Foster Provost

Bio

Foster Provost is Ira Rennert Professor of Entrepreneurship and Information Systems and Director, Fubon Center, Data Analytics & AI, at the Stern School of Business at New York University. He is Professor of Data Science at NYU and former interim Director of the NYU's Center for Data Science.  He is also a Distinguished Scientist for real-estate giant Compass. He previously was Editor-in-Chief of the journal Machine Learning and was elected as a founding board member of the International Machine Learning Society.  Foster stands out in data science for having made substantial contributions across research, practical applications, and business thought leadership.

Foster's research on data science and machine learning has won many awards, including (among others) the 2020 ACM SIGKDD Test of Time Award, the 2017 European Research Paper of the Year, Best Paper awards in the top research venues across four decades, the 2009 INFORMS Design Science Award from the top professional society for operations research, IBM Faculty Awards, and a President’s Award from NYNEX Science and Technology (now Verizon).

For more than 25 years, Prof. Provost has helped leaders in business and government understand how data science, artificial intelligence, and machine learning technologies can add value. His book Data Science for Business is required reading in many of the top business schools, and was listed as one of Fortune Magazine's "must read books for MBAs." He has designed AI/machine learning systems for some of the largest companies in the world and worked with the DoD on the application of AI/machine learning to counter-terrorism.

Foster also has had substantial experience helping to found startups. He was the founding chief scientist for adtech data science powerhouse Dstillery, designing the original machine learning algorithms and building the founding data science teams for both Media6Degrees and Everyscreen Media (which merged to form Dstillery). He also was a coFounder of Detectica (acquired by Compass), Belgium's Predicube (acquired by Var), and most notably, "baby unicorn" Integral Ad Science, which was acquired by Vista Equity Partners for a reported $850M.

Talks

• Causal Targeting - It's not Causal Effect Estimation (and Why it Matters)

  Keynote
  10-06-2021 - 14:00-15:00
  Abstract
  

  Causal decision making (CDM) at scale has become a routine part of business, and increasingly CDM is based on statistical models and machine learning algorithms. For example, businesses target offers, incentives, and recommendations algorithmically with the goal of affecting consumer behavior.  In this talk I highlight something important: targeting for causal effect is not the same as causal effect estimation.  In fact, accurate causal effect estimation is not necessary for accurate CDM.  I will discuss three implications of this distinction.  (1) We should consider carefully the objective function of causal machine learning, and if possible we should optimize for accurate causal targeting rather than for accurate effect-size estimation.  (2) For causal targeting it may be just as good or even better to learn with confounded data as with unconfounded data, because confounding has different effects on the two.  Finally, (3) causal statistical modeling may not be necessary at all to support CDM, because there may be (and perhaps often is) a proxy target for statistical modeling that can do as well or better.  This observation helps to explain at least one broad common CDM practice that seems ``wrong" at first blush---the widespread use of non-causal models for targeting interventions like advertisements and retention incentives.  The last two implications are particularly important in practice, as acquiring (unconfounded) data on both ``sides" of the counterfactual for modeling can be quite costly and often is impracticable.  Understanding causal targeting is vital to modern data science practice, and is fertile ground for new data science research (there's been surprisingly little until just the past few years).