Trjtdtk

Were there two appearances of Stan Lee?

What is the strongest case that can be made in favour of the UK regaining some control over fishing policy after Brexit?

Reverse the word in a string with the same order in javascript

Is it possible to measure lightning discharges as Nikola Tesla?

Python "triplet" dictionary?

Binary Numbers Magic Trick

Does a creature that is immune to a condition still make a saving throw?

Possible to set `foldexpr` using a function reference?

Historically, were women trained for obligatory wars? Or did they serve some other military function?

Pulling the rope with one hand is as heavy as with two hands?

Unexpected email from Yorkshire Bank

Can a creature tell when it has been affected by a Divination wizard's Portent?

Help, my Death Star suffers from Kessler syndrome!

Past Perfect Tense

Why does Bran Stark feel that Jon Snow "needs to know" about his lineage?

What does 「再々起」mean?

Need help understanding harmonic series and intervals

Minimum value of 4 digit number divided by sum of its digits

Why is the origin of “threshold” uncertain?

Why was Germany not as successful as other Europeans in establishing overseas colonies?

How to stop co-workers from teasing me because I know Russian?

Phrase for the opposite of "foolproof"

Does the EU Common Fisheries Policy cover British Overseas Territories?

Do I have to worry about players making “bad” choices on level up?

Use deep reinforcement learning with recursive actions?

Theano in deep learning researchKeyword/phrase extraction from Text using Deep Learning librariesCombining Neural Network with Reinforcement Learning in a Continuous SpaceAlphaGo (and other game programs using reinforcement-learning) without human databaseWhat is missing from the following Curriculum Learning implementation in a Deep Neural Net?Weights in neural networkDefining State Representation in Deep Q-LearningHandling actions with delayed effect (Reinforcement learning)Boundaries of Reinforcement LearningReinforcement learning - How to deal with varying number of actions which do number approximation

1

$begingroup$

Can we use recursive actions in deep reinforcement learning ? If yes how ?

For example, in a reassignment problem, for a task i the neural networks returns a node m, and then there will be some condition : If the node m is busy (If there's already a task j assigned to that node), then the task j will be reassigned to another node and so on until a vacant node will be returned (We suppose that there's always at least one vacant node).

deep-learning reinforcement-learning

share|improve this question

asked Apr 8 at 16:29

User732User732

61

$endgroup$

add a comment | 

1

$begingroup$

Can we use recursive actions in deep reinforcement learning ? If yes how ?

For example, in a reassignment problem, for a task i the neural networks returns a node m, and then there will be some condition : If the node m is busy (If there's already a task j assigned to that node), then the task j will be reassigned to another node and so on until a vacant node will be returned (We suppose that there's always at least one vacant node).

deep-learning reinforcement-learning

share|improve this question

asked Apr 8 at 16:29

User732User732

61

$endgroup$

add a comment | 

$begingroup$

Can we use recursive actions in deep reinforcement learning ? If yes how ?

For example, in a reassignment problem, for a task i the neural networks returns a node m, and then there will be some condition : If the node m is busy (If there's already a task j assigned to that node), then the task j will be reassigned to another node and so on until a vacant node will be returned (We suppose that there's always at least one vacant node).

deep-learning reinforcement-learning

share|improve this question

asked Apr 8 at 16:29

User732User732

61

$endgroup$

share|improve this question

asked Apr 8 at 16:29

User732User732

61

share|improve this question

asked Apr 8 at 16:29

User732User732

61

asked Apr 8 at 16:29

User732User732

61

asked Apr 8 at 16:29

User732User732

61

1 Answer
1

active

oldest

votes

0

$begingroup$

Your description doesn't seem so much like a "recurrent action" as a "failed to progress" situation.

The simplest ways to deal with possibility of a blocked/failed action are:

• Leave state and time step unchanged and pick next best action, without any new policy evaluations. This is effectively the approach used by many turn-based game playing RL, such as AlphaGo, although the rejection of non-valid actions is done using the game engine and not technically tried repeatedly in the environment, that does not make a big difference in all cases.

Or

• Update the time step, receive a reward (maybe zero) and try again on next time step. The state information might usefully include blocked actions (or they can be applied programmatically as in the first option). You would typically use this approach if there was a real cost to attempting the action (i.e. a negative reward for losing time) and/or it is possible from the state information to predict which actions could be blocked, and you want to include these factors in the agent's behaviour.

The second option is the approach you might use in a maze solver, if it could only sense walls by bumping in to them.

If the blocked actions are completely unpredictable from the agent's observations, then it doesn't matter too much which approach you use. However, if there is a cost for failing to act, then the second approach would give you more accurate action values in case you are using a value-based approach or need to assess the performance of your agent numerically.

share|improve this answer

edited Apr 8 at 19:39

answered Apr 8 at 18:58

Neil SlaterNeil Slater

17.9k33265

$endgroup$

• 
  
  
  

  
  

  
  
  
  
  
  

  
  $begingroup$
  I don't necessarily want the task to be assigned to only vacant nodes, meaning that reassigning the task j while reassigning task i could produce a more interesting scenario, that's while I talked about "recursive actions".
  $endgroup$
  – User732
  Apr 8 at 21:17
  

  
  
  
  


• 
  
  
  

  
  

  
  
  
  
  
  

  
  $begingroup$
  @User732: Sorry I don't understand. Maybe if you add some worked through examples in the question, it would help. The example you give so far matches this answer as far as I can see = "I want to try an action, and if there is some rule triggered where the action is not valid, I want to try a different action".
  $endgroup$
  – Neil Slater
  Apr 8 at 21:51
  

  
  
  
  

add a comment | 

Your Answer

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%2f48889%2fuse-deep-reinforcement-learning-with-recursive-actions%23new-answer', 'question_page');

);

Post as a guest

Name

Email

Required, but never shown

0

$begingroup$

Your description doesn't seem so much like a "recurrent action" as a "failed to progress" situation.

