biologically inspired sparse restricted boltzmann machines
play

Biologically-Inspired Sparse Restricted Boltzmann Machines Pablo - PowerPoint PPT Presentation

Biologically-Inspired Sparse Restricted Boltzmann Machines Pablo Tostado Michael Wiest Alice Yepremyan 1 Content 1. Motivation 2. Background a. Restricted Boltzmann Machines b. Sparsity 3. Methods a. Pruning algorithm b.


  1. Biologically-Inspired Sparse Restricted Boltzmann Machines Pablo Tostado Michael Wiest Alice Yepremyan 1

  2. Content 1. Motivation 2. Background a. Restricted Boltzmann Machines b. Sparsity 3. Methods a. Pruning algorithm b. Evaluation criteria 4. Results 5. Discussion 6. Future directions 2

  3. Motivation 3

  4. Sparsity In Biological systems: 1 In Artificial systems: 2 Reduces computational Increases computational ● ● complexity efficiency ● Increases speed ● Less prone to overfitting ● Yield to higher probabilities ● Can often lead to better solutions in neural networks Total energy consumed (AlexNet and dropout) ● decreases with increasing sparsity 4

  5. Background 5

  6. Restricted Boltzmann Machines ● Generative, stochastic Artificial Neural Networks ● Fully connected bipartite graphs Able to learn probability distributions over its set of inputs ● Building block of deeper neural networks ● ● Hebbian nature in the learning algorithm 6

  7. RBM: Structure [3] 7

  8. RBM: Energy of Network (v,h) Z = v,h 8

  9. Methods 9

  10. RBM architecture Visible Units: Hidden Units: 784 units 1) 100 nodes 2) 500 nodes MNIST (28x28) [3] 10

  11. Our Algorithm 1. Do an initial round of training (1000 epochs). 2. Prune the P* lowest weight connection (set weight to zero) 3. Train again (400 epochs) 4. Repeat 2 and 3 until desired amount of pruning done 5. Do a final round of training (1000 epochs) 11

  12. Data ● Training: 1000 MNIST images (100 of each character) ● Testing: 100 MNIST images (10 of each character) *Due to computing time. 12

  13. Evaluation Criteria for Pruned and Unpruned ● Accuracy* of image reconstruction from: 1. Noisy image 2. Occluded image Altering the parameters ● Hidden nodes: 100 and 500 ○ Percent noisy/occluded: 5, 10, 25, 50% ○ ● Visual nodes represented by hidden nodes Pruning over time ● 13

  14. Evaluation Criteria: Noise 20% Noise 14

  15. Evaluation Criteria: Square Occlusion 20% occlusion 15

  16. Image Recovery Example N Steps 16

  17. Noise Convergence (20% noise, 25% pruning)

  18. Occlusion Convergence (20% occlude, 25% prune) 18

  19. Image recovery Scoring 10 examples of each image ● For each image do 100 iterations of random noise/occlusion ● ● For some iteration i : 19

  20. Results 20

  21. Pruning preference (100 hidden nodes) 0% 5% 10% 25% 50% 80% Percentage Pruned 21

  22. Pruning preference (500 hidden nodes) 0% 5% 10% 25% 50% Percentage Pruned 22

  23. 23

  24. 24

  25. 100 Hidden Nodes 25

  26. 100 Hidden Nodes 26

  27. 500 Hidden Nodes 27

  28. 500 Hidden Nodes 28

  29. Discussion: Example of Denoising Input Img 50% noise 10% pruning 0% pruning 29

  30. Number of Training Epochs 30

  31. Training Error 31

  32. Future directions: ● Train over more images Alternate pruning heuristics (L1 norm) ● Train over more hidden nodes ● ● Computational efficiency evaluation ● Build a classifier on top of MNIST pixels 32

  33. Thanks! 33

  34. Additional Slides 34

  35. 100 Hidden Nodes 35

  36. 100 Hidden Nodes 36

  37. 500 Hidden Nodes 37

  38. 500 Hidden Nodes 38

  39. Works Cited [1] B. A. Olshausen and D. J. Field. Sparse coding with an overcomplete basis set: A strategy employed by V1? Vision Research, 37:3311–3325, 1997. [2] S. Changpinyo, M. Sandler, and A. Zhmoginov. The power of sparsity in convolutional neural networks. CoRR, abs/1702.06257, 2017. [3] Chris Nicholson, Adam Gibson, Skymind team. “A Beginner's Tutorial for Restricted Boltzmann Machines.” A Beginner's Tutorial for Restricted Boltzmann Machines - Deeplearning4j: Open-Source, Distributed Deep Learning for the JVM , deeplearning4j.org/restrictedboltzmannmachine. 39

  40. Measure of Error: KL-Divergence 40

Download Presentation
Download Policy: The content available on the website is offered to you 'AS IS' for your personal information and use only. It cannot be commercialized, licensed, or distributed on other websites without prior consent from the author. To download a presentation, simply click this link. If you encounter any difficulties during the download process, it's possible that the publisher has removed the file from their server.

Recommend


More recommend