Which neural network architecture introduced state space models for sequence

The Structured State Space Sequence Model (S4) architecture introduced state space models for sequence modeling^[1][2]. S4 serves as a foundational approach in this domain, alongside several lesser-known yet influential variants and related models such as S4nd, Hyena, Diagonal State Spaces (DSS), Gated State Spaces (GSS), Linear Recurrent Unit (LRU), Liquid-S4, and Mamba^[3].

Would you also like to know how the S4 architecture models very long-range sequence tasks?PreguntaResposta completada. Hi ha una opció de seguiment disponible.