Prof. Dimitris J Bertsimas

Boeing Leaders for Global Operations Professor of Management
Professor of Operations Research

Primary DLC

MIT Sloan School of Management

MIT Room: E62-560

Assistant

Shalane Hutchins
shalane@mit.edu

Areas of Interest and Expertise

Air Safety
eCommerce
Financial Engineering
Information Technology
Operations Research
Optimization
Revenue Management
Statistics

Research Summary

While much of Bertsimas' work is mathematical in nature, he is also keenly interested in operations research applications in a variety of contexts, including finance, revenue management, supply chain management, and transportation. He is working on optimising large portfolios, the revenue streams of airlines, and the operations at Panama Canal.

Recent Work

  • Video

    Dimitris Bertsimas - 2019 Citi-NY

    November 6, 2019Conference Video Duration: 42:49

    Interpretable AI and its Applications in Financial Services

    This talk introduces a new generation of machine learning methods that provide state of the art performance and are very interpretable. Optimal classification (OCT) and regression (ORT) trees are introduced for prediction and prescription with and without hyperplanes. It will be shown that (a) Trees are very interpretable, (b) They can be calculated in large scale in practical times, and (c) In a large collection of real world data sets, they give comparable or better performance than random forests or boosted trees. Their prescriptive counterparts have a significant edge on interpretability and comparable or better performance than causal forests. These optimal trees with hyperplanes have at least as much modeling power as (feedforward, convolutional and recurrent) neural networks and comparable performance in a variety of real world data sets. Finally, a variety of optimal trees applications in financial services will be discussed.

    2019 MIT Citi Conference in NYC

    AI in LIfe Science 2018 - Dimitris Bertsimas

    December 4, 2018Conference Video Duration: 28:23

    Interpretable AI

    This talk introduces a new generation of machine learning methods that provide state of the art performance and are very interpretable, introducing optimal classification (OCT) and regression (ORT) trees for prediction and prescription with and without hyperplanes. This talk shows that (a) Trees are very interpretable, (b) They can be calculated in large scale in practical times, and (c) In a large collection of real world data sets, they give comparable or better performance than random forests or boosted trees. Their prescriptive counterparts have a significant edge on interpretability and comparable or better performance than causal forests. Finally, we show that optimal trees with hyperplanes have at least as much modeling power as (feedforward, convolutional, and recurrent) neural networks and comparable performance in a variety of real world data sets. These results suggest that optimal trees are interpretable, practical to compute in large scale, and provide state of the art performance compared to black box methods.

    2018 MIT AI in Life Sciences and Healthcare Conference

    Dimitris Bertsimas

    October 18, 2016MIT Faculty Feature Duration: 27:28

    MIT