Reinforcement Learning - III


CSE 440: Introduction to Artificial Intelligence

Vishnu Boddeti


Content Credits: CMU AI, http://ai.berkeley.edu

Today

  • Approximate $Q$-Learning
  • Applications of RL

Reinforcement Learning

  • Still assume a Markov decision process (MDP):
    • A set of states $s \in S$
    • A set of states per state $A$
    • A model $T(s,a,s')$
    • A reward function $R(s,a,s')$
  • Still looking for a policy $\pi(s)$
  • New twist: do not know $T$ or $R$
  • Big idea: Compute all averages over $T$ using sample outcomes

Recap: $Q$-Learning

  • $Q$-Learning: sample-based $Q$-value iteration
  • $$\begin{equation} Q_{k+1}(s,a) \leftarrow \sum_{s'}T(s,a,s')[R(s,a,s') + \gamma \max_{a'}Q_k(s',a')] \end{equation}$$
  • Learn $Q(s,a)$ values as you go
    • Receive a sample $(s,a,s',r)$
    • Consider your old estimate: $Q(s,a)$
    • Consider your new sample estimate:
    • $$\begin{equation} sample = R(s,a,s') + \gamma \max_{a'}Q(s',a') \end{equation}$$
    • Incorporate the new estimate into a running average:
    • $$\begin{equation} Q(s,a) \leftarrow (1-\alpha)Q(s,a) + (\alpha) sample \end{equation}$$

Approximate $Q$-Learning

Generalizing Across States

  • Basic $Q$-Learning keeps a table of all q-values

  • In realistic situations, we cannot possibly learn about every single state.
    • Too many states to visit them all in training
    • Too many states to hold the q-tables in memory

  • Instead, we want to generalize:
    • Learn about some small number of training states from experience
    • Generalize that experience to new, similar situations
    • This is a fundamental idea in machine learning, and we will see it over and over again

Example: Pacman

Let us say we discover through experience that this state is bad:
In naïve $q$-learning, we know nothing about this state:

Feature-Based Representations

  • Solution: describe a state using a vector of features (properties)
    • Features are functions from states to real numbers (often 0/1) that capture important properties of the state
    • Example features:
      • Distance to closest ghost
      • Distance to closest dot
      • Number of ghosts
      • $\frac{1}{(\mbox{dist to dots})^2}$
      • Is Pacman in a tunnel? (0/1)
      • $\dots$ etc.
      • Is it the exact state on this slide?
    • Can also describe a $q$-state (s, a) with features (e.g. action moves closer to food)

Linear Value Functions

  • Using a feature representation, we can write a $q$ function (or value function) for any state using a few weights:
  • $$\begin{equation} \begin{aligned} V(s) = w_1f_1(s) + w_2f_2(s) + \dots + w_nf_n(s) \nonumber \\ Q(s,a) = w_1f_1(s,a) + w_2f_2(s,a) + \dots + w_nf_n(s,a) \nonumber \\ \end{aligned} \end{equation}$$
  • Advantage: our experience is summed up in a few powerful numbers
  • Disadvantage: states may share features but actually be very different in value.

Approximate $Q$-Learning

$Q(s,a) = w_1f_1(s,a) + w_2f_2(s,a) + \dots + w_nf_n(s,a)$
  • Q-learning with linear Q-functions:
  • $$\begin{equation} transition = (s,a,r,s') \end{equation}$$ $$\begin{equation} difference = [r + \gamma \max_{a'}Q(s',a')] - Q(s,a) \end{equation}$$ $$\begin{equation} Q(s,a) \leftarrow Q(s,a) + \alpha[difference] \mbox{ Exact Q's} \end{equation}$$ $$\begin{equation} w_i \leftarrow w_i + \alpha[difference]f_i(s,a) \mbox{ Approximate Q's} \end{equation}$$
  • Intuitive interpretation:
    • Adjust weights of active features
    • E.g., if something unexpectedly bad happens, blame the features that were on: do not prefer all states with that state's features
  • Formal justification: online least squares

Example $Q$-Pacman

$Q$-Learning and Least Squares

Linear Approximation: Regression

Prediction: $\hat{y} = w_0 + w_1f_1(x)$

Optimization: Least Squares

$$\begin{equation} \mbox{total error } = \sum_i (y_i-\hat{y}_i)^2 = \sum_i \left(y_i - \sum_k w_kf_k(x)\right)^2 \end{equation}$$
Prediction: $\hat{y} = w_0 + w_1f_1(x)$

Minimizing Error

  • Imagine we had only one point x, with features f(x), target value y, and weights w:
    $$\begin{equation} error(w) = \frac{1}{2}\left(y-\sum_k w_kfk(x)\right)^2 \end{equation}$$ $$\begin{equation} \frac{\partial error(w)}{\partial w_m} = -\left(y-\sum_k w_kfk(x)\right)f_m(x) \end{equation}$$ $$\begin{equation} w_m \leftarrow w_m + \alpha\left(y-\sum_k w_kfk(x)\right)f_m(x) \end{equation}$$
  • Approximate q update explained:
  • $$\begin{equation} w_m \leftarrow w_m + \alpha[\underbrace{r + \gamma\max_{a}Q(s',a')}_{"target"}-\underbrace{Q(s,a)}_{"prediction"}]f_m(s,a) \nonumber \end{equation}$$

Applications

RL: Helicopter Flight

RL: Learning Locomotion

RL: Learning Soccer

RL: Learning Manipulation

RL: NASA SUPERBall

RL: In-Hand Manipulation

Conclusion

  • We are done with Part I: Search and Planning
  • We have seen how AI methods can solve problems in:
    • Search
    • Constraint Satisfaction Problems
    • Optimization
    • Markov Decision Problems
    • Reinforcement Learning
  • Next Up: Part II: Uncertainty

Q & A

Image
XKCD