Computable multiclass PAC learnability for finite label spaces is characterized exactly by the finiteness of computable distinguisher dimensions, including a new computable Natarajan dimension.
Multiclass boosting and the cost of weak learning
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
On the Computability of Multiclass PAC Learning
Computable multiclass PAC learnability for finite label spaces is characterized exactly by the finiteness of computable distinguisher dimensions, including a new computable Natarajan dimension.