Eric Siegel

Data Analytics - Bodily Professor

Bicentennial Bodily Professor

Delivering On the Promise of Machine Learning

Darden's inaugural Bodily Bicentennial Professor in Analytics, Eric Siegel, is a leading consultant and former Columbia University professor who focuses on machine learning. He is the founder of the Machine Learning Week conference series, executive editor of The Machine Learning Times and author of the bestselling book, Predictive Analytics: The Power to Predict Who Will Click, Buy, Lie, or Die.

When it comes to machine learning projects, there’s a great opportunity to improve. Currently, the overall track record of this highly visible technology is compromised: the realized value doesn’t match the excitement... yet. And that misalignment spells opportunity.

Machine learning’s popularity exploded because of its great potential business value and sheer scientific clout — learning from data to predict impresses the best of us. The praise is not all empty hype, but surveys show a stark reality: most machine learning projects fail to deploy. The number-crunching part is solid, but using it to improve operations turns out to be a greater change-management challenge than most foresee.

Those who proactively address this dilemma by adopting a collaborative, end-to-end management practice will be ahead of the curve. They will empower their organizations to overcome impediments so that the fruits of their data scientists’ labor come to fruition.

To right the misconceptions that sabotage machine learning and write the strategic playbook to launch it effectively, Eric has spent his one-year Darden appointment developing guidance and curricula to help future graduates run machine learning projects that successfully deploy. Here is an outline of his work:

Lectures. Several guest appearances at UVA courses and student clubs, including the School of Data Science courses "Business Analytics for Data Scientists" and "Big Data Ethics," and the Darden Tech Club. And a final presentation, "Machine Learning Is Notoriously Difficult to Deploy – But Here's How" (view the slide deck and the writeup in The Darden Report)

Curriculum supplement. Developed and "field tested" course materials to expand introductory data science courses so that they cover the business-side execution of machine learning projects – the known-how needed to ensure deployment is achieved. This supplement fulfills a critical unmet learner need.

Research. Interviewing industry leaders at companies like UPS and FICO as well as leading professors about machine learning deployment and helping run an expanded industry survey focused on the topic.

Writing. A series of nearly a dozen articles that define a framework for effective enterprise machine learning deployment, including:

The AI Hype Cycle Is Distracting Companies (Harvard Business Review)

To Deploy Machine Learning, You Must Manage Operational Change - Here Is How UPS Got It Right (Harvard Data Science Review)

How Machine Learning Can Improve the Customer Experience (Harvard Business Review)

Predictions: Seeing the Future (podcast interview: Trailblazers with Walter Isaacson)

To Avoid Wasting Money on Artificial Intelligence, Business Leaders Need More AI Acumen (Analytics Magazine, coauthored with Darden's Michael Albert)

"When AI Fails to Launch: Introduction to the Series on Machine Learning Leadership"

"Models Are Rarely Deployed: Machine Learning’s Industry-wide Dilemma"

"The Data Disconnect: A Key Challenge for Machine Learning Deployment"

This webpage will be updated as these and other articles are published.