wit ith im image clu lustering
play

Wit ith Im Image Clu lustering Jianwei Yang Devi Parikh Dhruv - PowerPoint PPT Presentation

Unsuperv rvised Learning Jointly Wit ith Im Image Clu lustering Jianwei Yang Devi Parikh Dhruv Batra Vir irgin inia ia Tech 1 https://filebox.ece.vt.edu/~jw2yang/ 2 Huge amount of images!!! 3 Huge amount of images!!! Learning


  1. Unsuperv rvised Learning Jointly Wit ith Im Image Clu lustering Jianwei Yang Devi Parikh Dhruv Batra Vir irgin inia ia Tech 1 https://filebox.ece.vt.edu/~jw2yang/

  2. 2

  3. Huge amount of images!!! 3

  4. Huge amount of images!!! Learning without annotation efforts 4

  5. Huge amount of images!!! Learning without annotation efforts What we need to learn? 5

  6. Huge amount of images!!! Learning without annotation efforts What we need to learn? An open problem 6

  7. Huge amount of images!!! Learning without annotation efforts What we need to learn? An open problem A hot problem 7

  8. Huge amount of images!!! Learning without annotation efforts What we need to learn? An open problem A hot problem Various methodologies 8

  9. Learning distribution (structure) Clustering 9 Jain, Anil K., M. Narasimha Murty, and Patrick J. Flynn. "Data clustering: a review." ACM computing surveys (CSUR) 31.3 (1999): 264-323.

  10. Learning distribution (structure) Clustering K-means (Image Credit: Jesse Johnson) 10 Jain, Anil K., M. Narasimha Murty, and Patrick J. Flynn. "Data clustering: a review." ACM computing surveys (CSUR) 31.3 (1999): 264-323.

  11. Learning distribution (structure) Clustering K-means (Image Credit: Jesse Johnson) Hierarchical Clustering 11 Jain, Anil K., M. Narasimha Murty, and Patrick J. Flynn. "Data clustering: a review." ACM computing surveys (CSUR) 31.3 (1999): 264-323.

  12. Learning distribution (structure) Clustering K-means (Image Credit: Jesse Johnson) Spectral Clustering Hierarchical Clustering Manor et al, NIPS’04 12 Jain, Anil K., M. Narasimha Murty, and Patrick J. Flynn. "Data clustering: a review." ACM computing surveys (CSUR) 31.3 (1999): 264-323.

  13. Learning distribution (structure) Clustering K-means (Image Credit: Jesse Johnson) Spectral Clustering Hierarchical Clustering Graph Cut Manor et al, NIPS’04 Shi et al, TPAMI’00 13 Jain, Anil K., M. Narasimha Murty, and Patrick J. Flynn. "Data clustering: a review." ACM computing surveys (CSUR) 31.3 (1999): 264-323.

  14. Learning distribution (structure) Clustering K-means (Image Credit: Jesse Johnson) Spectral Clustering Hierarchical Clustering Graph Cut Manor et al, NIPS’04 Shi et al, TPAMI’00 DBSCAN, Ester et al, KDD’96 (Image Credit: Jesse Johnson) 14 Jain, Anil K., M. Narasimha Murty, and Patrick J. Flynn. "Data clustering: a review." ACM computing surveys (CSUR) 31.3 (1999): 264-323.

  15. Learning distribution (structure) Clustering K-means (Image Credit: Jesse Johnson) Spectral Clustering Hierarchical Clustering Graph Cut Manor et al, NIPS’04 Shi et al, TPAMI’00 DBSCAN, Ester et al, KDD’96 (Image Credit: Jesse Johnson) EM Algorithm, Dempster et al, JRSS’77 15 Jain, Anil K., M. Narasimha Murty, and Patrick J. Flynn. "Data clustering: a review." ACM computing surveys (CSUR) 31.3 (1999): 264-323.

  16. Learning distribution (structure) Clustering K-means (Image Credit: Jesse Johnson) Spectral Clustering Hierarchical Clustering Graph Cut Manor et al, NIPS’04 Shi et al, TPAMI’00 DBSCAN, Ester et al, KDD’96 (Image Credit: Jesse Johnson) EM Algorithm, Dempster et al, JRSS’77 NMF, Xu et al, SIGIR‘03 (Image Credit: Conrad Lee) 16 Jain, Anil K., M. Narasimha Murty, and Patrick J. Flynn. "Data clustering: a review." ACM computing surveys (CSUR) 31.3 (1999): 264-323.

