Hfrxdjt

Doesn't the system of the Supreme Court oppose justice?

How can I write humor as character trait?

Is there a RAID 0 Equivalent for RAM?

I found an audio circuit and I built it just fine, but I find it a bit too quiet. How do I amplify the output so that it is a bit louder?

What is Cash Advance APR?

Is this part of the description of the Archfey warlock's Misty Escape feature redundant?

Circuit Analysis: Obtaining Close Loop OP - AMP Transfer function

Change the color of a single dot in `ddot` symbol

When were female captains banned from Starfleet?

US tourist/student visa

What is the difference between lands and mana?

What (the heck) is a Super Worm Equinox Moon?

Microchip documentation does not label CAN buss pins on micro controller pinout diagram

A variation to the phrase "hanging over my shoulders"

What does Apple's new App Store requirement mean

What are some good ways to treat frozen vegetables such that they behave like fresh vegetables when stir frying them?

A Trivial Diagnosis

How to convince somebody that he is fit for something else, but not this job?

Is there any evidence that Cleopatra and Caesarion considered fleeing to India to escape the Romans?

Will number of steps recorded on FitBit/any fitness tracker add up distance in PokemonGo?

What is the English pronunciation of "pain au chocolat"?

Permission on Database

Does the reader need to like the PoV character?

Shouldn’t conservatives embrace universal basic income?

What activation function should I use for a specific regression problem?

How flexible is the link between objective function and output layer activation function?Technology stack for linear regression on (not so) large datasetParameterization regression of rotation angleAlternatives to linear activation function in regression tasks to limit the outputWhat is the Time Complexity of Linear Regression?Activation function vs Squashing functionUsing Keras to Predict a Function Following a Normal Distributionbest activation function for ensemble?Regression with -1,1 target range - Should we use a tanh activation in the last 1 unit dense layer?Images Score Regression only regresses to the average of the target values

1

$begingroup$

Which is better for regression problems create a neural net with tanh/sigmoid and exp(like) activations or ReLU and linear? Standard is to use ReLU but it's brute force solution that requires certain net size and I would like to avoid creating a very big net, also sigmoid is much more prefered but in my case regression will output values from range (0, 1e7)... maybe also sigmoid net with linear head will work? I am curious about your take on the subject.

machine-learning neural-network deep-learning regression activation-function

share|improve this question

edited 1 hour ago

Media

7,36762161

asked 4 hours ago

questerquester

313

$endgroup$

add a comment | 

1

$begingroup$

Which is better for regression problems create a neural net with tanh/sigmoid and exp(like) activations or ReLU and linear? Standard is to use ReLU but it's brute force solution that requires certain net size and I would like to avoid creating a very big net, also sigmoid is much more prefered but in my case regression will output values from range (0, 1e7)... maybe also sigmoid net with linear head will work? I am curious about your take on the subject.

machine-learning neural-network deep-learning regression activation-function

share|improve this question

edited 1 hour ago

Media

7,36762161

asked 4 hours ago

questerquester

313

$endgroup$

add a comment | 

$begingroup$

Which is better for regression problems create a neural net with tanh/sigmoid and exp(like) activations or ReLU and linear? Standard is to use ReLU but it's brute force solution that requires certain net size and I would like to avoid creating a very big net, also sigmoid is much more prefered but in my case regression will output values from range (0, 1e7)... maybe also sigmoid net with linear head will work? I am curious about your take on the subject.

machine-learning neural-network deep-learning regression activation-function

share|improve this question

edited 1 hour ago

Media

7,36762161

asked 4 hours ago

questerquester

313

$endgroup$

share|improve this question

edited 1 hour ago

Media

7,36762161

asked 4 hours ago

questerquester

313

share|improve this question

edited 1 hour ago

Media

7,36762161

asked 4 hours ago

questerquester

313

edited 1 hour ago

Media

7,36762161

edited 1 hour ago

Media

7,36762161

asked 4 hours ago

questerquester

313

asked 4 hours ago

questerquester

313

2 Answers
2

active

oldest

votes

1

$begingroup$

The issue with sigmoid and tanh activations is that their gradients saturate for extreme values of their arguments. This may occur if you do not normalize your inputs. If the learned weights of the unit are such that the gradient of its activation is close to zero, it will take longer for any updates to be reflected in the unit's weights. A final layer with no non-linearity will help you scale up your hidden layers' outputs.

