User Green Falcon

47 votes

Accepted

Does gradient descent always converge to an optimum?

answered Nov 9, 2017 at 17:56

28 votes

Keras difference beetween val_loss and loss during training

answered Nov 30, 2017 at 20:03

28 votes

Why do convolutional neural networks work?

answered Jan 4, 2018 at 16:04

28 votes

How to set the number of neurons and layers in neural networks

answered Jan 15, 2018 at 11:59

20 votes

Accepted

Can the number of epochs influence overfitting?

answered Feb 7, 2018 at 14:00

18 votes

Dropout on which layers of LSTM?

answered Sep 13, 2018 at 13:35

17 votes

What are the pros and cons of Keras and TFLearn?

answered Dec 2, 2017 at 11:21

16 votes

Accepted

How to add non-image features along side images as the input of CNNs

answered May 8, 2018 at 15:17

15 votes

Accepted

What is LSTM, BiLSTM and when to use them?

answered Dec 14, 2017 at 6:06

15 votes

Accepted

cosine_similarity returns matrix instead of single value

answered Jan 15, 2018 at 13:26

14 votes

Accepted

When are weights updated in CNN?

answered Feb 3, 2018 at 15:42

14 votes

What are the differences between Convolutional1D, Convolutional2D, and Convolutional3D?

answered May 6, 2019 at 7:14

13 votes

Should we apply normalization to test data as well?

answered Feb 8, 2018 at 17:07

13 votes

Accepted

What is the input size of Alex net

answered Mar 19, 2018 at 6:01

13 votes

Why is ReLU used as an activation function?

answered Jan 10, 2018 at 13:32

12 votes

How to maximize recall?

answered Mar 9, 2018 at 16:43

12 votes

How does Sigmoid activation work in multi-class classification problems

answered Oct 6, 2018 at 19:55

11 votes

Accepted

The differences between SVM and Logistic Regression

answered May 10, 2018 at 2:00

11 votes

Is there any proven disadvantage of transfer learning for CNNs?

answered Apr 26, 2018 at 12:11

10 votes

Accepted

Does Tensorflow support a Decision Tree Classifier?

answered Feb 20, 2018 at 12:38

10 votes

Accepted

Beginner math books for Machine Learning

answered Jan 10, 2018 at 3:55

10 votes

Accepted

Input normalization for ReLu?

answered Dec 20, 2017 at 15:23

9 votes

Accepted

Are CNNs insensitive to rotations and shifts in images?

answered Mar 25, 2018 at 10:47

9 votes

What is difference between Fully Connected layer and Bilinear layer in CNN?

answered May 14, 2018 at 13:19

9 votes

What are the cases where it is fine to initialize all weights to zero

answered Apr 29, 2018 at 5:28

9 votes

Accepted

Batch normalization vs batch size

answered Nov 29, 2018 at 20:32

8 votes

Accepted

Why does the naive bayes algorithm make the naive assumption that features are independent to each other?

answered Sep 2, 2018 at 11:40

8 votes

Accepted

Is there a person class in ImageNet? Are there any classes related to humans?

answered Feb 11, 2018 at 8:40

8 votes

Are there any differences between Recurrent Neural Networks and Residual Neural Networks?

answered Dec 28, 2017 at 9:56

8 votes

In supervised learning, why is it bad to have correlated features?

answered Nov 7, 2017 at 16:01

Stack Exchange Network

346 Answers

Does gradient descent always converge to an optimum?

Keras difference beetween val_loss and loss during training

Why do convolutional neural networks work?

How to set the number of neurons and layers in neural networks

Can the number of epochs influence overfitting?

Dropout on which layers of LSTM?

What are the pros and cons of Keras and TFLearn?

How to add non-image features along side images as the input of CNNs

What is LSTM, BiLSTM and when to use them?

cosine_similarity returns matrix instead of single value

When are weights updated in CNN?

What are the differences between Convolutional1D, Convolutional2D, and Convolutional3D?

Should we apply normalization to test data as well?

What is the input size of Alex net

Why is ReLU used as an activation function?

How to maximize recall?

How does Sigmoid activation work in multi-class classification problems

The differences between SVM and Logistic Regression

Is there any proven disadvantage of transfer learning for CNNs?

Does Tensorflow support a Decision Tree Classifier?

Beginner math books for Machine Learning

Input normalization for ReLu?

Are CNNs insensitive to rotations and shifts in images?

What is difference between Fully Connected layer and Bilinear layer in CNN?

What are the cases where it is fine to initialize all weights to zero

Batch normalization vs batch size

Why does the naive bayes algorithm make the naive assumption that features are independent to each other?

Is there a person class in ImageNet? Are there any classes related to humans?

Are there any differences between Recurrent Neural Networks and Residual Neural Networks?

In supervised learning, why is it bad to have correlated features?