Machine Learning
Techniques
Papers
Definitions
ML refers to the field that describes techniques and algorithms that give computers the ability to learn without being explicitly programmed.
The goal is to be able to generalise, that is to make accurate predictions on new, unseen data.
We train the model to find the best values for the trainable parameters of the chosen model for the given task and dataset. This is achieved by minimising the chosen loss function.
Supervised ML relies on labelled input and output data to learn and make predictions, while unsupervised learning does not.
Machine learning problems should be well-defined into their elements.
Training
When a model is trained too well on training data, it performs poorly on new, unseen data, with a phenomenon called overfitting.
When a model is not capable enough of capturing underlying pattern, it is called underfitting.
The ultimate goal of ML models is to be able to generalise.
Identifying overfitting or underfitting can be done.
Train and test split is used to both train the model and confirm that it can generalise and make accurate predictions on new data.
Regularisation are techniques to prevent models from overfitting.
Hyperparameters are manually decided parameters.