Author Topic: Q-Learning interesting Problem  (Read 614 times)

binharoun_7

  • *
  • Roomba
  • Posts: 1
Q-Learning interesting Problem
« on: November 14, 2013, 09:54:22 AM »
Hello everyone, i have just started to study Learning and see the possibilities of using Learning to solve my problem.

Problem: I am supposed to detect a certain combination of data, i have four matrices that acts as an input to my system, i have already categorised the inputs ( each input can either be Low (L) , or High (H) ). I need to detect certain types of input for example LLLH, LLHH, HHHH etc

NOTE:
1)LLLH means the first input in L, second input is L, third input is L and the fourth input is H!
2)I have labelled each type of input type as state, for example LLLL is state 1, LLLH is state 2, so on.

What i have studied in Learning is that most of the time you have one goal (only one state as a goal) which makes it easier for the agent to learn and create the Q-matrix from the R-matrix . Now in my problem i have many goal ( many states act as goal and need to be detected). I don't know how to design the states, how to create the Reward-matrix by having many goals and how the agent will learn. Can you please help me how can i use Learning in this kind of situation. Taking into account i have like 16 goals in 20+ states!

Freddy

  • *
  • Deep Thought
  • *********************
  • Posts: 5204
  • Mostly Harmless
Re: Q-Learning interesting Problem
« Reply #1 on: November 15, 2013, 04:42:50 PM »
Not something I know about. I can usually think of someone to ask but I draw a blank with this one.

Maybe you could try www.chatbots.org - the programmers tend to hang out there more.

I just wanted to say welcome too, so welcome :)

 

Welcome

Please login or register.



Login with username, password and session length
Speak, toy! Welcome to the age of chatting with your teddy bear
by Tyler (AI News )
May 24, 2015, 10:52:17 PM
Google patents 'creepy' cuddly toys
by Art (AI News )
May 24, 2015, 07:55:42 PM
Robots could soon take over the world and wipe out humanity says Martin Rees
by Art (AI News )
May 24, 2015, 07:54:15 PM
This Artificial Intelligence Pioneer Has a Few Concerns
by Tyler (AI News )
May 24, 2015, 04:50:32 PM
Home robot cleaner that can be remotely controlled
by Tyler (AI News )
May 24, 2015, 10:50:40 AM
Robots Need Your Brain Power To Get Smarter
by Tyler (AI News )
May 24, 2015, 04:50:55 AM
Neuroscientists Are Making an Artificial Brain for Everyone
by Tyler (AI News )
May 23, 2015, 04:50:44 PM
SuperUROP class of 2015 graduates with high accomplishment and promise
by Tyler (AI News )
May 23, 2015, 10:50:55 AM

Users Online

12 Guests, 0 Users

Most Online Today: 23. Most Online Ever: 208 (August 27, 2008, 08:26:54 AM)

Articles