Book Image

Creators of Intelligence

By : Dr. Alex Antic
Book Image

Creators of Intelligence

By: Dr. Alex Antic

Overview of this book

A Gartner prediction in 2018 led to numerous articles stating that "85% of AI and machine learning projects fail to deliver.” Although it's unclear whether a mass extinction event occurred for AI implementations at the end of 2022, the question remains: how can I ensure that my project delivers value and doesn't become a statistic? The demand for data scientists has only grown since 2015, when they were dubbed the new “rock stars” of business. But how can you become a data science rock star? As a new senior data leader, how can you build and manage a productive team? And what is the path to becoming a chief data officer? Creators of Intelligence is a collection of in-depth, one-on-one interviews where Dr. Alex Antic, a recognized data science leader, explores the answers to these questions and more with some of the world's leading data science leaders and CDOs. Interviews with: Cortnie Abercrombie, Edward Santow, Kshira Saagar, Charles Martin, Petar Veličković, Kathleen Maley, Kirk Borne, Nikolaj Van Omme, Jason Tamara Widjaja, Jon Whittle, Althea Davis, Igor Halperin, Christina Stathopoulos, Angshuman Ghosh, Maria Milosavljevic, Dr. Meri Rosich, Dat Tran, and Stephane Doyen.
Table of Contents (23 chapters)
1
Chapter 1: Introducing the Creators of Intelligence

Balancing research and consulting

AA: Speaking of research more broadly, you’re an active researcher in this field and you’re a consultant, and I find it amazing that you can juggle the two. I’m keen to understand some of the current projects that you’re working on and what excites you the most. What areas of your expertise do you love to dabble in and produce actual products and solutions to?

CM: Well, the main thing I’ve been working on is something I call the WeightWatcher project, which is an AI model monitoring system. What we’re able to do is use techniques from theoretical physics and chemistry to analyze the performance of an AI model – such as a deep neural network – without looking at any of the data. I can take a model that’s been trained by somebody and feed it into my theory. I can look at some pretty pictures and some metrics, and tell you whether certain layers are overtrained and certain layers are undertrained...