rcpinto
  • Member for 7 years, 11 months
  • Last seen more than a month ago
The meaning of discount factor on reinforcement learning
6 votes

The discount factor does not represent the likelihood to reach the state $s′ $from the state $s$. That would be $p(s'|s,a)$, which is not used in Q-Learning, since it is model-free (only model-based ...

View answer
What is the difference between apprenticeship and imitation learning?
Accepted answer
4 votes

In general, yes, they are the same thing, which means to learn from demonstration (LfD). But, usually, apprenticeship learning is mentioned in the context of "Apprenticeship learning via inverse ...

View answer
Is it possible to solve the Mountain Car reinforcement learning task with linear Q-Learning using the state as direct input?
3 votes

I just found out the answer and it's actually pretty simple: while there is a good linear policy for the mountain car task, the value function itself is non-linear. The state space of this task is ...

View answer
Disadvantages to using simple step functions for activation in neural networks?
3 votes

The disadvantage would be that your fitness landscape would look less smooth (a small change in your weights may result in a big change in the output or no change at all). This may affect the ...

View answer
Which matrix of Q values is being used here?
Accepted answer
2 votes

Reinforcement learning works by bootstrapping, which means that Q values are updated based on previous estimates. Also, MDP's don't generate Q values, just rewards (Q values are long-term estimates of ...

View answer
Lamarckian and Genetic algorithm
2 votes

In fact, you're talking about Memetic Algorithms. In the case of a neural network, supposing you have training samples, the individual learning may be done through an iteration of backpropagation. If ...

View answer
The complexity of the algorithm with loops
Accepted answer
1 votes

Your "stuff" will get executed $N(N - 1)/2 = 0.5N^2 - 0.5N$ times. When analyzing the asymptotic complexity, only the highest order term is kept, and multiplicative constants are removed, leaving you ...

View answer
Reinforcement Learning - Q Learning
Accepted answer
1 votes

At t=1, you just took action a = d at state s = S and ended in s' = S again with 0 reward. So: $Q(S,d) = (1 - 1/2)(1) + 1/2[0 + (1)(1)] = 1$ ($maxQ(s',a')$ is the maximum value you can get from state ...

View answer
How to choose proper activation functions for hidden and output layers of a perceptron neural network?
1 votes

Input: it is not common to use activation functions in the input layer. Just rescale your data to the [-1;1] interval; Hidden: the $tanh$ activation used to be the most popular. Now, this role has ...

View answer
weights in a simple neural network
1 votes

It's just that the logistic function reaches its saturated regime around +-4.6, as in the image. So, any absolute values of x much lower than this will give you results around 0.5. Increasing the ...

View answer