Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

The Neural Network Zoo - Xavier Giro-i-Nieto - UPC Barcelona 2018

304 views

Published on

https://telecombcn-dl.github.io/2018-dlcv/

Deep learning technologies are at the core of the current revolution in artificial intelligence for multimedia data analysis. The convergence of large-scale annotated datasets and affordable GPU hardware has allowed the training of neural networks for data analysis tasks which were previously addressed with hand-crafted features. Architectures such as convolutional neural networks, recurrent neural networks and Q-nets for reinforcement learning have shaped a brand new scenario in signal processing. This course will cover the basic principles and applications of deep learning to computer vision problems, such as image classification, object detection or image captioning.

Published in: Data & Analytics
  • Be the first to comment

The Neural Network Zoo - Xavier Giro-i-Nieto - UPC Barcelona 2018

  1. 1. Xavier Giro-i-Nieto xavier.giro@upc.edu Associate Professor Universitat Politecnica de Catalunya Technical University of Catalonia The Neural Network Zoo Day 1 Lecture 2 #DLUPC http://bit.ly/dlcv2018
  2. 2. bit.ly/DLCV2018 #DLUPC 2 Acknowledgements Fjodor Van Veen, “The Neural Network Zoo” The Asimov Institute (2016)
  3. 3. bit.ly/DLCV2018 #DLUPC 3 The Perceptron J. Alammar, “A visual and interactive guide to the Basics of Neural Networks” (2016) F. Van Veen, “The Neural Network Zoo” (2016)
  4. 4. bit.ly/DLCV2018 #DLUPC 4 Neural Network = Multi-layer Perceptron 2 layers of perceptrons F. Van Veen, “The Neural Network Zoo” (2016)
  5. 5. bit.ly/DLCV2018 #DLUPC 5 Deep Neural Network (DNN) F. Van Veen, “The Neural Network Zoo” (2016)
  6. 6. bit.ly/DLCV2018 #DLUPC 6 Recurrent Neural Network (RNN) Alex Graves, “Supervised Sequence Labelling with Recurrent Neural Networks” The hidden layers and the output depend from previous states of the hidden layers
  7. 7. bit.ly/DLCV2018 #DLUPC 7 Recurrent Neural Network (RNN) F. Van Veen, “The Neural Network Zoo” (2016)
  8. 8. bit.ly/DLCV2018 #DLUPC 8 Recurrent Neural Network (RNN) F. Van Veen, “The Neural Network Zoo” (2016)
  9. 9. bit.ly/DLCV2018 #DLUPC 9 Recurrent Neural Network (RNN) Deep Learning TV, “Recurrent Neural Networks - Ep. 9”
  10. 10. bit.ly/DLCV2018 #DLUPC 10 Convolutional Neural Network (CNN) LeCun, Y., Bottou, L., Bengio, Y., & Haffner, P. (1998). Gradient-based learning applied to document recognition. Proceedings of the IEEE, 86(11), 2278-2324. LeNet-5
  11. 11. bit.ly/DLCV2018 #DLUPC 11 Convolutional Neural Network (CNN) Orange A Krizhevsky, I Sutskever, GE Hinton “Imagenet classification with deep convolutional neural networks” NIPS 2012. AlexNet
  12. 12. bit.ly/DLCV2018 #DLUPC 12 Convolutional Neural Network (CNN) Deep Learning TV, “Convolutional Neural Networks - Ep. 8”
  13. 13. bit.ly/DLCV2018 #DLUPC 13 Deconvolutional Neural Network Junting Pan, SalGAN (2017) F. Van Veen, “The Neural Network Zoo” (2016)
  14. 14. bit.ly/DLCV2018 #DLUPC 14 Deep Residual Network F. Van Veen, “The Neural Network Zoo” (2016) He, Kaiming, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. "Deep Residual Learning for Image Recognition." CVPR 2016 [slides]
  15. 15. bit.ly/DLCV2018 #DLUPC 15 Skip Connections Ronneberger, Olaf, Philipp Fischer, and Thomas Brox. "U-net: Convolutional networks for biomedical image segmentation." MICCAI 2015 U-Net
  16. 16. bit.ly/DLCV2018 #DLUPC 16 Skip Connections Van Den Oord, Aaron, Sander Dieleman, Heiga Zen, Karen Simonyan, Oriol Vinyals, Alex Graves, Nal Kalchbrenner, Andrew Senior, and Koray Kavukcuoglu. "Wavenet: A generative model for raw audio." ISCA 2016.
