Download presentation
Presentation is loading. Please wait.
Published byLindsay Gregory Modified over 9 years ago
1
REINFORCEMENT LEARNING LEARNING TO PERFORM BEST ACTIONS BY REWARDS Tayfun Gürel
2
ROADMAP: Introduction to the problem Introduction to the problem Models of Optimal Behavior Models of Optimal Behavior An Immediate Reward Example and Solution An Immediate Reward Example and Solution Markov Decision Process Markov Decision Process Model Free Methods Model Free Methods Model Based Methods Model Based Methods Partially Observable Markov Decision Process Partially Observable Markov Decision Process Applications Applications Conclusion Conclusion
3
RL brings a way of programming agents by reward and punishment without specifying how the task is to be achieved. (Kaelbling,1996) Based on trial-error interactions A set of problems rather than a set of techniques
4
The standard reinforcement-learning model i: input r: reward s: state a: action
5
The reinforcement learning model consists of : a discrete set of environment states: S ; a discrete set of environment states: S ; a discrete set of agent actions A ; and a discrete set of agent actions A ; and a set of scalar reinforcement signals; typically {0,1}, or the real numbers a set of scalar reinforcement signals; typically {0,1}, or the real numbers (different from supervised learning) (different from supervised learning)
6
An example dialog for agent environment relationship: Environment: You are in state 65. You have 4 possible actions. Agent: I'll take action 2. Environment: You received a reinforcement of 7 units. You are now in state 15. You have 2 possible actions. Agent: I'll take action 1. Environment: You received a reinforcement of -4 units. You are now in state 65. You have 4 possible actions. Agent: I'll take action 2. Environment: You received a reinforcement of 5 units. You are now in state 44. You have 5 possible actions..
7
Some Examples Bioreactor Bioreactor actions: stirring rate, temperature control states: sensory readings related to chemicals reward: instant production rate of target chemical Recycling robot Recycling robot actions: search for a can, wait, or recharge states: low battery, high battery reward: + for having a can, - for running out of battery
8
ROADMAP: Introduction to the problem Introduction to the problem Models of Optimal Behavior Models of Optimal Behavior An Immediate Reward Example and Solution An Immediate Reward Example and Solution Markov Decision Process Markov Decision Process Model Free Methods Model Free Methods Model Based Methods Model Based Methods Partially Observable Markov Decision Process Partially Observable Markov Decision Process Applications Applications Conclusion Conclusion
9
Models of Optimal Behavior: Agent tries to maximize one of the following: finite-horizon Model: infinite-horizon discounted model: average-reward model:
10
ROADMAP: Introduction to the problem Introduction to the problem Models of Optimal Behavior Models of Optimal Behavior An Immediate Reward Example and Solution An Immediate Reward Example and Solution Markov Decision Process Markov Decision Process Model Free Methods Model Free Methods Model Based Methods Model Based Methods Partially Observable Markov Decision Process Partially Observable Markov Decision Process Applications Applications Conclusion Conclusion
11
k-armed bandit k gambling machines k gambling machines h pulls are allowed h pulls are allowed Machines are not equivalent: Machines are not equivalent: Trying to learn paying probabilities Trying to learn paying probabilities Tradeoff between exploitation-exploration
12
Solution Strategies 1: Dynamic Programming Approach A belief state : Expected pay-off remaining: Probability of action i being paid:
13
Dynamic Programming Approach:, then Because there are no remaining pulls. So, all values can be recursively computed: Update probabilities after each action
14
ROADMAP: Introduction to the problem Introduction to the problem Models of Optimal Behavior Models of Optimal Behavior An Immediate Reward Example and Solution An Immediate Reward Example and Solution Markov Decision Process Markov Decision Process Model Free Methods Model Free Methods Model Based Methods Model Based Methods Partially Observable Markov Decision Process Partially Observable Markov Decision Process Applications Applications Conclusion Conclusion
15
MARKOV DECISION PROCESS k-armed bandit gives immediate reward k-armed bandit gives immediate reward DELAYED REWARD? Characteristics of MDP: a set of states : S a set of actions : A a reward function : R : S x A R A state transition function: T: S x A ∏ ( S) T (s,a,s’): probability of transition from s to s’ using action a
16
MDP EXAMPLE: Transition function States and rewards Bellman Equation : (Greedy policy selection)
17
Value Iteration Algorithm AN ALTERNATIVE ITERATION: (Singh,1993) (Important for model free learning) Stop Iteration when V(s) differs less than є. Policy difference ratio =< 2єγ / (1-γ ) ( Williams & Baird 1993b)
18
Policy Iteration Algorithm Policies converge faster than values. Why faster convergence?
