Rescaled mirror flows converge to a limit whose primal variable incrementally minimizes quadratic loss over a subdifferential-defined time-dependent hypothesis set.
Implicit regularization for group sparsity
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
math.OC 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Incremental Learning in Mirror Flows
Rescaled mirror flows converge to a limit whose primal variable incrementally minimizes quadratic loss over a subdifferential-defined time-dependent hypothesis set.