Machine Learning is defined as a specific subset of AI, representing an approach to building AI applications where models are trained to make predictions. Unlike the broad scope of AI, ML has a narrower goal, focused on training machines to autonomously learn patterns and insights from data to improve performance or increase prediction accuracy over time.

ML tasks are generally categorized into supervised learning (when labeled data is available) or unsupervised learning (when the goal is to uncover patterns without labels). These foundational tasks further break down into specific functions like classification, regression, or clustering. The process relies on algorithms for these learning types to create predictive models.