The simplest ways to deal with possibility of a blocked/failed action are:

• Leave state and time step unchanged and pick next best action, without any new policy evaluations. This is effectively the approach used by many turn-based game playing RL, such as AlphaGo, although the rejection of non-valid actions is done using the game engine and not technically tried repeatedly in the environment, that does not make a big difference in all cases.

Or

• Update the time step, receive a reward (maybe zero) and try again on next time step. The state information might usefully include blocked actions (or they can be applied programmatically as in the first option). You would typically use this approach if there was a real cost to attempting the action (i.e. a negative reward for losing time) and/or it is possible from the state information to predict which actions could be blocked, and you want to include these factors in the agent's behaviour.

The second option is the approach you might use in a maze solver, if it could only sense walls by bumping in to them.

If the blocked actions are completely unpredictable from the agent's observations, then it doesn't matter too much which approach you use. However, if there is a cost for failing to act, then the second approach would give you more accurate action values in case you are using a value-based approach or need to assess the performance of your agent numerically.

share|improve this answer

edited Apr 8 at 19:39

answered Apr 8 at 18:58

Neil SlaterNeil Slater

17.9k33265

$endgroup$

• 
  
  
  

  
  

  
  
  
  
  
  

  
  $begingroup$
  I don't necessarily want the task to be assigned to only vacant nodes, meaning that reassigning the task j while reassigning task i could produce a more interesting scenario, that's while I talked about "recursive actions".
  $endgroup$
  – User732
  Apr 8 at 21:17
  

  
  
  
  


• 
  
  
  

  
  

  
  
  
  
  
  

  
  $begingroup$
  @User732: Sorry I don't understand. Maybe if you add some worked through examples in the question, it would help. The example you give so far matches this answer as far as I can see = "I want to try an action, and if there is some rule triggered where the action is not valid, I want to try a different action".
  $endgroup$
  – Neil Slater
  Apr 8 at 21:51
  

  
  
  
  

add a comment | 

0

$begingroup$

Your description doesn't seem so much like a "recurrent action" as a "failed to progress" situation.

The simplest ways to deal with possibility of a blocked/failed action are:

• Leave state and time step unchanged and pick next best action, without any new policy evaluations. This is effectively the approach used by many turn-based game playing RL, such as AlphaGo, although the rejection of non-valid actions is done using the game engine and not technically tried repeatedly in the environment, that does not make a big difference in all cases.

Or

• Update the time step, receive a reward (maybe zero) and try again on next time step. The state information might usefully include blocked actions (or they can be applied programmatically as in the first option). You would typically use this approach if there was a real cost to attempting the action (i.e. a negative reward for losing time) and/or it is possible from the state information to predict which actions could be blocked, and you want to include these factors in the agent's behaviour.

The second option is the approach you might use in a maze solver, if it could only sense walls by bumping in to them.

If the blocked actions are completely unpredictable from the agent's observations, then it doesn't matter too much which approach you use. However, if there is a cost for failing to act, then the second approach would give you more accurate action values in case you are using a value-based approach or need to assess the performance of your agent numerically.

share|improve this answer

edited Apr 8 at 19:39

answered Apr 8 at 18:58

Neil SlaterNeil Slater

17.9k33265

$endgroup$

• 
  
  
  

  
  

  
  
  
  
  
  

  
  $begingroup$
  I don't necessarily want the task to be assigned to only vacant nodes, meaning that reassigning the task j while reassigning task i could produce a more interesting scenario, that's while I talked about "recursive actions".
  $endgroup$
  – User732
  Apr 8 at 21:17
  

  
  
  
  


• 
  
  
  

  
  

  
  
  
  
  
  

  
  $begingroup$
  @User732: Sorry I don't understand. Maybe if you add some worked through examples in the question, it would help. The example you give so far matches this answer as far as I can see = "I want to try an action, and if there is some rule triggered where the action is not valid, I want to try a different action".
  $endgroup$
  – Neil Slater
  Apr 8 at 21:51
  

  
  
  
  

add a comment | 

$begingroup$

Your description doesn't seem so much like a "recurrent action" as a "failed to progress" situation.

The simplest ways to deal with possibility of a blocked/failed action are:

• Leave state and time step unchanged and pick next best action, without any new policy evaluations. This is effectively the approach used by many turn-based game playing RL, such as AlphaGo, although the rejection of non-valid actions is done using the game engine and not technically tried repeatedly in the environment, that does not make a big difference in all cases.

Or

• Update the time step, receive a reward (maybe zero) and try again on next time step. The state information might usefully include blocked actions (or they can be applied programmatically as in the first option). You would typically use this approach if there was a real cost to attempting the action (i.e. a negative reward for losing time) and/or it is possible from the state information to predict which actions could be blocked, and you want to include these factors in the agent's behaviour.

The second option is the approach you might use in a maze solver, if it could only sense walls by bumping in to them.

If the blocked actions are completely unpredictable from the agent's observations, then it doesn't matter too much which approach you use. However, if there is a cost for failing to act, then the second approach would give you more accurate action values in case you are using a value-based approach or need to assess the performance of your agent numerically.

share|improve this answer

edited Apr 8 at 19:39

answered Apr 8 at 18:58

Neil SlaterNeil Slater

17.9k33265

$endgroup$

share|improve this answer

edited Apr 8 at 19:39

answered Apr 8 at 18:58

Neil SlaterNeil Slater

17.9k33265

answered Apr 8 at 18:58

Neil SlaterNeil Slater

17.9k33265

answered Apr 8 at 18:58

Neil SlaterNeil Slater

17.9k33265