An analytic theory of generalization dynamics and transfer learning in deep linear networks.
Andrew K. LampinenSurya GanguliPublished in: ICLR (Poster) (2019)
Keyphrases
- transfer learning
- learning tasks
- knowledge transfer
- cross domain
- labeled data
- collaborative filtering
- algorithmic information theory
- active learning
- semi supervised learning
- manifold alignment
- machine learning
- multi task
- transfer knowledge
- domain adaptation
- reinforcement learning
- text classification
- machine learning algorithms
- multi task learning
- structure learning
- transferring knowledge
- text categorization
- databases
- unlabeled data
- information extraction
- high dimensional
- learning algorithm
- neural network
- network structure
- dynamical systems
- text mining
- decision trees