In the end, the performance is application specific. You should try out both kinds of activations on a subset of your data and see which performs better.

share|improve this answer

answered 4 hours ago

hazrmardhazrmard

1264

$endgroup$

add a comment | 

0

$begingroup$

There are two points that have to be considered.

1. Take care of the output of your network. If that's a Real number and can take any value, you have to use linear activation as the output.


2. The inner activations highly depend on your task and the size of the network that you use. What I'm going to tell you is based on experience. If you don't have a very deep network, $tanh$ and $ReLU$ may not differ very much in convergence time. If you're using very deep networks, don't use $tahn$ at all. $ReLU$ is also not recommended in some contexts. You can employ $PReLU$ in very deep networks. It does not add too many parameters to learn. You can also use $leaky-ReLU$ in order to avoid dying ReLU problem which may occur.

Finally, about the other nonlinearity that you've referred; try not to use $Sigmoid$ due to that fact that it's expected value is not equal to zero but half. It's a bit statistical stuff, but you can consider it's roughly hard for a network to learn shifted weights.

share|improve this answer

edited 1 hour ago

answered 1 hour ago

MediaMedia

7,36762161

$endgroup$

add a comment | 

Your Answer

StackExchange.ifUsing("editor", function ()
return StackExchange.using("mathjaxEditing", function ()
StackExchange.MarkdownEditor.creationCallbacks.add(function (editor, postfix)
StackExchange.mathjaxEditing.prepareWmdForMathJax(editor, postfix, [["$", "$"], ["\\(","\\)"]]);
);
);
, "mathjax-editing");

StackExchange.ready(function()
var channelOptions =
tags: "".split(" "),
id: "557"
;
initTagRenderer("".split(" "), "".split(" "), channelOptions);

StackExchange.using("externalEditor", function()
// Have to fire editor after snippets, if snippets enabled
if (StackExchange.settings.snippets.snippetsEnabled)
StackExchange.using("snippets", function()
createEditor();
);

else
createEditor();

);

function createEditor()
StackExchange.prepareEditor(
heartbeatType: 'answer',
autoActivateHeartbeat: false,
convertImagesToLinks: false,
noModals: true,
showLowRepImageUploadWarning: true,
reputationToPostImages: null,
bindNavPrevention: true,
postfix: "",
imageUploader:
brandingHtml: "Powered by u003ca class="icon-imgur-white" href="https://imgur.com/"u003eu003c/au003e",
contentPolicyHtml: "User contributions licensed under u003ca href="https://creativecommons.org/licenses/by-sa/3.0/"u003ecc by-sa 3.0 with attribution requiredu003c/au003e u003ca href="https://stackoverflow.com/legal/content-policy"u003e(content policy)u003c/au003e",
allowUrls: true
,
onDemand: true,
discardSelector: ".discard-answer"
,immediatelyShowMarkdownHelp:true
);



);

draft saved

draft discarded

Sign up or log in

StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);

Sign up using Google

Sign up using Facebook

Sign up using Email and Password

Post as a guest

Name

Email

Required, but never shown

StackExchange.ready(
function ()
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fdatascience.stackexchange.com%2fquestions%2f47751%2fwhat-activation-function-should-i-use-for-a-specific-regression-problem%23new-answer', 'question_page');

);

Post as a guest

Name

Email

Required, but never shown

1

$begingroup$

The issue with sigmoid and tanh activations is that their gradients saturate for extreme values of their arguments. This may occur if you do not normalize your inputs. If the learned weights of the unit are such that the gradient of its activation is close to zero, it will take longer for any updates to be reflected in the unit's weights. A final layer with no non-linearity will help you scale up your hidden layers' outputs.

In the end, the performance is application specific. You should try out both kinds of activations on a subset of your data and see which performs better.

share|improve this answer

answered 4 hours ago

hazrmardhazrmard

1264

$endgroup$

add a comment | 

1

$begingroup$

The issue with sigmoid and tanh activations is that their gradients saturate for extreme values of their arguments. This may occur if you do not normalize your inputs. If the learned weights of the unit are such that the gradient of its activation is close to zero, it will take longer for any updates to be reflected in the unit's weights. A final layer with no non-linearity will help you scale up your hidden layers' outputs.

