The Optimal Noise in Noise-Contrastive Learning Is Not What You Think.
Omar ChehabAlexandre GramfortAapo HyvärinenPublished in: CoRR (2022)
Keyphrases
- noisy data
- noise level
- learning algorithm
- random noise
- learning scheme
- learning process
- knowledge acquisition
- signal to noise ratio
- missing data
- noise sensitivity
- neural network
- image noise
- noise reduction
- unsupervised learning
- active learning
- optimal solution
- reinforcement learning
- worst case
- learning community
- training data
- minimum mean square error
- machine learning