For general convex learning tasks whose optima share a low-rank or clustered structure, this paper gives sample-complexity bounds for recovering that structure, including a one-sample-per-task regime where the number of tasks must be exponential in the rank.
Trace norm regularization for multi-task learning with scarce data
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
-
Meta-learning of shared linear representations beyond well-specified linear regression
For general convex learning tasks whose optima share a low-rank or clustered structure, this paper gives sample-complexity bounds for recovering that structure, including a one-sample-per-task regime where the number of tasks must be exponential in the rank.