Deep Learning · TEE · index

Questions and complete answers

  1. Model answer: explain an artificial neuron
  2. Model answer: why a network needs nonlinear activation
  3. Model answer: forward propagation, backpropagation and updating
  4. Model answer: batch, stochastic and mini-batch training
  5. Model answer: momentum and Nesterov
  6. Model answer: AdaGrad, RMSProp and Adam
  7. Model answer: regularization and the L1–L2 difference
  8. Model answer: generalization and early stopping
  9. Model answer: dropout with training and inference explained
  10. Model answer: CNN structure and parameter sharing
  11. Model answer: shape, activations and parameters
  12. Model answer: feature extraction and fine-tuning
  13. Model answer: AlexNet, Inception and ResNet
  14. Model answer: why recurrent networks use a hidden state
  15. Model answer: backpropagation through time
  16. Model answer: LSTM and GRU memory
  17. MT1A Q1a
  18. MT1A Q1b
  19. MT1A Q2
  20. MT1A Q3
  21. MT1A Q4
  22. MT1B Q1a
  23. MT1B Q1b
  24. MT1B Q2
  25. MT1B Q3
  26. MT1B Q4
  27. F25 Q1a
  28. F25 Q1b
  29. F25 Q1c
  30. F25 Q1d
  31. F25 Q2a
  32. F25 Q2b
  33. F25 Q3a
  34. F25 Q3b
  35. F25 Q4a
  36. F25 Q4b
  37. F25 Q5a
  38. F25 Q5b
  39. F25 Q6a
  40. F25 Q6b
  41. F25 Q7a
  42. F25 Q7b
  43. R26 Q1a
  44. R26 Q1b
  45. R26 Q1c
  46. R26 Q1d
  47. R26 Q2a
  48. R26 Q2b
  49. R26 Q3a
  50. R26 Q3b
  51. R26 Q4a
  52. R26 Q4b
  53. R26 Q5a
  54. R26 Q5b
  55. R26 Q6a
  56. R26 Q6b
  57. R26 Q7a
  58. R26 Q7b
  59. F23 Q1a
  60. F23 Q1b
  61. F23 Q1c
  62. F23 Q1d
  63. F23 Q2a
  64. F23 Q2b
  65. F23 Q3a
  66. F23 Q3b
  67. F23 Q4a
  68. F23 Q4b
  69. F23 Q5a
  70. F23 Q5b
  71. F23 Q6a
  72. F23 Q6b
  73. F23 Q7a
  74. F23 Q7b
  75. R24 Q1a
  76. R24 Q1b
  77. R24 Q1c
  78. R24 Q1d
  79. R24 Q2a
  80. R24 Q2b
  81. R24 Q3a
  82. R24 Q3b
  83. R24 Q4a
  84. R24 Q4b
  85. R24 Q5a
  86. R24 Q5b
  87. R24 Q6a
  88. R24 Q6b
  89. R24 Q7a
  90. R24 Q7b
  91. F22 Q1a
  92. F22 Q1b
  93. F22 Q1c
  94. F22 Q1d
  95. F22 Q1e
  96. F22 Q2a
  97. F22 Q2b
  98. F22 Q3a
  99. F22 Q3b
  100. F22 Q4a
  101. F22 Q4b
  102. F22 Q5a
  103. F22 Q5b
  104. F22 Q6a
  105. F22 Q6b
  106. F22 Q7a
  107. F22 Q7b
  108. R23 Q1a
  109. R23 Q1b
  110. R23 Q1c
  111. R23 Q1d
  112. R23 Q1e
  113. R23 Q2a
  114. R23 Q2b
  115. R23 Q3a
  116. R23 Q3b
  117. R23 Q4a
  118. R23 Q4b
  119. R23 Q5a
  120. R23 Q5b
  121. R23 Q6a
  122. R23 Q6b
  123. R23 Q7a
  124. R23 Q7b
  125. Drill 1 Signs and brackets
  126. Drill 2 Half sum versus mean
  127. Drill 3 Natural log and probability
  128. Drill 4 Dot product and bias
  129. Drill 5 Matrix rows receive inputs
  130. Drill 6 Chain rule on a tiny loss
  131. Drill 7 Sigmoid with binary cross entropy
  132. Drill 8 Activation contrast
  133. Drill 9 Softmax and true class loss
  134. Drill 10 Sigmoid neuron gradient
  135. Drill 11 One perceptron learning update
  136. Drill 12 XOR using two hidden units
  137. Drill 13 Full ReLU backpropagation
  138. Drill 14 Two output sigmoid backpropagation
  139. Drill 15 A negative ReLU blocks this path
  140. Drill 16 Batch and sequential SGD
  141. Drill 17 Incomplete mini batch
  142. Drill 18 GD variation 1
  143. Drill 19 GD variation 2
  144. Drill 20 Momentum variation 1
  145. Drill 21 Momentum variation 2
  146. Drill 22 Nesterov variation 1
  147. Drill 23 Nesterov variation 2
  148. Drill 24 AdaGrad variation 1
  149. Drill 25 AdaGrad variation 2
  150. Drill 26 RMSProp variation 1
  151. Drill 27 RMSProp variation 2
  152. Drill 28 Adam variation 1
  153. Drill 29 Adam variation 2
  154. Drill 30 Regularized objective
  155. Drill 31 L1 update on positive and negative weights
  156. Drill 32 L2 update
  157. Drill 33 Inverted dropout
  158. Drill 34 Convolution with a negative coefficient
  159. Drill 35 Padding stride and channels
  160. Drill 36 Max and average pooling
  161. Drill 37 Complete small CNN
  162. Drill 38 Reverse a spatial formula
  163. Drill 39 Bottleneck cost with new values
  164. Drill 40 Residual addition
  165. Drill 41 Two RNN steps
  166. Drill 42 LSTM with a negative candidate
  167. Drill 43 Calculate gates before the cell
  168. Drill 44 GRU reset then mix
  169. Drill 45 BPTT shared weights
  170. Drill 46 Mixed convolution pool dense and loss
  171. Drill 47 Mixed backpropagation and L2
  172. Drill 48 Mixed dropout and a momentum update
  173. Drill 49 Mixed CNN dimensions softmax and loss
  174. Drill 50 Three step LSTM retention and replacement
  175. Drill 51 Inception concatenation and parameter count
  176. Drill 52 Projection shortcut