In this paper, we propose a new step size rule to accelerate Yan et al.’s self-adaptive projection method. Under the new step size strategy, the superiority of modified projection method is verified through theory to numerical experiments.

Catastrophic forgetting is a key obstacle to continual learning. One of the state-of-the-art approaches orthogonal projection. The idea this approach learn each task by updating network parameters or weights only in direction subspace spanned all previous inputs. This ensures no interference with tasks that have been learned. system OWM uses performs very well against other systems. In paper, w...

We show that indirect pole-zero placement adaptive controllers are robust for systems with time-varying parameters as well as unmodeled dynamics and disturbances. We use a parameter estimator with projection. No special signal normalization is employed to ensure robustness. The nominal system parameters need only be bounded, and their variations need only be small in an average sense. This allo...