What is machine learning?
Machine learning is the process that helps a computer learn without direct instruction. Machine learning uses algorithms to discover patterns in large datasets. Patterns are then used to create a comprehensive AI model, allowing for predictions with high accuracy that match your specific needs.2
During the creation of AI models, machine learning models are carefully trained to make predictions using smaller portions of data. The accuracy of the predictions depends on how well the smaller datasets represent the larger dataset. When the sample data is a good representation of the larger dataset (from which it is sampled), the machine learning model yields more reliable outcomes.2
A helpful analogy to the machine learning process is how an athlete learns the skills to play soccer. A soccer player starts by practicing basic skills like dribbling and passing. Through continuous practice, the soccer player begins to recognize playing patterns, anticipate the movements of opponents and teammates, and develop better coordination. Over time, the soccer player becomes more skilled at making split-second decisions during a match and adjusting strategies based on the evolving dynamics of the game.
Similarly, a machine learning model starts with learning patterns and relationships in data during the training process. The model then uses these learned patterns to make predictions or perform tasks that it was designed for. Just as the soccer player's expertise grows through practice and experiencing various game scenarios, the machine learning model's proficiency evolves through iterative learning and adaptation within the parameters set for the model.
2: See the Bibliography in this module's Summary section.