The ECL-RR method combines a revised mean/variance regularizer, a parameter-learning-network compression, and a common/distinctive subspace decomposition, and reports top accuracy on four continual learning benchmarks.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2024 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
Revised Regularization for Efficient Continual Learning through Correlation-Based Parameter Update in Bayesian Neural Networks
The ECL-RR method combines a revised mean/variance regularizer, a parameter-learning-network compression, and a common/distinctive subspace decomposition, and reports top accuracy on four continual learning benchmarks.