Workshop on Inverse Problems and Machine LearningMay 27-29, 2019Program |
|||
|
||||||||||
Monday, May 27, 2019 |
||||||||
|
|
||||||||
|
08:30 - 09:00
|
Registration (Room 5345) and Coffee & Croissants (Room 6245)
|
|
|
|
09:00 - 09:40
|
Mikhail (Misha) Belkin
(The Ohio State University) Rethinking the bias-variance trade-off | Abstract |
|
09:45 - 10:25
|
Dirk Lorenz
(Technische Universität Braunschweig) Unrolling of primal-dual algorithms | Abstract |
|
10:30 - 11:10
|
Coffee break
|
|
|
11:10 - 11:50
|
Dejan Slepcev
(Carnegie Mellon University) Proper weighted Laplacian for semi-supervised learning | Abstract |
|
12:00 - 13:30
|
Lunch break
|
|
|
13:30 - 14:10
|
Michael Mahoney
(UC Berkeley) Why deep learning works: traditional and heavy-tailed implicit self-regularization in deep neural networks | Abstract |
|
14:15 - 14:55
|
Christoph Brune
(University of Twente) Deep learning decomposition for inverse problems | Abstract |
|
15:00 - 15:30
|
Coffee break
|
|
|
15:30 - 16:10
|
Venkat Chandrasekaran
(California Institute of Technology) A geometric perspective on false discovery control | Abstract |
|
16:15 - 16:55
|
Giovanni S. Alberti
(Universit`a di Genova) Adversarial deformations in deep neural networks | Abstract |
Tuesday, May 28, 2019 |
||||||||
|
|
||||||||
|
08:30 - 09:00
|
Coffee & Croissants
|
|
|
|
09:00 - 09:40
|
Jeffrey Willam Calder
(University of Minnesota) PDE continuum limits for prediction with expert advice | Abstract |
|
09:45 - 10:25
|
Christoph Schwab
(ETH Zurich) Deep neural network expression in Bayesian PDE constrained data assimilation and inversion | Abstract |
|
10:30 - 11:10
|
Coffee break
|
|
|
11:10 - 11:50
|
Bamdad Hosseini
(California Institute of Technology) Consistency of semi-supervised learning algorithms on graphs | Abstract |
|
12:00 - 13:30
|
Lunch break
|
|
|
13:30 - 14:10
|
Nicolas Le Roux
(Google Brain - Montréal) On the interplay between noise and curvature in deep learning | Abstract |
|
14:15 - 14:55
|
Matteo Santacesaria
(University of Genoa) Inverse problems for PDEs via compressed sensing | Abstract |
|
15:00 - 15:30
|
Coffee break
|
|
|
15:30 - 16:10
|
Matthew Thorpe
(University of Cambridge) On the well posedness of graph Laplacian regularisation in semi-supervised learning | Abstract |
|
16:15 - 16:55
|
Nikola Kovachki
(California Institute of Technology) Continuous time limits for momentum methods as implemented in machine learning | Abstract |
Wednesday, May 29, 2019 |
||||||||
|
|
||||||||
|
08:30 - 09:00
|
Coffee & Croissants
|
|
|
|
09:00 - 09:40
|
Mauro Maggioni
(Johns Hopkins University) Statistical learning & dynamical systems: exploiting hidden low-dimensional structures | Abstract |
|
09:45 - 10:25
|
Bharath Sriperumbudur
(The Pennsylvania State University) Distribution regression: theory and applications |
|
10:30 - 11:10
|
Coffee break
|
|
|
11:10 - 11:50
|
Daniel Sanz-Alonso
(University of Chicago) Function-space inverse problems and semi-supervised learning | Abstract |
|
12:00 - 13:30
|
Lunch break
|
|
|
13:30 - 14:10
|
Franca Hoffmann
(California Institute of Technology) Geometric insights into spectral clustering by graph Laplacian embeddings | Abstract |
|
14:15 - 14:55
|
Eldad Haber
(The University of British Columbia) Efficient architectures for deep neural networks | Abstract |
|
15:00 - 15:30
|
Coffee break
|
|
|
15:30 - 16:10
|
Alfredo Garbuno Inigo
(Caltech) Optimize, learn, sample | Abstract |
|
16:15 - 16:55
|
Adam M. Oberman
(McGill University) |
Last modified : Monday, May 27, 2019 14:07