19
POLICY ITERATION ON GRID WORLD
20
0 0 0 0 0 0.8 -0.8 0 0
21
POLICY ITERATION ON GRID WORLD 0 0 0 0 0 0.8 -0.8 0 0 0 0 0 0.64 0 0.8 0.46 0 0
22
POLICY ITERATION ON GRID WORLD 0 0 0 0.64 0 0.8 0.46 0 0
23
POLICY ITERATION ON GRID WORLD 0 0 0 0.64 0 0.8 0.46 0 0 0.51 0 0 0.77 0 0.93 0.59 0.36 0
24
POLICY ITERATION ON GRID WORLD 0 0 0 0.64 0 0.8 0.46 0 0 0.51 0 0 0.77 0 0.93 0.59 0.36 0 0.66 0.41 0 0.89 0.32 0.95 0.70 0.480.19
25
MDP Graphical Representation β, α : T (s, action, s’ ) Similarity to HMMs
26
ROADMAP: Introduction to the problem Introduction to the problem Models of Optimal Behavior Models of Optimal Behavior An Immediate Reward Example and Solution An Immediate Reward Example and Solution Markov Decision Process Markov Decision Process Model Free Methods Model Free Methods Model Based Methods Model Based Methods Partially Observable Markov Decision Process Partially Observable Markov Decision Process Applications Applications Conclusion Conclusion
27
Model Free Methods Models of the environment: T: S x A ∏ ( S) and R : S x A R Do we know them? Do we have to know them? Monte Carlo Methods Monte Carlo Methods Adaptive Heuristic Critic Adaptive Heuristic Critic Q Learning Q Learning
28
Monte Carlo Methods Idea: Idea: Hold statistics about rewards for each state Hold statistics about rewards for each state Take the average Take the average This is the V(s) This is the V(s) Based only on experience Based only on experience Assumes episodic tasks Assumes episodic tasks (Experience is divided into episodes and all episodes will terminate regardless of the actions selected.) (Experience is divided into episodes and all episodes will terminate regardless of the actions selected.) Incremental in episode-by-episode sense not step-by- step sense. Incremental in episode-by-episode sense not step-by- step sense.
29
Problem: Unvisited pairs (problem of maintaining exploration) For every make sure that: P( selected as a start state and action) >0 (Assumption of exploring starts )
30
Monte Carlo Control How to select Policies: (Similar to policy evaluation)
31
ADAPTIVE HEURISTIC CRITIC & TD(λ) How the AHC learns, TD(0) algorithm: AHC : TD Algorithm
32
Q LEARNING Q values in Value Iteration: But we don’t know and Instead use the following : Decayed α properly? Q values will converge. (Singh 1994)
33
Q-LEARNING CRITICS: Simpler than AHC learning Simpler than AHC learning Q-Learning is exploration sensitive Q-Learning is exploration sensitive Analog to value iteration in MDP Analog to value iteration in MDP Most popular Model free learning algorithm Most popular Model free learning algorithm
34
ROADMAP: Introduction to the problem Introduction to the problem Models of Optimal Behavior Models of Optimal Behavior An Immediate Reward Example and Solution An Immediate Reward Example and Solution Markov Decision Process Markov Decision Process Model Free Methods Model Free Methods Model Based Methods Model Based Methods Partially Observable Markov Decision Process Partially Observable Markov Decision Process Applications Applications Conclusion Conclusion
35
Model Based Methods Model free methods do not learn the model parameters Model free methods do not learn the model parameters Inefficient use of data! Learn the model.
