Tribes
The five machine-learning paradigms described by Pedro Domingos in The Master Algorithm:
| Tribe | Origins | Problem | Solution |
|---|---|---|---|
| Symbolists | Logic, philosophy | Knowledge composition | Induction |
| Connectionists | Neuroscience | Credit assignment | Backpropagation |
| Evolutionaries | Evolutionary biology | Structure discovery | Genetic programming |
| Bayesians | Statistics | Uncertainty | Probabilistic inference |
| Analogizers | Psychology | Similarity | Kernel machines |
Symbolists
TODO
Connectionists
TODO
Evolutionaries
TODO
Bayesians
TODO
Analogizers
TODO
Essays
- Ways to think about machine learning (2018) by Ben Evans
Books
Courses
- Mathematics for Machine Learning at Coursera
Glossary
- parameter: a value that the algorithm itself mutates over time.
- hyperparameter: a value specified prior to learning, e.g., the learning rate.
News
See Also
- Arto's Notes re: artificial intelligence (AI), deep learning (DL), neural networks (NNs), Markov logic network (MLN), fast.ai, NuPIC, OpenAI, TensorFlow