Home

Transport-methods in signal processing & machine learning

Weekly plan

Week 1: introduction
    • Transport methods and engineering applications (8/25), class instructions slides, overview slides, reading
    • Linear transforms (8/27), reading1, reading2 (sections 2.1, 2.2, 2.3)
Week 2: introduction (cont.)
Week 3: One dimensional optimal transport and the cumulative distribution transform
    • Homework 1 Discussion (9/8)
    • 1D transport: cumulative distribution transform and Wasserstein distance slidesadditional notes (9/10)
Week 4: Applications
Week 5:  CDT and Convexity, introduction to the Radon-CDT
    • Convexity properties discussion(9/22)
    • Radon transform notes, Radon-CDT notes  (9/24)
Week 6:  R-CDT and Convexity, Sliced Wasserstein distance
    • Sliced Wasserstein distance, convexity properties of R-CDT(9/29) notes
    • Homework 2 discussion (10/1)
Week 7: N-D optimal transport, continuous and discrete formulation
Week 8: Mid semester 
    • Linear optimal transport (LOT) notes, discrete transport slides (10/13)
    • Project presentations (10/15)
Week 9: Linear optimal transport (LOT)
    • Project presentation part 2 (10/20)
    • Discrete LOT (10/22)
Week 10: Measure theory formulation
    • Measure and integration and wrap-up of OT theory slides, annotated version(10/27)
    • Measure and integration discussion (10/29)
Week 11: Applications: Signal processing
Week 12: Applications: Machine learning
    • Gaussian mixture model estimation (11/10), materials (slides, data, python code)
    • Statistical generative models (11/12), slides, code
Week 13: Project presentations
    • project presentations (11/17)
    • project presentations (11/19)
Week 14: Numerical methods
    • numerical methods, slides (11/24)
Ċ
Gustavo Rohde,
Aug 13, 2020, 4:57 PM