Semi-Supervised Learning via New Deep Network Inversion

Anonymous

Nov 03, 2017 (modified: Nov 03, 2017) ICLR 2018 Conference Blind Submission readers: everyone Show Bibtex
  • Abstract: We exploit a recently derived inversion scheme for arbitrary deep neural networks to develop a new semi-supervised learning framework that applies to a wide range of systems and problems. The approach outperforms current state-of-the-art methods on MNIST for small label set. Experiments with one-dimensional signals highlight the generality of the method. Importantly, our approach is simple, efficient, and requires no change in the deep network architecture.
  • TL;DR: We exploit an inversion scheme for arbitrary deep neural networks to develop a new semi-supervised learning framework that outperforms current state-of-the-art methods on MNIST.
  • Keywords: inversion scheme, deep neural networks, semi-supervised learning, MNIST

Loading