User user3658307

397 votes

11 answers

173k views

Explaining to laypeople why bootstrapping works

Alan H.

4,949

asked Apr 8, 2012 at 21:04

337 votes

8 answers

131k views

Why is Euclidean distance not a good metric in high dimensions?

machine-learning clustering distance-functions metric high-dimensional

teaLeef

3,587

asked May 18, 2014 at 17:50

153 votes

8 answers

155k views

What is the difference between off-policy and on-policy learning?

machine-learning reinforcement-learning artificial-intelligence

cgo

8,025

asked Dec 2, 2015 at 14:21

132 votes

19 answers

107k views

Books for self-studying time series analysis?

time-series self-study references

CuriousMind

2,173

asked Jan 3, 2012 at 1:22

120 votes

6 answers

172k views

What loss function for multi-class, multi-label classification tasks in neural networks?

neural-networks python loss-functions keras cross-entropy

aKzenT

1,301

asked Apr 17, 2016 at 14:28

89 votes

11 answers

10k views

Why should I be Bayesian when my model is wrong?

bayesian modeling philosophical misspecification

Guillaume Dehaene

2,167

asked Apr 20, 2017 at 15:28

72 votes

11 answers

11k views

Is there any mathematical basis for the Bayesian vs frequentist debate?

probability bayesian frequentist philosophical kolmogorov-axioms

Chill2Macht

5,757

asked Aug 18, 2016 at 3:40

71 votes

4 answers

24k views

Assumptions regarding bootstrap estimates of uncertainty

bootstrap uncertainty

user4733

2,514

asked May 24, 2011 at 19:53

71 votes

5 answers

76k views

What's the difference between momentum based gradient descent and Nesterov's accelerated gradient descent?

optimization gradient-descent

applecider

1,185

asked Nov 3, 2015 at 6:51

62 votes

2 answers

15k views

Optimal number of folds in $K$-fold cross-validation: is leave-one-out CV always the best choice?

cross-validation bias-variance-tradeoff

Amelio Vazquez-Reina

18k

asked Jun 12, 2013 at 13:24

55 votes

6 answers

7k views

What are the main theorems in Machine (Deep) Learning?

machine-learning deep-learning mathematical-statistics

user188529

asked Jan 6, 2018 at 15:37

53 votes

3 answers

14k views

Consider the sum of $n$ uniform distributions on $[0,1]$, or $Z_n$. Why does the cusp in the PDF of $Z_n$ disappear for $n \geq 3$?

normal-distribution mathematical-statistics uniform-distribution central-limit-theorem

tetragrammaton

1,366

asked Oct 30, 2012 at 0:09

48 votes

3 answers

27k views

Intuitive difference between hidden Markov models and conditional random fields

machine-learning hidden-markov-model natural-language conditional-random-field

user1343318

1,251

asked May 5, 2013 at 22:58

36 votes

3 answers

9k views

What's the relation between hierarchical models, neural networks, graphical models, bayesian networks?

causality neural-networks multilevel-analysis graphical-model

cespinoza

762

asked Nov 13, 2010 at 5:43

34 votes

4 answers

54k views

Maximum Mean Discrepancy (distance distribution)

machine-learning distributions distance feature-engineering domain-adaptation

Mahsa

441

asked Apr 28, 2017 at 15:45

32 votes

1 answer

22k views

What is the standard error of the sample standard deviation?

sampling standard-deviation standard-error

Remi.b

4,742

asked Jun 11, 2015 at 15:42

32 votes

1 answer

7k views

What are variational autoencoders and to what learning tasks are they used?

machine-learning bayesian deep-learning autoencoders variational-bayes

user188529

asked Jan 6, 2018 at 14:31

30 votes

3 answers

29k views

Asymptotic distribution of sample variance of non-normal sample

mathematical-statistics variance central-limit-theorem asymptotics

Alecos Papadopoulos

54k

asked Jul 1, 2014 at 0:38

30 votes

1 answer

8k views

PCA, LDA, CCA, and PLS

pca discriminant-analysis partial-least-squares canonical-correlation

carlosdc

3,145

asked Feb 20, 2012 at 2:13

29 votes

3 answers

9k views

Connection between Fisher metric and the relative entropy

mathematical-statistics kullback-leibler fisher-information

Kumara

577

asked Mar 2, 2013 at 12:06

23 votes

3 answers

9k views

Does differential geometry have anything to do with statistics?

