SAGL learns subspace-preserving sparse attention graphs from heterogeneous multiview data via bilinear attention factorization, dynamic sparsity gating, and α-entmax projection, outperforming prior unsupervised transfer learning methods.
We adopt centered kernel alignment (CKA) [ 37] to measure the simi- larity between feature distributions produced by different pretrained model pairs
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
other 1
citation-polarity summary
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1roles
other 1polarities
unclear 1representative citing papers
citing papers explorer
-
Learning Subspace-Preserving Sparse Attention Graphs from Heterogeneous Multiview Data
SAGL learns subspace-preserving sparse attention graphs from heterogeneous multiview data via bilinear attention factorization, dynamic sparsity gating, and α-entmax projection, outperforming prior unsupervised transfer learning methods.