0
$\begingroup$

I have a problem where I need to rearrange a particular user's mobile home screen icon layout. Let's say that the social media app usage of a user is high compared to other app usage. So I need the reinforcement algorithm to process this information and send back the instructions to the android operating system as to how the icons needs to be arranged. To address this problem, I have chosen three algorithms:

  1. Q-learning.
  2. State-Action-Reward-State-Action.
  3. Deep Deterministic Policy Gradient.

I have decided to first consider only Q-learning, so I am trying to understand the states, rewards, and the actions I need to pass in order to make this algorithm work.

The principles I have considered are:

– The environment is the mobile device operating system.
– Moving the apps up in the list depending on their usage can be the action where the app can be moved left, right, up or down.
– The reward can be a periodic reward, check if the user has rearranged an app which was given prominence in the list by the algorithm and receive a negative feedback if the user has rearranged the icon position or if it is in the same position receive a positive reward.

The initial challenge I am facing is to understand what inputs/states I need to pass into the algorithm and is there any reinforcement learning library I can use to mimic such an environment?
Are there any resources or papers I can use to solve this problem I am facing?

$\endgroup$
4
  • 2
    $\begingroup$ That's a lot of questions. We generally want a question to address a specific subject, rather than multiple subjects. Please edit your post accordingly. $\endgroup$ Apr 30 at 5:32
  • $\begingroup$ Do I need to edit the question more ?, I actually have edited the question to ask 1 question $\endgroup$ May 2 at 11:12
  • $\begingroup$ That seems ok. I have tried to clean it up further. What does "While the reward can be based" mean? The sentence seems incomplete. $\endgroup$ May 2 at 11:17
  • $\begingroup$ Thank I completed it as well $\endgroup$ May 2 at 14:25

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.