Beyond Backpropagation: Uncertainty Propagation thumbnail
Pause
Mute
Subtitles not available
Playback speed
0.25
0.5
0.75
1
1.25
1.5
1.75
2
Full screen

Beyond Backpropagation: Uncertainty Propagation

Published on May 27, 20165655 Views

Deep learning is founded on composable functions that are structured to capture regularities in data and can have their parameters optimized by backpropagation (differentiation via the chain rule). Th

Related categories

Chapter list

Beyond Backpropagation: Uncertainty Propagation00:00
David00:02
Neural network - 101:30
Neural network - 202:38
Neural network - 304:03
Global information storage capacity in optimally compressed bytes04:24
Zero MeaZero Mean Gaussn Gaussian Process Sample05:57
Gaussian Processes09:13
Neural Network10:59
Bayesian Optimization - 111:41
Bayesian Optimization - 212:35
Open Data Science and Africa14:53
Disease Incidence for Malaria15:34
uganda18:01
Deployed with UN Global Pulse Lab18:25
Results18:44
Quote19:26
Algorithm21:00
Parametric Model - 122:24
Parametric Model - 223:05
Parametric Model - 323:57
Parametric Model - 425:32
Parametric Model - 527:01
Parametric Model - 627:49
Parametric Model - 729:16
Two Gaussian processes30:09
Render Gausian Non Gausian31:23
Stochastic Process Composition31:59
MLP (200 iterations)32:55
MLP (converged)33:10
GP33:40
DeepGP233:42
DeepGP333:46
Model33:54
Regression34:05
Classical Latent Variables34:07
Use Abstraction for Complex Systems37:19
Example: Motion Capture Modeling37:24
Modelling Digits37:30
Inferentia37:37
Numerical Issues37:48
Health38:07
To Find Out More38:17
Thank you39:18