Understanding geometry of encoder-decoder CNNs

Cited 1 time in webofscience Cited 0 time in scopus
  • Hit : 172
  • Download : 0
Encoder-decoder networks using convolutional neural network (CNN) architecture have been extensively used in deep learning literatures thanks to its excellent performance for various inverse problems. However, it is still difficult to obtain coherent geometric view why such an architecture gives the desired performance. Inspired by recent theoretical understanding on generalizability, expressivity and optimization landscape of neural networks, as well as the theory of deep convolutional framelets, here we provide a unified theoretical framework that leads to a better understanding of geometry of encoder-decoder CNNs. Our unified framework shows that encoder-decoder CNN architecture is closely related to nonlinear frame representation using combinatorial convolution frames, whose expressivity increases exponentially with the depth. We also demonstrate the importance of skipped connection in terms of expressivity, and optimization landscape.
Publisher
International Machine Learning Society (IMLS)
Issue Date
2019-06
Language
English
Citation

36th International Conference on Machine Learning, ICML 2019, pp.12245 - 12254

URI
http://hdl.handle.net/10203/280296
Appears in Collection
AI-Conference Papers(학술대회논문)
Files in This Item
There are no files associated with this item.
This item is cited by other documents in WoS
⊙ Detail Information in WoSⓡ Click to see webofscience_button
⊙ Cited 1 items in WoS Click to see citing articles in records_button

qr_code

  • mendeley

    citeulike


rss_1.0 rss_2.0 atom_1.0