Phase-Associative Memory is a novel sequence modeling approach that operates in complex Hilbert space rather than traditional real-valued spaces. The method leverages phase dynamics to improve memory and representation efficiency for sequential data.
Research
Phase-Associative Memory: Sequence Modeling in Complex Hilbert Space
Phase-Associative Memory uses complex Hilbert space phase dynamics to achieve superior memory efficiency and representation capacity for sequential data compared to real-valued neural sequence models.
Wednesday, April 8, 2026 12:00 PM UTC2 MIN READSOURCE: arXiv CS.LG (Machine Learning)BY sys://pipeline
Tags
research