36
Certainty Equivalent Methods: First learn the models of the environment by keeping statistics, then learn the actions to take. First learn the models of the environment by keeping statistics, then learn the actions to take. Objections: Arbitrary division between the learning phase and Arbitrary division between the learning phase and acting phase acting phase Initial data gathering. (How to choose exploration strategy without knowing the model) Initial data gathering. (How to choose exploration strategy without knowing the model) Changes in the environment? Changes in the environment? Better to learn the model and to act simultaneously
37
DYNA After action a, from s to s’ with reward r: 1.Update Transition model T and reward function R 2.Update Q values: 3.Do k more random updates (random s, a pairs ):
38
ROADMAP: Introduction to the problem Introduction to the problem Models of Optimal Behavior Models of Optimal Behavior An Immediate Reward Example and Solution An Immediate Reward Example and Solution Markov Decision Process Markov Decision Process Model Free Methods Model Free Methods Model Based Methods Model Based Methods Partially Observable Markov Decision Process Partially Observable Markov Decision Process Applications Applications Conclusion Conclusion
39
POMDPs What if state information (from sensors) is noisy? Mostly the case! MDP techniques are suboptimal! Two halls are not the same.
40
POMDPs – A Solution Strategy SE: Belief State Estimator (Can be based on HMM) П: MDP Techniques
41
ROADMAP: Introduction to the problem Introduction to the problem Models of Optimal Behavior Models of Optimal Behavior An Immediate Reward Example and Solution An Immediate Reward Example and Solution Markov Decision Process Markov Decision Process Model Free Methods Model Free Methods Model Based Methods Model Based Methods Generalization Generalization Partially Observable Markov Decision Process Partially Observable Markov Decision Process Applications Applications Conclusion Conclusion
42
APPLICATIONS Juggling robot: dynamic programming Juggling robot: dynamic programming (Schaal & Atkeson 1994) (Schaal & Atkeson 1994) Box pushing robot: Q-learning Box pushing robot: Q-learning (Mahadevan& Connel 1991a) (Mahadevan& Connel 1991a) Disk collecting robot: Q-learning Disk collecting robot: Q-learning (Mataric 1994) (Mataric 1994)
43
ROADMAP: Introduction to the problem Introduction to the problem Models of Optimal Behavior Models of Optimal Behavior An Immediate Reward Example and Solution An Immediate Reward Example and Solution Markov Decision Process Markov Decision Process Model Free Methods Model Free Methods Model Based Methods Model Based Methods Generalization Generalization Partially Observable Markov Decision Process Partially Observable Markov Decision Process Applications Applications Conclusion Conclusion
44
Conclusion RL is not supervised learning RL is not supervised learning Planning rather than classification Planning rather than classification Poor performance on large problems. Poor performance on large problems. New methods neded New methods neded (i.e. shaping, imitation, reflexes) (i.e. shaping, imitation, reflexes) RL/MDPs extends HMMs. How? RL/MDPs extends HMMs. How?
45
MDP as a graph Is it possible to represent it as an HMM?
46
Relation to HMMs Recycling robot example revisited as an HMM problem Battery t-1 Battery t+1 Battery t Action t-1Action t+1Action t Battery={ low, high} Action={wait, search, recharge}
47
Relation to HMMs Recycling robot example revisited as an HMM problem Battery t-1 Battery t+1 Battery t Action t-1Action t+1Action t Battery={ low, high} Action={wait, search, recharge} Not representable as an HMM
48
HMMs vs MDPs Once we have the MDP representation of the problem, We can do inferences just like HMM by converting it to probabilistic automaton. Vice versa is not possible. Once we have the MDP representation of the problem, We can do inferences just like HMM by converting it to probabilistic automaton. Vice versa is not possible. (Actions, rewards) (Actions, rewards) Use HMMs to do probabilistic reasoning over time Use HMMs to do probabilistic reasoning over time Use MDP/RL to optimize the behavior. Use MDP/RL to optimize the behavior.
Similar presentations
© 2024 SlidePlayer.com Inc.
All rights reserved.