  17. Learning distribution (structure) Sub-space Analysis PCA (Image Credit: Jesse Johnson) ICA (Image Credit: Shylaja et al) tSNE, Maaten et al, JMLR’08 Subspace Clustering, Vidal et al. 17 Sparse coding, Olshausen et al. Vision Research’97

  18. Learning representation (feature) Autoencoder , Hinton et al, Science’06 DBN, Hinton et al, Science’06 DBM, Salakhutdinov et al, AISTATS’09 (Image Credit: Jesse Johnson) Bengio et al, TPAMI’13 Yoshua Bengio, Aaron Courville, and Pierre Vincent. "Representation learning: A review and new perspectives." IEEE Transactions on Pattern Analysis and Machine Intelligence. 35.8 (2013): 1798-1828. 18

  19. Learning representation (feature) VAE, Kingma et al, arXiv’13 (Image Credit: Fast Forward Labs) GAN, Goodfellow et al, NIPS’14 DCGAN, Radford et al, arXiv’15 (Image Credit: Mike Swarbrick Jones) 19

  20. Most Recent CV Works Ego-motion, Jayaraman et al, ICCV’15 Spatial context, Doersch et al, ICCV’15 Temporal context, Wang et al, ICCV’15 Solving Jigsaw, Noroozi et al, ECCV’16 20 Context Encoder, Deepak et al, CVPR’16

  21. Most Recent CV Works Visual concept clustering, Huang et al, CVPR’16 TAGnet , Wang et al, SDM’16 Graph constraint, Li et al, ECCV’16 Deep Embedding, Xie et al, ICML’16 21

  22. Our Work Joint Unsupervised Learning (JULE) of Deep Representations and Image Clusters 22

  23. Outline • Intuition • Approach • Experiments • Extensions 23

  24. Intuition Meaningful clusters can provide supervisory signals to learn image representations 24

  25. Intuition Meaningful clusters can provide supervisory signals to learn image representations Good representations help to get meaningful clusters 25

  26. Intuition Cluster images first, and then learn representations 26

  27. Intuition Cluster images first, and then learn representations Learn representations first, and then cluster images 27

  28. Intuition Cluster images first, and then learn representations Learn representations first, and then cluster images Cluster images and learn representations progressively 28

  29. Intuition Good cluster Good representations Good clusters Poor clusters Good representations Poor representations 29

  30. Intuition Good cluster Good representations Good clusters Poor clusters Good representations Poor representations 30

  31. Intuition Good cluster Good representations Good clusters Poor clusters Good representations Poor representations 31

  32. Intuition Good cluster Good representations Good clusters Poor clusters Good representations Poor representations 32

  33. Approach • Framework • Objective • Algorithm & Implementation 33

  34. Approach: Framework  arg min ( | , ) L y I  Convolutional Neural Network  argmin ( , L y | ) I Representation Agglomerative  y , Learning Clustering Agglomerative Clustering  argmin ( | , ) L y I y 34

  35. Approach: Framework Convolutional Neural Network Agglomerative Clustering   argmin ( | , ) L y I arg min ( | , ) L y I  y 35

  36. Approach: Recurrent Framework 36

  37. Approach: Recurrent Framework 37

  38. Approach: Recurrent Framework 38

  39. Approach: Recurrent Framework 39

  40. Approach: Recurrent Framework 40

  41. Approach: Recurrent Framework 41

  42. Approach: Recurrent Framework Backward at each time-step is time-consuming and prone to over-fitting! 42

  43. Approach: Recurrent Framework Backward at each time-step is time-consuming and prone to over-fitting! How about updating once for multiple time-steps? 43

  44. Approach: Recurrent Framework Partially Unrolling: divide all T time-steps into P periods In each period, we merge clusters for multiple times and update CNN parameters at the end of period 44

  45. Approach: Recurrent Framework Partially Unrolling: divide all T time-steps into P periods In each period, we merge clusters for multiple times and update CNN parameters at the end of period 45

  46. Approach: Recurrent Framework Partially Unrolling: divide all T time-steps into P periods In each period, we merge clusters for multiple times and update CNN parameters at the end of period P is determined by a hyper-parameter will be introduced later 46

  47. Approach: Objective Function   argmin ( | , ) L y I arg min ( | , ) L y I  argmin ( , L y | ) I  y  y , Overall loss: 47

  48. Approach: Objective Function Loss at time-step t: Conventional Agg. Proposed Agg. Clustering Strategy Clustering Strategy 48

  49. Approach: Objective Function Loss at time-step t: Affinity measure Conventional Agg. Proposed Agg. Clustering Strategy Clustering Strategy 49

  50. Approach: Objective Function Loss at time-step t: i-th cluster Conventional Agg. Proposed Agg. Clustering Strategy Clustering Strategy 50

  51. Approach: Objective Function Loss at time-step t: K_c nearest neighbor clusters of i-th cluster Conventional Agg. Proposed Agg. Clustering Strategy Clustering Strategy 51

  52. Approach: Objective Function Loss at time-step t: Affinity between i-th cluster and its NN Conventional Agg. Proposed Agg. Clustering Strategy Clustering Strategy 52

  53. Approach: Objective Function Loss at time-step t: Affinity between i-th cluster and its NN Conventional Agg. Proposed Agg. Clustering Strategy Clustering Strategy Differences between two cluster affinities 53

  54. Approach: Objective Function Merge these two clusters Loss at time-step t: Affinity between i-th cluster and its NN Conventional Agg. Proposed Agg. Clustering Strategy Clustering Strategy Differences between two cluster affinities 54

Recommend


More recommend