Algebraic Machine Learning
- (Stanford University - Jaclyn Chen)
- Overview
Algebraic Machine Learning (AML) is a new mathematical approach to Artificial Intelligence (AI) that combines user-defined and self-generated symbols. The generated symbols provide internal representations that allow AML to learn from data and adapt to the world like neural networks.
AML is a new learning paradigm that builds on Abstract Algebra and Model Theory. Machine learning is all about math, which helps create algorithms that can learn from data to make accurate predictions.
Some topics that are important for ML math include: descriptive statistics, hypothesis testing, regression analysis, probability distributions, conditional probability, sampling and central limit theorem, bayes theorem.
AML is a new AI paradigm that combines user-defined symbols with self-generated symbols, enabling AML to learn from data and adapt to the world like a neural network, combined with the interpretability of Symbolic AI.
AML is a purely symbolic method that neither uses neurons nor is it a neurosymbolic method. Algebraic machine learning does not use parameters and does not rely on fitting, regression, backtracking, constraint satisfiability, logical rules, production rules, or error minimization.
[More to come ...]
