Workshop III: Geometry of Big Data

Part of the Long Program Geometry and Learning from Data in 3D and Beyond
April 29 - May 3, 2019

Schedule

All times in this Schedule are Pacific Time (PT)

Monday, April 29, 2019

Morning Session
8:00 - 8:55 Breakfast (hosted by IPAM)
8:55 - 9:00 Welcome & Opening Remarks: Dean Miguel García-Garibay (Dean of Physical Sciences, UCLA) and Dima Shlyakhtenko (Director, IPAM)
9:00 - 9:40
Pradeep Ravikumar (Carnegie Mellon University)

DAGs with NO TEARS: Continuous Optimization for Structure Learning
PDF Presentation

 
9:50 - 10:05 Break
10:05 - 10:45
Ronen Talmon (Technion - Israel Institute of Technology)

Data Analysis with the Riemannian Geometry of Symmetric Positive-Definite Matrices
PDF Presentation

 
10:55 - 11:10 Break
11:10 - 11:50
12:00 - 1:30 Lunch (on your own)

Afternoon Session
1:30 - 2:10
Dejan Slepcev (Carnegie Mellon University)

Proper regularizers for semi-supervised learning
PDF Presentation
2:20 - 2:35 Break
2:35 - 3:15
3:25 - 3:40 Break
3:40 - 4:20
Lorenzo Rosasco (Università di Genova)

A consistent framework for structure machine learning
PDF Presentation

 
4:30 - 6:00 Poster Session & Reception (Hosted by IPAM)

Tuesday, April 30, 2019

Morning Session
8:00 - 9:00 Breakfast (hosted by IPAM)
9:00 - 9:40
Frederic Chazal (Institut National de Recherche en Informatique et Automatique (INRIA))

On the density of expected persistence diagrams and its kernel based estimation
PDF Presentation

 
9:50 - 10:05 Break
10:05 - 10:45
 
10:55 - 11:10 Break
11:10 - 11:50
12:00 - 1:45 Lunch (on your own)

Afternoon Session
1:45 - 2:25
Andrea Montanari (Stanford University)

Some geometric surprises in modern machine learning

 
2:35 - 2:50 Break
2:50 - 3:30
3:40 - 3:55 Break
3:55 - 4:35
Rebecca Willett (University of Chicago)

Learning to Solve Inverse Problems in Imaging
PDF Presentation

 

Wednesday, May 1, 2019

Morning Session
8:00 - 9:00 Breakfast (hosted by IPAM)
9:00 - 9:40
9:50 - 10:05 Break
10:05 - 10:45
Marina Meila (University of Washington)

Manifold Learning for the Sciences
PDF Presentation

 
10:55 - 11:10 Break
11:10 - 11:50
Mahdi Soltanolkotabi (University of Southern California (USC))

Towards demystifying over-parameterization in deep learning

 
12:00 - 1:45 Lunch (on your own)

Afternoon Session
1:45 - 2:30
Stanley Osher (University of California, Los Angeles (UCLA))

PDE Approaches for Deep Learning
PDF Presentation

 
2:40 - 2:55 Break
2:55 - 3:35
Imre Risi Kondor (University of Chicago)

Covariant neural networks for learning physical systems
PDF Presentation

 
3:45 - 4:00 Break
4:00 - 4:40

Thursday, May 2, 2019

Morning Session
8:00 - 9:00 Breakfast (hosted by IPAM)
9:00 - 9:40
9:50 - 10:05 Break
10:05 - 10:45
Bin Yu (University of California, Berkeley (UC Berkeley))

PCS workflow, interpretable machine learning, and DeepTune

 
10:55 - 11:10 Break
11:10 - 11:50
12:00 - 1:45 Lunch (on your own)

Afternoon Session
1:45 - 2:25
Zuowei Shen (National University of Singapore)

Deep Approximation via Deep Learning
PDF Presentation

 
2:35 - 2:50 Break
2:50 - 3:30
Joel Tropp (California Institute of Technology)

SketchySVD
PDF Presentation

 
3:40 - 3:55 Break
3:55 - 4:35
Jianfeng Cai (Hong Kong University of Science and Technology)

Solving Systems of Phaseless Equations: Riemannian Optimization and Global Geometry

 

Friday, May 3, 2019

Morning Session
8:00 - 9:00 Breakfast (hosted by IPAM)
9:00 - 9:40
9:50 - 10:05 Break
10:05 - 10:45
 
10:55 - 11:10 Break
11:10 - 11:50
12:00 Conclusion