Machine learning (ML) encompasses a myriad of technical, business, social, and scientific methods and disciplines. Project managers and technology experts are faced with an ever-expanding field of new concepts, vocabulary, algorithms, and even legal and ethical issues. No single practitioner can claim complete expertise in this extensive area of knowledge, but should at least be literate in the terminology and tools of the ML trade. Thus, a dictionary or encyclopedia for this collection of practices is needed for those who manage large information technology (IT) projects that incorporate ML, in order to make informed decisions about such usage.
Kashyap, an expert in technology management, has provided an extensive guide to the world of ML. His book is not a technical how-to reference, but a reasonably comprehensive survey of the concepts and methods needed to understand the field and to guide decision makers.
The opening chapters begin with definitions of ML and related technologies, including supervised and unsupervised learning, big data analytics, cloud and Internet of Things (IoT) services and architecture, and artificial intelligence (AI).
To support readers, the author includes concept maps, case studies, best practice discussions, and jargon translations, along with relevant presentation templates for the chapters available on the publisher’s website.
There are quite readable sections on ML algorithms such as clustering and regression, along with recommendations for how to select appropriate models and programming tools. The author also provides a valuable chapter on future applications of ML, including some interesting speculation on its use in cognitive augmentation.
One significant omission is a thorough discussion of the ethics of ML. There are well-documented abuses of this technology (see O’Neil’s Weapons of math destruction, for example [1,2]), which should be included in any discussion of these powerful algorithms, along with best practices to avoid biases in ML results.
Nevertheless, Kashyap’s book does provide a thorough and understandable compendium of the current ML universe for those who manage and direct ML-enabled projects.