or activation statistics, which may overlook neurons that shape input-driven state transitions. We propose Dynamical Mode Pruning (DMP), a reservoir pruning method that ranks neurons by their contribution to dominant transition modes obtained from a trajectory-averaged Jacobian Gramian. DMP removes low-impact units and retrains only the readout. Experiments on chaotic and real-world time-series benchmarks show that DMP improves or preserves forecasting accuracy while reducing redundant reservoir components. Our results suggest that dynamical influence is a useful criterion for reservoir refinement beyond static structural importance alone.
