NeuralBlitz What I do
- Build Bayesian neural networks
- Quantify uncertainty in DL
- Apply variational inference
- Use Monte Carlo dropout
When to use me
When working on uncertainty quantification in deep learning.
Key Concepts
- Bayesian NNs
- Variational inference
- Dropout as Bayesian
- Weight uncertainty
- Epistemic uncertainty
- Aleatoric uncertainty
- Model calibration