Chapter 1, Introducing Active Machine Learning, explores the fundamental principles of active machine learning, a highly effective approach that significantly differs from passive methods. This chapter also offers insights into its distinctive strategies and advantages.
Chapter 2, Designing Query Strategy Frameworks, presents a comprehensive exploration of the most effective and widely utilized query strategy frameworks in active machine learning and covers uncertainty sampling, query-by-committee, expected model change, expected error reduction, and density-weighted methods.
Chapter 3, Managing the Human in the Loop, discusses the best practices and techniques for the design of interactive active machine learning systems, with an emphasis on optimizing human-in-the-loop labeling. Aspects such as labeling interface design, the crafting of effective workflows, strategies for resolving model-label disagreements, the selection of suitable labelers, and their efficient management are covered.
Chapter 4, Applying Active Learning to Computer Vision, covers various techniques for harnessing the power of active machine learning to enhance computer vision model performance in tasks such as image classification, object detection, and semantic segmentation, also addressing the challenges in their application.
Chapter 5, Leveraging Active Learning for Big Data, explores the active machine learning techniques for managing big data such as videos, and acknowledges the challenges in developing video analysis models due to their large size and frequent data duplication based on frames-per-second rates, with a demonstration of an active machine learning method for selecting the most informative frames for labeling.
Chapter 6, Evaluating and Enhancing Efficiency, details the evaluation of active machine learning systems, encompassing metrics, automation, efficient labeling, testing, monitoring, and stopping criteria, aiming for accurate evaluations and insights into system efficiency, guiding informed improvements in the field.
Chapter 7, Utilizing Tools and Packages for Active ML, discusses the Python libraries, frameworks, and tools commonly used for active learning, highlighting their value in implementing various active learning techniques and offering an overview suitable for both beginners and experienced programmers.
