A token-space SVD-based method that separately resolves gradient conflicts in the range and null spaces of transformer tokens improves multi-task learning performance with minimal extra parameters.
Mod-squad: Designing mixtures of experts as modular multi-task learners
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
background 1representative citing papers
citing papers explorer
-
Resolving Token-Space Gradient Conflicts: Token Space Manipulation for Transformer-Based Multi-Task Learning
A token-space SVD-based method that separately resolves gradient conflicts in the range and null spaces of transformer tokens improves multi-task learning performance with minimal extra parameters.