| This study presents Memory enhanced Liquid Time- constant Networks (LTCs) to enhance the temporal and memory dynamics of recurrent models. LTCs, characterized by liquid time constants offer adaptive and stable representations for time-series prediction by modulating hidden states with input- dependent dynamics. Integrating memory features that utilize attractor dynamics to model memory mechanisms explicitly further, augment the system’s representational and computational efficiency. By leveraging the inherent strengths of memory retention and LTCs’ expressivity in capturing continuous time processes, the proposed framework is designed for advanced time-series modeling tasks |
*** Title, author list and abstract as submitted during Camera-Ready version delivery. Small changes that may have occurred during processing by Springer may not appear in this window.