Calendar

🚧 Work in progress. This site is online so you can get a feel for the course, but I'm still finalizing the details — the schedule, papers, and policies may still change. Check back for updates!

Part I · Instructor lectures

Mon Sep 14 · Distances between distributions

Course intro & logistics. Integral probability metrics, Maximum Mean Discrepancy, and the Wasserstein-1 distance.

Released Welcome assignment

Background reading:

• A Kernel Two-Sample Test (Gretton et al., 2012)
• Generative Moment Matching Networks (Li et al., 2015)

Mon Sep 21 · Coupling distributions

Monge vs. Kantorovich, entropic regularization & Sinkhorn, and Brenier’s theorem.

Due Welcome assignment

Background reading:

• Sinkhorn Distances (Cuturi, 2013)
• Interpolating between Optimal Transport and MMD using Sinkhorn Divergences (Feydy et al., 2019)

Mon Sep 28 · Interpolating between distributions

Dynamic OT / Benamou–Brenier, Wasserstein barycenters, and continuous normalizing flows → flow matching.

Background reading:

• Neural Ordinary Differential Equations (Chen et al., 2018)
• Flow Matching for Generative Modeling (Lipman et al., 2023)

---

Part II · Reading group

Papers in this part are presented by students in the reading group. See Role-play Seminar for how the sessions are organized, the roles, and expectations.

Mon Oct 5 · OT for domain adaptation

Papers to present:

• Optimal Transport for Domain Adaptation (Courty et al., 2017)
• DeepJDOT: Deep Joint Distribution Optimal Transport for Unsupervised Domain Adaptation (Damodaran et al., 2018)

Wed Oct 14 · Distributions as objects

Legislative Day — Monday schedule.

Papers to present:

• Wasserstein Wormhole: Scalable Optimal Transport Distance with Transformers (Haviv et al., 2024)
• Generative Distribution Embeddings: Lifting Autoencoders to the Space of Distributions (Fishman et al., 2025)

Mon Oct 19 · Supervised Monge maps

Papers to present:

• Monge Gap (Uscidda & Cuturi, 2023)
• Conditional Monge Gap (Driessen et al., 2025)

Mon Oct 26 · Unbalanced OT

Due Project — preliminary proposal

Papers to present:

• Scaling Algorithms for Unbalanced Optimal Transport (Chizat et al., 2018)
• Unbalancedness in Neural Monge Maps (Eyring et al., 2024)

Mon Nov 2 · Gromov–Wasserstein

Papers to present:

• Gromov–Wasserstein Averaging of Kernel and Distance Matrices (Peyré et al., 2016)
• Gromov–Wasserstein Alignment of Word Embedding Spaces (Alvarez-Melis & Jaakkola, 2018)

Mon Nov 9 · OT couplings inside flow matching

Papers to present:

• OT-CFM: Conditional Flow Matching with minibatch OT (Tong et al., 2023)
• MMFM: Flow Matching across time and conditions (Rohbeck et al., 2025)

Mon Nov 16 · Flow matching on structured domains

Papers to present:

• Riemannian Flow Matching on General Geometries (Chen & Lipman, 2023)
• Discrete Flow Matching (Gat et al., 2024)

Mon Nov 23 · No class

Thanksgiving week.

Mon Nov 30 · Few-step flow matching & guidance

Papers to present:

• Flow Map Matching (Boffi et al., 2024)
• Mean Flow (Geng et al., 2025)
• Guided Flows for Generative Modeling and Decision Making (Zheng et al., 2023)
• CFG-Zero*: Improved Classifier-Free Guidance (Fan et al., 2025)

Mon Dec 7 · Applications: single-cell & music

Papers to present:

• Optimal-Transport Analysis of Single-Cell Gene Expression / Waddington-OT (Schiebinger et al., 2019)
• GENOT: Generative Entropic Neural Optimal Transport (Klein et al., 2024)
• Music generation — papers TBD

Mon Dec 14 · Final project presentations

Due Final project — report & presentation

Last day of classes (possibly remote). See Project Logistics.