AI Blueprints
上QQ阅读APP看书,第一时间看更新

Foreword

In retrospect, this book has been in the making for years. Stetson University, where I serve as an Assistant Professor of Computer Science, hosted its first annual hackathon in Spring 2016. The student organizers wanted faculty to give some lectures about programming and practical application development. AI was a hot topic then just as it is now, so I reached into my background in AI, including a PhD in the subject and multiple years of teaching an AI course in colleges, to develop a presentation I called AI/ML IRL, that is, artificial intelligence and machine learning in real life (AI/ML IRL, J. Eckroth, sudo HackStetson presentation, 2016, https://www2.stetson.edu/~jeckroth/downloads/eckroth-ai-irl-stetson-hackathon-2016.pdf). I covered current applications of AI, the fears and promises of AI, and concluded with advice about how to use AI in real-world applications. In this presentation may be found the seeds of the AI workflow developed in Chapter 1, The AI Workflow, and the discussion of the hype cycle in Chapter 8, Preparing for Your Future and Surviving the Hype Cycle.

Around the same time, my colleague and CEO at i2k Connect, Dr. Reid Smith, was awarded the Robert S. Engelmore Memorial Award, sponsored by the Innovative Applications in Artificial Intelligence conference and AI Magazine. Dr. Smith gave a presentation for this award (A Quarter Century of AI Applications: What we knew then vs. what we know now, R. G. Smith, Robert S. Engelmore Memorial Lecture Award, presented at the Twenty-Eighth Conference on Innovative Applications of Artificial Intelligence (IAAI-16), Phoenix, AZ, 15 February, 2016, http://www.reidgsmith.com/2016-02-15_Engelmore_Lecture.pdf), where he discussed numerous examples of successful AI applications and the lessons learned.

Dr. Smith and I discussed our separate contributions about the topic of programming AI applications and came together to write the cover article of the Spring 2017 AI Magazine (Building AI Applications: Yesterday, Today, and Tomorrow, R. G. Smith and J. Eckroth, AI Magazine 38(1): 6-22, Spring 2017, https://www.aaai.org/ojs/index.php/aimagazine/article/view/2709). This article examined a series of important deployed applications that made significant use of AI and machine learning. We also showed the increasing interest in AI, which I have updated for Chapter 5, A Blueprint for Detecting Your Logo in Social Media when I discuss deep learning. Most importantly, this article introduced some of the essential features of the AI workflow, including some checklist items to pay attention to when developing your own applications.

Jumping back to 2014 momentarily, Frank Pohlmann contacted me to write a book for another publisher. I agreed, and we developed an outline, but as I was just starting at Stetson University, I was swamped with other activities and had to cancel the plan. Three years later, Mr. Pohlmann was now a Managing Acquisition Editor for Packt and contacted me again. All of the developments I described above had occurred in the intervening time and I had more practice managing my time as a professor. The timing was right.

This book is different than any of those prior works. It focuses on programming realistic and useful AI applications with state-of-the-art software libraries and techniques. It also teaches the reader the fundamentals of the techniques we use throughout the book. But this book is the spiritual successor, and the culmination, of several years of thinking, writing, and speaking.

AI Blueprints could not have been written without the encouragement, insights, and close editorial attention from Frank Pohlmann, Reid Smith, and the staff at Packt. Another colleague at i2k Connect, Dr. Eric Schoen, graciously served as the technical reviewer of this work.

His decades of software engineering experience, including more than 30 years at Schlumberger and most recently as their Chief Software Architect, as well as a PhD in AI from Stanford University, helped refine the technical sophistication of the examples and explanations in this book.

As I'm sure every reader knows at some fundamental level, the time is right for AI. It has been for some time and will be for the foreseeable future. This book is designed to help you capture a bit of the magic of AI in your own applications.