Deep Learning · TEE · chapter

Neural-network foundations

Technical definitions, neuron models, activations, losses, forward propagation and backpropagation. These foundations support the MT2 calculations.

1. What does learning from data mean
2. Artificial intelligence machine learning and deep learning
3. Classification regression and learning types
4. Training validation and test data
5. Parameters and hyperparameters
6. What one artificial neuron does
7. Biological neurons and the McCulloch-Pitts model
8. Perceptrons and the threshold
9. Drill 11 One perceptron learning update
10. Why XOR needs more than one perceptron
11. Drill 12 XOR using two hidden units
12. Activation functions and how to calculate them
13. Drill 8 Activation contrast
14. Choosing a loss
15. Drill 2 Half sum versus mean
16. Drill 7 Sigmoid with binary cross entropy
17. Drill 9 Softmax and true class loss
18. Cross-entropy for probability distributions
19. Forward propagation through multiple neurons
20. Your first complete training step
21. Complete sigmoid training example
22. Drill 10 Sigmoid neuron gradient
23. Backpropagation with one hidden neuron
24. More outputs and vectorized backpropagation
25. Drill 13 Full ReLU backpropagation
26. Drill 14 Two output sigmoid backpropagation
27. Drill 15 A negative ReLU blocks this path
28. Counting parameters in a dense network
29. Teacher vectorized ReLU example
30. Teams photos full sigmoid example
31. MT1A Q1a
32. MT1A Q1b
33. MT1A Q2
34. MT1A Q3
35. MT1A Q4
36. MT1B Q1a
37. MT1B Q1b
38. MT1B Q2
39. MT1B Q3
40. MT1B Q4
41. F25 Q1a
42. F25 Q1b
43. F25 Q2a
44. F25 Q3a
45. F25 Q6b
46. F25 Q7a
47. R26 Q1a
48. R26 Q1b
49. R26 Q2a
50. R26 Q3a
51. R26 Q6b
52. R26 Q7a
53. F23 Q1a
54. F23 Q2a
55. F23 Q2b
56. F23 Q3a
57. F23 Q5b
58. R24 Q1a
59. R24 Q1b
60. R24 Q2a
61. R24 Q3a
62. R24 Q4a
63. F22 Q1c
64. F22 Q1d
65. F22 Q2b
66. F22 Q7b
67. R23 Q1c
68. R23 Q1e
69. R23 Q3b
70. R23 Q5b