Flow-based Spectral Kernel Learning for Nonstationary Attention
Kernelized attention provides a useful perspective for understanding self-attention as a similarity-based aggregation mechanism, but existing approaches often rely on fixed or stationary kernels, limiting their ability to adapt the similarity structure to complex token interactions.
By Zi Yang, Ying Li, Alejandro Lancho, Dar\'{i}o C. Larese, Pablo M. Olmos, Michael Minyi ZhangarXiv - stat.ML
Read the full story
Originally published by arXiv - stat.ML on Oct 23, 2026.
Comments
Sign in to join the discussion.