The Variational Fair Autoencoder thumbnail
Pause
Mute
Subtitles not available
Playback speed
0.25
0.5
0.75
1
1.25
1.5
1.75
2
Full screen

The Variational Fair Autoencoder

Published on May 27, 20162612 Views

We investigate the problem of learning representations that are invariant to certain nuisance or sensitive factors of variation in the data while retaining as much of the remaining information as poss

Related categories

Chapter list

The Variational Fair Autoencoder00:00
Motivation - 100:09
Motivation - 200:39
Tackling such problems01:17
Related work01:46
Contribution02:33
Unsupervised Variational Autoencoder02:58
Semi-Supervised VAE[4] for invariant representations04:12
Further invariance via posterior regulariaztion 05:44
Maximum Mean Discrepancy06:23
Fast MMD via Random Fourier Features06:53
Variational Fair Autoencoder07:45
Experiments - 108:14
Evaluation criteria08:20
Fair Classification09:18
Fair classification results - 109:47
Fair classification results - 211:36
Experiments - 212:19
Domain Adaptation12:22
Domain adaptation results12:41
Experiments - 313:01
Invariant feature learning13:03
Invariant feature learning results13:25
Conclusion & future work14:15
Thank you15:04