Part 1: Fundamentals of Active Machine Learning
In the rapidly evolving landscape of machine learning (ML), the concept of active ML has emerged as a transformative approach that optimizes the learning process by selectively querying the most informative data points from unlabeled datasets. This part of the book is dedicated to laying the foundational principles, strategies such as uncertainty sampling, query-by-committee, expected model change, expected error reduction, and density-weighted methods, and considerations essential for understanding and implementing active ML effectively. Through a structured exploration, we aim to equip readers with a solid grounding of the best practices for managing the human in the loop by exploring labeling interface design, effective workflows, strategies for handling model-label disagreements, finding adequate labelers, and managing them efficiently.
This part includes the following chapters:
