Subject to: Dorit Hochbaum

2025-11-10 · Source: Subject to · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Data Science & Analytics, Software Development & Engineering · Depth: Advanced, extended

Summary

Dorit Hochbaum, a distinguished professor at UC Berkeley, shares her extensive career in industrial engineering and operations research, highlighting her contributions to discrete optimization, network flow techniques, and image segmentation. The interview covers her early life in Tel Aviv, her parents' World War II experiences, and her mandatory military service in Israel. Hochbaum recounts her academic journey from Tel Aviv University to the University of Pennsylvania and Carnegie Mellon, where she faced significant misogyny. She details her groundbreaking research, including work on the K-center problem with David Shmoys, monotone integer programs with two and three variables per inequality, and the pseudoflow algorithm for max flow min cut problems, which found significant application in open-pit mining. The discussion also touches upon her research in neuroscience and the application of Markov Random Fields to diverse problems like NSF project evaluation and semiconductor yield prediction. She emphasizes the importance of serendipity and persistence in research.

Key takeaway

For Machine Learning Engineers and Research Scientists developing optimization solutions, consider re-framing complex problems into monotone integer programs. Your ability to identify and apply these structures, particularly with two or three variables per inequality, can lead to polynomial-time algorithms and significantly faster, higher-quality solutions compared to spectral or traditional machine learning methods, as demonstrated in image segmentation and resource allocation.

Key insights

Serendipity and persistence are crucial for groundbreaking research and overcoming academic challenges.

Principles

• Optimal solutions often arise from unexpected connections.
• Theoretical insights can drive practical, real-time applications.

Method

Formulate problems as monotone integer programs with two or three variables per inequality to leverage efficient minimum cut algorithms for polynomial-time solutions, even for complex applications like image segmentation.

In practice

• Explore parametric cut procedures for efficient optimization.
• Apply Markov Random Fields for group decision-making analysis.

Topics

• Approximation Algorithms
• Network Flow Algorithms
• Monotone Integer Programs
• Image Segmentation
• Markov Random Fields

Best for: AI Scientist, Research Scientist, Machine Learning Engineer

Related on AIssential

• Geometrical fairness in graph neural networks — A new graph diffusion method modifies the Laplacian operator to mitigate bias and improve fairness.
• Largest empty shapes — Finding largest empty shapes in point sets requires diverse optimization models depending on the shape.
• Subject to: Bissan Ghaddar — Resilience and adaptability are crucial for navigating complex academic and professional paths, especially in optimization.
• Selecting Samples on Graphs: A Unified Dataset Pruning Framework for Lossless Training Acceleration — A unified graph-based framework prunes datasets by modeling samples as nodes and relations as edges, solving a Maximum Weight Clique Problem.
• Polynomial Dice Loss for Medical Image Segmentation — Polynomial Dice Loss enhances medical image segmentation by controlling loss shape via Taylor expansion for improved performance.
• Have I Solved This Before? Retrieving Similar Segmentation Problems for Evolutionary Learning — Storing abstract system models enables retrieving and refining existing solutions for new segmentation problems.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Subject to.