  17. 17. bit.ly/DLCV2018 #DLUPC 17 Dense Connections Huang, Gao, Zhuang Liu, Kilian Q. Weinberger, and Laurens van der Maaten. "Densely connected convolutional networks." CVPR 2017. [code] Dense Block of 5-layers with a growth rate of k=4 Connect every layer to every other layer of the same filter size.
  18. 18. bit.ly/DLCV2018 #DLUPC 18 Dense Connections Huang, Gao, Zhuang Liu, Kilian Q. Weinberger, and Laurens van der Maaten. "Densely connected convolutional networks." CVPR 2017. [code]
  19. 19. bit.ly/DLCV2018 #DLUPC 19Huang, Gao, Zhuang Liu, Kilian Q. Weinberger, and Laurens van der Maaten. "Densely connected convolutional networks." CVPR 2017. [code] Dense Connections
  20. 20. bit.ly/DLCV2018 #DLUPC 20 Autoencoder (AE) F. Van Veen, “The Neural Network Zoo” (2016)
  21. 21. bit.ly/DLCV2018 #DLUPC 21 Autoencoder (AE) “Deep Learning Tutorial”, Dept. Computer Science, Stanford Autoencoders: ● Predict at the output the same input data. ● Do not need labels:
  22. 22. bit.ly/DLCV2018 #DLUPC 22 Autoencoder (AE) “Deep Learning Tutorial”, Dept. Computer Science, Stanford Dimensionality reduction: ● Use hidden layer as a feature extractor of any desired size. Application #1
  23. 23. bit.ly/DLCV2018 #DLUPC 23 Autoencoder (AE) Encoder W1 Decoder W2 hdata reconstruction Loss (reconstruction error) Latent variables (representation/features) Figure: Kevin McGuinness (DLCV UPC 2017) Pretraining: 1. Initialize a NN solving an autoencoding problem. Application #2
  24. 24. bit.ly/DLCV2018 #DLUPC 24 Autoencoder (AE) Pretraining: 1. Initialize a NN solving an autoencoding problem. 2. Train for final task with “few” labels. Figure: Kevin McGuinness (DLCV UPC 2017) Encoder W1 hdata Classifier WC Latent variables (representation/features) prediction y Loss (cross entropy) Application #2
  25. 25. bit.ly/DLCV2018 #DLUPC 25 Autoencoder (AE) Deep Learning TV, “Autoencoders - Ep. 10”
  26. 26. bit.ly/DLCV2018 #DLUPC 26 Variational Autoencoder (AE) F. Van Veen, “The Neural Network Zoo” (2016)
  27. 27. bit.ly/DLCV2018 #DLUPC 27 Adversarial Networks Goodfellow, Ian, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, Sherjil Ozair, Aaron Courville, and Yoshua Bengio. "Generative adversarial nets." NIPS 2014 Goodfellow, Ian. "NIPS 2016 Tutorial: Generative Adversarial Networks." arXiv preprint arXiv:1701.00160 (2016). F. Van Veen, “The Neural Network Zoo” (2016)
  28. 28. bit.ly/DLCV2018 #DLUPC 28 Differentiable Neural Computers (DNC) F. Van Veen, “The Neural Network Zoo” (2016) Graves, Alex, Greg Wayne, and Ivo Danihelka. "Neural turing machines." arXiv preprint arXiv:1410.5401 (2014). [slides] [code]
  29. 29. bit.ly/DLCV2018 #DLUPC 29 Differentiable Neural Computers (DNC) von Neumann architecture (1952) Neural Turing Machine (2014) Graves, Alex, Greg Wayne, and Ivo Danihelka. "Neural turing machines." arXiv preprint arXiv:1410.5401 (2014). [slides] [code] End-to-end trainable
  30. 30. bit.ly/DLCV2018 #DLUPC 30 Differentiable Neural Computers (DNC) [Siraj Raval on YouTube]
  31. 31. bit.ly/DLCV2018 #DLUPC 31 MAC Network Hudson, Drew A., and Christopher D. Manning. "Compositional attention networks for machine reasoning." ICLR 2018.
  32. 32. bit.ly/DLCV2018 #DLUPC 32 More architectures to come... Yann LeCun, “A Path to AI”, Beneficial AI 2017. von Neumann architecture (1952)
  33. 33. bit.ly/DLCV2018 #DLUPC 33 The Full Story F. Van Veen, “The Neural Network Zoo” (2016)
  34. 34. bit.ly/DLCV2018 #DLUPC 34 The Prequel Fjodor Van Veen, “Neural Network Zoo Prequel: Cells and Layers”(2017)
  35. 35. bit.ly/DLCV2018 #DLUPCQuestions 35

×