How to contributeΒΆ
Please contribute by opening a pull request at https://github.com/cjratcliff/ml-compiled.
List of sections that need expanding in rough priority order (high at the top):
- Transformer
- Deconvolution layer
- Lambda loss
- Wasserstein distance
- Rademacher complexity
- VC dimension
- Evidence lower-bound
- Normalizing flows
- Change of variables
- ZCA
- Gradient boosting
- Softmax bottleneck
- CRFs
- Codomain
- Continuous
- Image
- SVMs - dual form, primal form and training
- Matrix differentiation
- Weibull distribution
- Vanishing/exploding gradient problem
- Feature scaling
- Object tracking
- Positive-unlabeled learning
- Logit
- Question answering
- Translation
- Manifold hypothesis
- Variational Bayes
- Parametric and non-parametric models
- Discriminative model
- Energy-based model
- Expectation-maximisation (EM)
- Inductive bias
- Metric learning
- Overcomplete representation
- Sparsity
- Advantage function
- Counterfactual regret minimization
- Direct policy search
- Information set
- Regret matching
- Reward clipping
- Reward sparsity
- Value iteration
- Mean field approach
- Partition function
List of sections to be added:
- Imitation learning
- Catastrophic forgetting
- Trust region policy optimization
- Self-attention
- Lipschitz smoothness
- Lipschitz constant
- Canonical Correlation Analysis (CCA)
- Compositionality
- Bayesian neural networks
- Bayesian optimisation
- Deep Belief Networks/Machines
- Restricted Boltzmann Machines
- Regression - p-values
- Empirical risk