• Graduate program
    • Why Tinbergen Institute?
    • Research Master
    • Admissions
    • Course Registration
    • Facilities
    • PhD Vacancies
    • Selected PhD Placements
    • Research Master Business Data Science
  • Research
  • Browse our Courses
  • Events
    • Summer School
      • Applied Public Policy Evaluation
      • Deep Learning
      • Economics of Blockchain and Digital Currencies
      • Economics of Climate Change
      • Foundations of Machine Learning with Applications in Python
      • From Preference to Choice: The Economic Theory of Decision-Making
      • Gender in Society
      • Machine Learning for Business
      • Marketing Research with Purpose
      • Sustainable Finance
      • Tuition Fees and Payment
      • Business Data Science Summer School Program
    • Events Calendar
    • Events Archive
    • Tinbergen Institute Lectures
    • 16th Tinbergen Institute Annual Conference
    • Annual Tinbergen Institute Conference
  • News
  • Alumni
Home | Events Archive | From Explainable AI to Generative Modeling with Tree-Based Machine Learning
Seminar

From Explainable AI to Generative Modeling with Tree-Based Machine Learning


  • Series
  • Speaker(s)
    Marvin N. Wright (Leibniz Institute for Prevention Research and Epidemiology and University of Bremen, Germany)
  • Field
    Econometrics, Operations Analytics, Data Science and Econometrics
  • Location
    Erasmus University Rotterdam, Campus Woudestein, Langeveld 1.16
    Rotterdam
  • Date and time

    May 15, 2025
    12:00 - 13:00

Abstract

Despite the success of deep learning, tree-based machine learning is still competitive in many application domains. Recent literature shows that, e.g. on tabular data, tree-based methods still outperform neural networks, while being faster, easier to apply and requiring less tuning. However, this research focuses exclusively on discriminative models and their prediction performance. In this talk, I present two recent tree-based methods that go beyond predictive performance. First, an explanation method that provides a global representation of a prediction function by decomposing it into the sum of main and interaction components of arbitrary order. Our method extends Shapley values to higher-order interactions and is applicable to tree-based methods that consist of ensembles of low dimensional structures such as gradient-boosted trees. Second, I present adversarial random forests, a provably consistent method for density estimation and generative modeling. With the new method, we achieve comparable or superior performance to state-of-the-art deep learning models on various tabular data benchmarks while executing about two orders of magnitude faster.