Reliable Uncertainty Estimates in Deep Neural Networks using Noise Contrastive Priors.
Danijar HafnerDustin TranAlex IrpanTimothy P. LillicrapJames DavidsonPublished in: CoRR (2018)
Keyphrases
- neural network
- measurement error
- sensor noise
- measurement noise
- learned from training data
- measurement errors
- pattern recognition
- noise level
- noisy data
- neural network model
- estimation error
- gradient estimation
- belief nets
- genetic algorithm
- fuzzy logic
- uncertain data
- confidence intervals
- noise model
- image processing
- artificial neural networks
- fault diagnosis
- additive noise
- cost effective
- feed forward
- parametric models
- decision theory
- fuzzy systems
- signal to noise ratio
- noisy environments
- noise free
- robust optimization
- prior model
- missing data
- bayesian framework
- prior knowledge
- bayesian approaches
- associative memory
- noise reduction