Another achievement of ML is the Support Vector Machine (SVM), and I'd say that it represents better the tradeoff between space and computation. In fact, SVMs

Another achievement of ML is the Support Vector Machine (SVM), and I'd say that it represents better the tradeoff between space and computation. In fact, SVMs literally make that tradeoff by using the Kernel trick to save up on the computational costs of projecting to higher, possibly infinite dimensions. As we know, projecting up to higher dimensions can provide us with a 100% accuracy on the training data, but of course, that's overfitting and we use constrains/support vectors to regularize the learner.