Embedded Files
Workshop on Equivariance and Data Augmentation
online, hosted by the University of Pennsylvania
Exploiting symmetry in structured data is a powerful way to improve the learning and generalization ability of AI systems, and extract higher quality information, in applications ranging from vision, imaging, and NLP to robotics. This is exemplified by convolutional neural nets, which are an ubiquitous architecture. Recently, there has been a great deal of progress to develop improved equivariant and invariant learning architectures, as well as improved data augmentation methods. There has also been progress on the theoretical foundations of the area, from the perspectives of statistics and optimization. The notion of adding data via data augmentation also arises in problems such as adversarial robustness. This workshop will bring together leading researchers in the area to discuss the state of the art of the field. The activity is part of the Center for Foundations of Information Processing at Penn, supported by NSF TRIPODS.
Due to Covid-19, the workshop will be held online as a Zoom webinar. While this means less in-person interaction, it also means that the talks are accessible to everyone for free, without the costs of travel.
Due to Covid-19, the workshop will be held online as a Zoom webinar. While this means less in-person interaction, it also means that the talks are accessible to everyone for free, without the costs of travel.
Recorded talks are available as a Youtube playlist. Individual videos are also linked below. The link below takes you to the playlist in the "play all" mode.
The Speakers
The Speakers
Christine Allen-Blanchette, Princeton University
Fabio Anselmi, Baylor College of Medicine/MIT
Pratik Chaudhari, University of Pennsylvania
Taco Cohen, Qualcomm AI Research
Carlos Esteves, University of Pennsylvania
Jane H Lee, Twitter/Yale University
Chelsea Finn, Stanford University
Haggai Maron, NVIDIA Research
Danilo J. Rezende, Google DeepMind
Alejandro Ribeiro, University of Pennsylvania
Alexander Robey, University of Pennsylvania
Alexander Robey, University of Pennsylvania
Tess E Smidt, Lawrence Berkeley National Laboratory
Greg Valiant, Stanford University
Kilian Q. Weinberger, Cornell University
Mark van der Wilk, Imperial College London
Hongyang R. Zhang, Northeastern University
Schedule: Friday, September 4, 2020
Schedule: Friday, September 4, 2020
all times EST (Eastern Time/New York Time)
9:00 - 9:30
Leveraging permutation group symmetries for designing equivariant neural networks
Leveraging permutation group symmetries for designing equivariant neural networks
Haggai Maron, NVIDIA Research
10:30 - 11:00
Learning with Marginalized Augmenation
Learning with Marginalized Augmenation
Kilian Q. Weinberger, Cornell University
11:00 - 11:20
Learning Invariances through Backprop with Bayesian Model Selection
Learning Invariances through Backprop with Bayesian Model Selection
Mark van der Wilk, Imperial College London
Break 11:20 - 11:40
11:40 - 12:00
Model-based Robust Deep Learning
Model-based Robust Deep Learning
Alexander Robey, University of Pennsylvania
12:00 - 12:20
Algebraic Neural Networks: Symmetry and Stability
Algebraic Neural Networks: Symmetry and Stability
Alejandro Ribeiro, University of Pennsylvania
12:20 - 12:40
Learning with few labeled data
Learning with few labeled data
Pratik Chaudhari, University of Pennsylvania
12:40 - 12:55
Spin-Weighted Spherical CNNs
Spin-Weighted Spherical CNNs
Carlos Esteves, University of Pennsylvania
12:55 - 1:10
A group-theoretic framework for data augmentation
A group-theoretic framework for data augmentation
Jane H Lee, Twitter/Yale University
1:10 - 1:55
Panel
Panel
Lunch Break 1:55 -2:30
2:30 - 3:00
Neurally plausible mechanisms for learning selective and invariant representations
Neurally plausible mechanisms for learning selective and invariant representations
Fabio Anselmi, Baylor College of Medicine/MIT
3:00 - 3:30
LagNetViP: A Lagrangian Neural Network for Video Prediction
LagNetViP: A Lagrangian Neural Network for Video Prediction
Christine Allen-Blanchette, Princeton University
3:30 - 4:00
Unintended features of Euclidean symmetry equivariant neural networks
Unintended features of Euclidean symmetry equivariant neural networks
Tess E Smidt, Lawrence Berkeley National Laboratory
5:00 - 5:30
Generalization Effects of Linear Transformations in Data Augmentation
Generalization Effects of Linear Transformations in Data Augmentation
Hongyang R. Zhang, Northeastern University
Contacts: Please contact Edgar Dobriban or Kostas Daniilidis with any questions.
Page updated
Google Sites
Report abuse