In the end, the performance is application specific. You should try out both kinds of activations on a subset of your data and see which performs better.

share|improve this answer

answered 4 hours ago

hazrmardhazrmard

1264

$endgroup$

add a comment | 

$begingroup$

The issue with sigmoid and tanh activations is that their gradients saturate for extreme values of their arguments. This may occur if you do not normalize your inputs. If the learned weights of the unit are such that the gradient of its activation is close to zero, it will take longer for any updates to be reflected in the unit's weights. A final layer with no non-linearity will help you scale up your hidden layers' outputs.

In the end, the performance is application specific. You should try out both kinds of activations on a subset of your data and see which performs better.

share|improve this answer

answered 4 hours ago

hazrmardhazrmard

1264

$endgroup$

share|improve this answer

answered 4 hours ago

hazrmardhazrmard

1264

answered 4 hours ago

hazrmardhazrmard

1264

answered 4 hours ago

hazrmardhazrmard

1264

0

$begingroup$

There are two points that have to be considered.

1. Take care of the output of your network. If that's a Real number and can take any value, you have to use linear activation as the output.


2. The inner activations highly depend on your task and the size of the network that you use. What I'm going to tell you is based on experience. If you don't have a very deep network, $tanh$ and $ReLU$ may not differ very much in convergence time. If you're using very deep networks, don't use $tahn$ at all. $ReLU$ is also not recommended in some contexts. You can employ $PReLU$ in very deep networks. It does not add too many parameters to learn. You can also use $leaky-ReLU$ in order to avoid dying ReLU problem which may occur.

Finally, about the other nonlinearity that you've referred; try not to use $Sigmoid$ due to that fact that it's expected value is not equal to zero but half. It's a bit statistical stuff, but you can consider it's roughly hard for a network to learn shifted weights.

share|improve this answer

edited 1 hour ago

answered 1 hour ago

MediaMedia

7,36762161

$endgroup$

add a comment | 

0

$begingroup$

There are two points that have to be considered.

1. Take care of the output of your network. If that's a Real number and can take any value, you have to use linear activation as the output.


2. The inner activations highly depend on your task and the size of the network that you use. What I'm going to tell you is based on experience. If you don't have a very deep network, $tanh$ and $ReLU$ may not differ very much in convergence time. If you're using very deep networks, don't use $tahn$ at all. $ReLU$ is also not recommended in some contexts. You can employ $PReLU$ in very deep networks. It does not add too many parameters to learn. You can also use $leaky-ReLU$ in order to avoid dying ReLU problem which may occur.

Finally, about the other nonlinearity that you've referred; try not to use $Sigmoid$ due to that fact that it's expected value is not equal to zero but half. It's a bit statistical stuff, but you can consider it's roughly hard for a network to learn shifted weights.

share|improve this answer

edited 1 hour ago

answered 1 hour ago

MediaMedia

7,36762161

$endgroup$

add a comment | 

$begingroup$

There are two points that have to be considered.

1. Take care of the output of your network. If that's a Real number and can take any value, you have to use linear activation as the output.


2. The inner activations highly depend on your task and the size of the network that you use. What I'm going to tell you is based on experience. If you don't have a very deep network, $tanh$ and $ReLU$ may not differ very much in convergence time. If you're using very deep networks, don't use $tahn$ at all. $ReLU$ is also not recommended in some contexts. You can employ $PReLU$ in very deep networks. It does not add too many parameters to learn. You can also use $leaky-ReLU$ in order to avoid dying ReLU problem which may occur.

Finally, about the other nonlinearity that you've referred; try not to use $Sigmoid$ due to that fact that it's expected value is not equal to zero but half. It's a bit statistical stuff, but you can consider it's roughly hard for a network to learn shifted weights.

share|improve this answer

edited 1 hour ago

answered 1 hour ago

MediaMedia

7,36762161

$endgroup$

share|improve this answer

edited 1 hour ago

answered 1 hour ago

MediaMedia

7,36762161

answered 1 hour ago

MediaMedia

7,36762161

answered 1 hour ago

MediaMedia

7,36762161