Cloud computing vendors, such as Amazon, IBM, Google, Oracle, Microsoft, and others, not only provide storage and compute resources for applications, but data analytics services as well. These platforms support various dataset exploration and modeling tools, including traditional human-guided data mining techniques and automated machine learning algorithms.

Gupta and Sehgal’s Introduction to machine learning in the cloud with Python brings together the three title subjects for advanced students and for software developers who want to learn how to exploit these technologies. The authors are well qualified for the task. Pramod Gupta is a noted artificial intelligence (AI)/machine learning researcher at the University of California, Berkeley, and a principal data scientist for NovaSignal, a medical technology company providing AI-enabled healthcare analytics and services. Naresh Sehgal is the vice president of cloud engineering at NovaSignal and has authored several books on cloud computing and security.

The subtitle, concepts and practice, reflects the book’s two main parts, and the focus and priority of the subject matter. It begins with a comprehensive introduction to machine learning, suitable for those needing some context and background. Chapters in this section also cover the most commonly used machine learning algorithms, deep learning methods and architectures, and a brief review of cloud computing concepts. Part 2, “Practices,” gets to the meat of the subject, nicely detailing the critical topic of machine learning data preparation, a necessary chapter on cloud data security, and a fascinating chapter on machine learning case studies taken from the authors’ work at NovaSignal on cloud-based, AI/machine learning-enhanced brain imaging. While the medical terminology and methods might prove challenging for some readers, these cases illustrate the power of the discussed analysis techniques.

The inclusion of “Python” in the book’s title might imply a tutorial on implementing machine learning within cloud services. That content, however, is covered later in three fairly detailed appendices that include Python source code excerpts for some common use cases: sentiment prediction and letter image recognition. Source files are available on Sehgal’s GitHub site. Instructions for setting up the Python environments, program libraries, and datasets are presented for Amazon Web Services (AWS) and for Google Cloud.

Individual chapters include exercises called “Points to Ponder,” which provoke helpful questions about the content for readers. The authors include an appendix with suggested responses and an additional appendix with more detailed “Questions for Practice.” Considering the technical content, the authors’ prose and presentation style is conversational and easy to follow. There is a useful and extensive glossary at the book’s conclusion, which will also be helpful to readers new to the subject matter. Each chapter ends with a list of relevant references, some in standard format but many as bare uniform resource locators (URLs) which could stand some improvement.

Overall, the book represents a valuable resource for students and practitioners who need a well-organized and somewhat in-depth overview of how machine learning can be performed using existing cloud resources.

More reviews about this item: Amazon