Reducing the Variance of Variational Estimates of Mutual Information by Limiting the Critic's Hypothesis Space to RKHS.
P. Aditya SreekarUjjwal TiwariAnoop M. NamboodiriPublished in: CoRR (2020)
Keyphrases
- hypothesis space
- mutual information
- hypothesis spaces
- learning algorithm
- low variance
- learning tasks
- image registration
- version space
- reproducing kernel hilbert space
- image segmentation
- feature selection
- kernel methods
- real valued
- loss function
- similarity measure
- machine learning
- positive examples
- data mining
- pairwise