Robert Munro
MEAP began July 2019 Publication in April 2020 (estimated)
ISBN 9781617296741 325 pages (estimated) printed in black & white
I've learned a lot of new things about Machine Learning I would never even have considered before
.
——Michiel Trimpe
Most machine learning systems worldwide learn from human feedback. However, most machine learning courses only focus on algorithms, not the human-computer interaction part of the system. This creates a huge knowledge gap when data science is applied to machine learning tasks in the real world. Data science spends more time on data management than algorithm construction. Human-in-the-loop machine learning is a practical guide for optimizing the machine learning process. It consists of annotation technology, active learning, transfer learning, and how to use machine learning to transform each step.
about the technology
"Man-in-the-loop machine learning" is a written application for the interaction requirements between humans and machine learning systems to improve the performance of humans or machines, or both. Most machine learning projects are not about the time or cost of human input for all data, so strategies are needed to determine important data for human inspection. The continuous participation of humans with correct interfaces accelerates the efficiency of marking complex or novel data that machines cannot handle, and reduces the possibility of data-related errors.
about the book
"Man-in-the-loop Machine Learning" is a guide for optimizing the human and machine parts of a machine learning system to ensure that the data and models are correct, relevant, and cost-effective. Robert Munro, a 20-year machine learning veteran, has proposed strategies for allowing machines and humans to work together efficiently, including building a reliable user interface for data annotation, active learning strategies for sampling human feedback, and transfer learning. After completing this book, you will be able to design a machine learning system that independently selects the correct data for human inspection to ensure the correctness and usefulness of those annotations.
what’s inside
- Active learning: sampling correct data and labeling it to humans;
- Marking strategy: Provide the best interface for human feedback;
- Supervised machine learning design and query strategies to support human-in-the-loop systems;
- Advanced adaptive learning methods: use machine learning to optimize every step of the human loop;
- Examples of data science applications in the real world where people are in the loop.
About the author
Robert Munro builds annotation, active learning and machine learning systems with startups focusing on machine learning and large companies including Amazon, Google, IBM and most major mobile phone manufacturers. If you communicate with your mobile phone by voice, your car parks autonomously, your music app adapts to your taste, or automatically recommends news for you, then Robert Munro is likely to contribute to these.
Robert Munro has a Ph.D. from Stanford University, focusing on human-in-the-loop machine learning for medical and disaster response. It is not only a machine learning expert, but also a disaster response expert. He has helped categorize disaster-related data from real disasters in the past.