Minimax regret is characterized for misspecified universal learning with log-loss, yielding the optimal universal learner as a unified framework for any uncertainty in the data-generating process.
Information- theoretic asymptotics of bayes methods
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.IT 1years
2026 1verdicts
UNVERDICTED 1roles
background 1polarities
background 1representative citing papers
citing papers explorer
-
Misspecified Universal Learning
Minimax regret is characterized for misspecified universal learning with log-loss, yielding the optimal universal learner as a unified framework for any uncertainty in the data-generating process.