User meTchaikovsky

49 votes

1 answer

54k views

Why KL divergence is non-negative?

asked Mar 18, 2018 at 10:43

15 votes

4 answers

10k views

What causes lasso to be unstable for feature selection?

asked Sep 8, 2018 at 8:53

9 votes

2 answers

13k views

Why does larger perplexity tend to produce clearer clusters in t-SNE?

asked Mar 28, 2019 at 7:38

8 votes

1 answer

520 views

Sparse linear regression 0-norm and 1-norm

asked Aug 28, 2018 at 1:22

6 votes

1 answer

653 views

Linear regression with $l_0$ regularization

asked Aug 27, 2018 at 4:33

6 votes

1 answer

128 views

A question about the inversion method

asked May 7, 2018 at 4:13

5 votes

1 answer

205 views

What can we learn from this visualization?

asked May 9, 2018 at 10:43

5 votes

2 answers

3k views

KL divergence invariant to affine transformation?

asked Apr 21, 2018 at 14:38

4 votes

1 answer

3k views

Data sparsity becomes a problem

asked Jul 10, 2018 at 6:08

4 votes

1 answer

892 views

Variational autoencoder obtain high likelihood but produce low quality sample?

asked Aug 4, 2018 at 23:28

4 votes

1 answer

1k views

How to continuously rotate a bivariate normal distribution by changing its correlation matrix?

asked Jan 11, 2019 at 6:28

3 votes

1 answer

516 views

Train an ensemble of neural networks on different datasets, what is the best way to scale the inputs?

asked Sep 15, 2019 at 2:35

3 votes

0 answers

29 views

Increase the number of samples when the PDF is invariant

asked Sep 27, 2018 at 6:03

3 votes

1 answer

1k views

Can we still use conjugate gradient descent when we don't know the Hessian matrix?

asked Oct 7, 2018 at 13:18

3 votes

1 answer

536 views

Feature filtering with LASSO and cross validation

asked Sep 3, 2018 at 10:31

2 votes

1 answer

1k views

What is $\lambda_{min}$ that makes all regression coefficients zero in LASSO regression [duplicate]

asked Sep 4, 2018 at 1:39

2 votes

0 answers

65 views

Select 3 features from 4000 correlated features

asked Sep 5, 2018 at 10:24

2 votes

1 answer

102 views

$\lambda \Vert k \Vert_0$ or $\Vert k \Vert_0 \leqslant n$

asked Sep 2, 2018 at 1:24

2 votes

0 answers

447 views

How do we train variational autoencoders in practice?

asked Aug 20, 2018 at 17:01

2 votes

1 answer

648 views

Joint distribution in latent Dirichlet model

asked Jul 12, 2018 at 2:13

2 votes

0 answers

2k views

Prove linear transformation of gaussian random vectors are still gaussian random vectors

asked May 30, 2018 at 11:22

2 votes

1 answer

3k views

How to perform batch training using L-BFGS?

asked Dec 24, 2018 at 10:57

2 votes

1 answer

63 views

How to differentiate through the optimization process in Model-Agnostic Meta-Learning?

asked Jan 7, 2019 at 8:41

2 votes

1 answer

3k views

Can t-SNE help feature selection?

asked Jan 8, 2019 at 11:35

2 votes

1 answer

71 views

Should I discard a feature whose max - min is smaller than 1e-5?

asked Jan 15, 2019 at 14:34

1 vote

0 answers

70 views

Why it is hard for `xgboost` to learn periodic functions?

asked Mar 10, 2019 at 8:23

1 vote

1 answer

807 views

Does it make sense to run lasso to select features for neural network training?

asked Nov 24, 2018 at 7:42

1 vote

2 answers

281 views

Can a neural network learned function $f$ be used for calculating its derivate $f'$?

asked Oct 5, 2018 at 2:24

1 vote

1 answer

4k views

Can PCA be used for detecting multicollinearity?

asked Sep 8, 2018 at 14:47

1 vote

1 answer

235 views

$\lambda_{\min}$ is $ \frac{1}{N}\max|\sum_{j}y^TX_j|$ rather than $\max|\sum_{j}y^TX_j|$?

asked Sep 11, 2018 at 4:58

Stack Exchange Network

35 Questions

Why KL divergence is non-negative?

What causes lasso to be unstable for feature selection?

Why does larger perplexity tend to produce clearer clusters in t-SNE?

Sparse linear regression 0-norm and 1-norm

Linear regression with $l_0$ regularization

A question about the inversion method

What can we learn from this visualization?

KL divergence invariant to affine transformation?

Data sparsity becomes a problem

Variational autoencoder obtain high likelihood but produce low quality sample?

How to continuously rotate a bivariate normal distribution by changing its correlation matrix?

Train an ensemble of neural networks on different datasets, what is the best way to scale the inputs?

Increase the number of samples when the PDF is invariant

Can we still use conjugate gradient descent when we don't know the Hessian matrix?

Feature filtering with LASSO and cross validation

What is $\lambda_{min}$ that makes all regression coefficients zero in LASSO regression [duplicate]

Select 3 features from 4000 correlated features

$\lambda \Vert k \Vert_0$ or $\Vert k \Vert_0 \leqslant n$

How do we train variational autoencoders in practice?

Joint distribution in latent Dirichlet model

Prove linear transformation of gaussian random vectors are still gaussian random vectors

How to perform batch training using L-BFGS?

How to differentiate through the optimization process in Model-Agnostic Meta-Learning?

Can t-SNE help feature selection?

Should I discard a feature whose max - min is smaller than 1e-5?

Why it is hard for `xgboost` to learn periodic functions?

Does it make sense to run lasso to select features for neural network training?

Can a neural network learned function $f$ be used for calculating its derivate $f'$?

Can PCA be used for detecting multicollinearity?

$\lambda_{\min}$ is $ \frac{1}{N}\max|\sum_{j}y^TX_j|$ rather than $\max|\sum_{j}y^TX_j|$?