ICLR2017: Identity and Invariance

Identity reproduction (Sunday Morning Insight: A Quick Panorama of Sensing from Direct Imaging to Machine Learning) and invariance are very important topics in engineering and science. It is therefore no surprise these themes appear in models trying to figure out the world. Here are a few I noticed in the the 500 submissions at the ICLR 2017 conference that are in the open review process.

• Identity Matters in Deep Learning Moritz Hardt, Tengyu Ma

• Learning Invariant Representations Of Planar Curves Gautam Pai, Aaron Wetzler, Ron Kimmel
• Learning Invariant Feature Spaces to Transfer Skills with Reinforcement Learning Abhishek Gupta, Coline Devin, YuXuan Liu, Pieter Abbeel, Sergey Levine
• Improving Invariance and Equivariance Properties of Convolutional Neural Networks Christopher Tensmeyer, Tony Martinez
• Warped Convolutions: Efficient Invariance to Spatial Transformations Joao F. Henriques, Andrea Vedaldi
• Linear Time Complexity Deep Fourier Scattering Network and Extension to Nonlinear Invariants Randall Balestriero, Herve Glotin

Join the CompressiveSensing subreddit or the Google+ Community or the Facebook page and post there !
Liked this entry ? subscribe to Nuit Blanche's feed, there's more where that came from. You can also subscribe to Nuit Blanche by Email, explore the Big Picture in Compressive Sensing or the Matrix Factorization Jungle and join the conversations on compressive sensing, advanced matrix factorization and calibration issues on Linkedin.