Task-Relevant Null-Space Residuals for Non-Injective Neural Mappings
The proposed Task-Relevant Null-Space Residuals (NSR) combines null-space component extraction from pre-mapping representations, member-level encoding and gating, and application-specific integration to exploit potentially task-relevant information under downstream supervision while preserving the original aggregation...
Bi-Zu Feng, Zhi-Mu Yang, Shu-Ming Wang et al.
· 0 citations