Multi-task models trained in a learned low-dimensional random subspace need far fewer parameters per task, and this enables non-vacuous generalization bounds for deep multi-task learning.
PAC-Bayes Bounds for Meta-learning with Data-Dependent Prior
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
abstract
By leveraging experience from previous tasks, meta-learning algorithms can achieve effective fast adaptation ability when encountering new tasks. However it is unclear how the generalization property applies to new tasks. Probably approximately correct (PAC) Bayes bound theory provides a theoretical framework to analyze the generalization performance for meta-learning. We derive three novel generalisation error bounds for meta-learning based on PAC-Bayes relative entropy bound. Furthermore, using the empirical risk minimization (ERM) method, a PAC-Bayes bound for meta-learning with data-dependent prior is developed. Experiments illustrate that the proposed three PAC-Bayes bounds for meta-learning guarantee a competitive generalization performance guarantee, and the extended PAC-Bayes bound with data-dependent prior can achieve rapid convergence ability.
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
From Low Intrinsic Dimensionality to Non-Vacuous Generalization Bounds in Deep Multi-Task Learning
Multi-task models trained in a learned low-dimensional random subspace need far fewer parameters per task, and this enables non-vacuous generalization bounds for deep multi-task learning.