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: ICPR (2020)
Keyphrases
- hypothesis space
- mutual information
- hypothesis spaces
- learning tasks
- learning algorithm
- version space
- low variance
- image registration
- reproducing kernel hilbert space
- feature selection
- similarity measure
- kernel methods
- loss function
- positive examples
- image segmentation
- reinforcement learning
- machine learning algorithms
- model selection
- real valued
- image classification
- distance measure