User user3658307

50 votes

Accepted

Deriving the KL divergence loss for VAEs

answered Oct 4, 2018 at 6:24

8 votes

Why is the log derivative estimator considered of large variance?

answered Aug 18, 2019 at 18:43

6 votes

Bayesian nonparametric answer to deep learning?

answered Oct 10, 2018 at 7:21

5 votes

Regression as mutual information minimization

answered Jan 12, 2019 at 18:12

5 votes

Accepted

On Yolo, and its loss function

answered Jul 7, 2018 at 22:51

5 votes

Could anyone explain the terms "Hypothesis space" "sample space" "parameter space" "feature space in machine learning with one concrete example?

answered Aug 19, 2019 at 18:15

5 votes

ADAM Gradient descent oscillates close to minimum

answered Mar 14, 2019 at 3:09

5 votes

Accepted

Generating vector image from a hand drawn picture. Machine Learning

answered Mar 20, 2019 at 23:19

5 votes

Object detection : is deep learning the only way to go?

answered Jul 17, 2019 at 4:35

4 votes

Accepted

Ornstein-Uhlenbeck process

answered Aug 23, 2019 at 3:39

4 votes

Accepted

Is the first principal component the one with the largest eigenvalue and how to convert it to explained variance?

answered Aug 15, 2019 at 6:16

4 votes

How meaningful is the connection between MLE and cross entropy in deep learning?

answered Mar 26, 2018 at 5:55

4 votes

Accepted

What is a good loss function for object localisation and classification using a cnn?

answered Jul 7, 2018 at 22:03

4 votes

Accepted

covariance to correlation is like mutual information to --?

answered Oct 24, 2018 at 21:14

3 votes

How to choose an appropriate variational distribution?

answered Jan 14, 2019 at 3:59

3 votes

Accepted

Linear regression with error dispersion dependent on the independent variable

answered Feb 22, 2019 at 5:10

3 votes

Understanding the YOLO algorithm

answered Jul 7, 2018 at 23:39

2 votes

In VAEs, why don't we just use a fixed variance for the z distribution?

answered Oct 4, 2018 at 16:14

2 votes

Are analytical derivatives unambiguously superior to numerical derivatives in GMM?

answered Jul 17, 2018 at 22:10

2 votes

Accepted

Where's wrong in my reasoning behind upper bound for reconstruction error?

answered Mar 2, 2019 at 3:43

2 votes

Accepted

Choosing the number of labels in a multiclass classification problem

answered Sep 28, 2018 at 19:49

2 votes

Training an Object Detection Model Using with Artificial Data from Video Games

answered Aug 31, 2019 at 2:46

2 votes

Accepted

Why are exogenous variables not used in inference/recognition networks?

answered Jul 23, 2019 at 15:22

2 votes

What are the theoretical/practical reasons to use normal distribution to initialize the weights in Neural Networks?

answered Aug 15, 2019 at 3:03

2 votes

Accepted

What are the shortcomings of calculating the loss in pixel space vs. feature space

answered Apr 3, 2019 at 19:13

1 vote

What is the difference between having input parameters that are independent/dependent in ML tasks?

answered Jun 6, 2019 at 4:12

1 vote

Is it possible to obtain learn specific information about a distribution from a GAN?

answered Sep 6, 2019 at 6:36

1 vote

Accepted

Is this intuition behind understanding what the weights of a plain feedforward neural network are representing correct?

answered Feb 29, 2020 at 0:12

1 vote

Diversity of generated samples in VAE

answered Dec 16, 2018 at 21:05

1 vote

Accepted

Understand VAE with VamPrior

answered Dec 17, 2018 at 1:07

Stack Exchange Network

34 Answers

Deriving the KL divergence loss for VAEs

Why is the log derivative estimator considered of large variance?

Bayesian nonparametric answer to deep learning?

Regression as mutual information minimization

On Yolo, and its loss function

Could anyone explain the terms "Hypothesis space" "sample space" "parameter space" "feature space in machine learning with one concrete example?

ADAM Gradient descent oscillates close to minimum

Generating vector image from a hand drawn picture. Machine Learning

Object detection : is deep learning the only way to go?

Ornstein-Uhlenbeck process

Is the first principal component the one with the largest eigenvalue and how to convert it to explained variance?

How meaningful is the connection between MLE and cross entropy in deep learning?

What is a good loss function for object localisation and classification using a cnn?

covariance to correlation is like mutual information to --?

How to choose an appropriate variational distribution?

Linear regression with error dispersion dependent on the independent variable

Understanding the YOLO algorithm

In VAEs, why don't we just use a fixed variance for the z distribution?

Are analytical derivatives unambiguously superior to numerical derivatives in GMM?

Where's wrong in my reasoning behind upper bound for reconstruction error?

Choosing the number of labels in a multiclass classification problem

Training an Object Detection Model Using with Artificial Data from Video Games

Why are exogenous variables not used in inference/recognition networks?

What are the theoretical/practical reasons to use normal distribution to initialize the weights in Neural Networks?

What are the shortcomings of calculating the loss in pixel space vs. feature space

What is the difference between having input parameters that are independent/dependent in ML tasks?

Is it possible to obtain learn specific information about a distribution from a GAN?

Is this intuition behind understanding what the weights of a plain feedforward neural network are representing correct?

Diversity of generated samples in VAE

Understand VAE with VamPrior