What you always wanted to know about Deep Learning, but were afraid to ask

on YouTube ↗01h 04m

" summary

Covers core mechanics: weights, biases, activation functions (sigmoid/softmax/tanh/ReLU/leaky ReLU), loss and backprop with SGD/Adam and the math for dL/dw, plus batch size/epoch tradeoffs. Uses TensorFlow + Keras on MNIST (60,000 images) reaching 97% accuracy on MNIST and exports models to TensorFlow Lite via a Flutter APK to run handwriting recognition on-device. Shows IMDB sentiment demo and a CNN fruit classifier trained on a few hundred images per class (fruit classifier: banana / durian / strawberry), with webcam and MobileNet object-detection demos that work but still misclassify sometimes.

AI-written summary · about

" chapters (33)

1. 0:00Start
2. 0:10Is this session for you?
3. 1:45Agenda
4. 5:47What is Deep Learning?
5. 6:29Implementing Deep Learning using Neural Networks
6. 7:18Hidden Layer(s)
7. 7:43Weights and Biases
8. 9:09Calculating the Result of a Node (Forward Propagation)
9. 10:06Feeding the Result of a Node to an Activation Function
10. 11:50Categories of Activation Functions
11. 12:01Binary Step Function
12. 12:36Analogy
13. 13:45Non-Linear Activation
14. 20:06Evaluating Performance
15. 22:57Binary Cross Entropy
16. 23:10In Summary Activation Function and Loss Function
17. 24:42Use of Optimizers
18. 25:25Back Propagation
19. 26:09Expressing Loss Function
20. 26:59Derivative (Gradient)
21. 29:05A walkthrough
22. 30:23Initializing the Weights
23. 31:30Taking Partial Differentials
24. 32:42Significance of the Partial Differentials
25. 33:26Updating the Weights using Stochastic Gradient Descent
26. 35:54Batch and Epoch
27. 36:29Batch Size and Epoch
28. 38:13Types of Gradient Descent
29. 39:05In Summary Activation Function Optimier, and Los
30. 39:47TensorFlow and Keras
31. 40:421. Building an Artificial Neural Network using the MNIST Dataset
32. 48:32Handwriting Recognition
33. 58:36Object Detection

" related talks

• 35mDeconstructing Deep LearningMark West · Big Data Conference Europe 2019
• 47mDeep Learning: a practical introduction from an Android dev perspectiveJoaquim Verges · droidcon SF 2019
• 01h 29mHands-on introduction to Deep Learning with Keras and TensorflowRodrigo Agundez · PyData Amsterdam 2018
• 27mDemystifying Deep Neural NetworksRosie Campbell · Front-Trends 2017
• 01h 06mDeep Learning with TensorFlowSeth Juarez · NDC Sydney 2018
• 01h 07mDeep Learning with TensorFlowSeth Juarez · NDC Oslo 2018
• 44mMachine Learning ExposedJames Weaver · Codemotion Amsterdam 2017
• 45mThe most soft and fluffy intro into Machine Learning and Deep Neural NetworksAleksey Okhrimenko · DevFest Gorky 2019