mathematical-statistics information-geometry

LaTeXFan

1,106

asked Feb 16, 2014 at 4:28

23 votes

1 answer

7k views

Why is the cost function of neural networks non-convex?

machine-learning neural-networks optimization loss-functions convex

Luca

4,450

asked May 23, 2017 at 15:43

21 votes

4 answers

19k views

Calculating required sample size, precision of variance estimate?

estimation random-variable variance sampling sample-size

Abe

3,631

asked Feb 9, 2011 at 0:48

18 votes

1 answer

3k views

What are some distributions over the probability simplex?

distributions multinomial-distribution compositional-data

singelton

662

asked Aug 4, 2012 at 20:57

15 votes

4 answers

6k views

Estimate the Kullback–Leibler (KL) divergence with Monte Carlo

kullback-leibler

frelk

1,157

asked May 21, 2017 at 20:49

14 votes

2 answers

13k views

Pros and cons of weight normalization vs batch normalization

machine-learning neural-networks deep-learning conv-neural-network batch-normalization

doodledood

295

asked Sep 24, 2017 at 20:04

13 votes

3 answers

716 views

Using information geometry to define distances and volumes…useful?

model-selection information-geometry

Yaroslav Bulatov

5,403

asked Aug 12, 2010 at 23:53

13 votes

2 answers

7k views

Anchoring Faster RCNN

deep-learning computer-vision

BadProgrammer

487

asked Mar 6, 2017 at 23:30

13 votes

2 answers

9k views

What are the advantages of using a Bayesian neural network

bayesian neural-networks bayesian-network

fishiwhj

432

asked Mar 16, 2015 at 1:27

12 votes

1 answer

3k views

Covariance matrix for Gaussian Process and Wishart distribution

machine-learning normal-distribution covariance wishart-distribution nonparametric-bayes

steadyfish

1,802

asked Mar 31, 2013 at 14:41

Stack Exchange Network

current community

your communities

more stack exchange communities

59 Bookmarks

Explaining to laypeople why bootstrapping works

Why is Euclidean distance not a good metric in high dimensions?

What is the difference between off-policy and on-policy learning?

Books for self-studying time series analysis?

What loss function for multi-class, multi-label classification tasks in neural networks?

Why should I be Bayesian when my model is wrong?

Is there any mathematical basis for the Bayesian vs frequentist debate?

Assumptions regarding bootstrap estimates of uncertainty

What's the difference between momentum based gradient descent and Nesterov's accelerated gradient descent?

Optimal number of folds in $K$-fold cross-validation: is leave-one-out CV always the best choice?

What are the main theorems in Machine (Deep) Learning?

Consider the sum of $n$ uniform distributions on $[0,1]$, or $Z_n$. Why does the cusp in the PDF of $Z_n$ disappear for $n \geq 3$?

Intuitive difference between hidden Markov models and conditional random fields

What's the relation between hierarchical models, neural networks, graphical models, bayesian networks?

Maximum Mean Discrepancy (distance distribution)

What is the standard error of the sample standard deviation?

What are variational autoencoders and to what learning tasks are they used?

Asymptotic distribution of sample variance of non-normal sample

PCA, LDA, CCA, and PLS

Connection between Fisher metric and the relative entropy

Does differential geometry have anything to do with statistics?

Why is the cost function of neural networks non-convex?

Calculating required sample size, precision of variance estimate?

What are some distributions over the probability simplex?

Estimate the Kullback–Leibler (KL) divergence with Monte Carlo

Pros and cons of weight normalization vs batch normalization

Using information geometry to define distances and volumes…useful?

Anchoring Faster RCNN

What are the advantages of using a Bayesian neural network

Covariance matrix for Gaussian Process and Wishart distribution