Reinforcement Learning, Application and Q-Learning

In this slide we have coverd the following topices
What is Artificial Intelligence?
What is Machine Learning?
Relationship among AI, ML and DL.
Human Brain Learning Process
Learning Vs Recognition
Supervised Learning
Unsupervised Learning
Reinforcement Learning
Definition of Reinforcement Learning
Reinforcement Learning Application: AWS Deep racer
Markov Decision Process
Understanding Q-Learning Algorithm
Q-Learning Algorithm Example

1. Bangladesh University of Professionals (BUP) MCSE-1101: Advanced Artificial Intelligence Machine Learning:- Reinforcement learning Presentation Title Presented By Md. Abdullah al Mamun A.B.M. Nazibullah

2. What is Artificial Intelligence? What is Machine Learning? Relationship among AI, ML and DL. Human Brain Learning Process Learning Vs Recognition Supervised Learning Unsupervised Learning Reinforcement Learning Definition of Reinforcement Learning Reinforcement Learning Application: AWS Deep racer Markov Decision Process Understanding Q-Learning Algorithm Q-Learning Algorithm Example Outline

3. What is exactly Artificial Intelligence? Artificial Intelligence is a model/procedure/tool who has capability for self learning , dynamically detect the pattern/object and take decision by own knowledge just like human brain . “ So according to the definition, is it proved that AI i s really threat for human existence?”

4. Machine Learning? Machine Learning is a subset of Artificial Intelligence(AI) which provides machines the ability to learn automatically & improve from experience without being explicitly programmed.

5. The Relationship among AI, ML and DL Machine Learning is a sub-category of AI Deep Learning is a sub-category of ML That’s mean they are both forms of AI

6. Human Brain Learning Process Input Image Feature Extraction Learning Human Brain Neuron Recognition

7. Learning Vs Recognition Learning Learning is a search through the space of possible hypotheses for one that will perform well, even on new examples beyond the training set. To measure the accuracy of a hypothesis we give it a test set of examples that are distinct from the training set. Recognition According to the training dataset learning process is performed and engine is updated. By pass through the input sample over the engine and it will return an output according to the learning accuracy.

8. Supervised Learning Supervised Learning use of labeled datasets to train algorithms that to classify data or predict outcomes accurately. As input data is fed into the model, it adjusts its weights through a reinforcement learning process, which ensures that the model has been fitted appropriately. The model first learns from the given training data. The training data contains different patterns, which the model will learn. Application: classifying spam in a separate folder from your inbox Image- and object-recognition Predictive analytics

9. Unsupervised Learning Unsupervised learning has no training phase; instead, the algorithm is simply handed a dataset and uses the variables within the data to identify and separate out natural clusters. Application: Finding customer segments Feature selection

10. Reinforcement Learning Reinforcement Learning(RL) is a type of machine learning technique that enables an agent to learn in an interactive environment by trial and error using feedback from its own actions and experiences. Application: Robot deciding its path Next move in a chess game

11. Definition of Reinforcement Learning

12. A Taxonomy of RL Algorithms Model-based RL uses experience to construct an internal model of the transitions and immediate outcomes in the environment. Model-free RL , on the other hand, uses experience to learn directly one or both of two simpler quantities (state/ action values or policies) which can achieve the same optimal behavior but without estimation or use of a world model.

13. AWS Deepracer AWS DeepRacer gives you an interesting and fun way to get started with reinforcement learning (RL). RL is an advanced machine learning (ML) technique that takes a very different approach to training models than other machine learning methods. Its super power is that it learns very complex behaviors without requiring any labeled training data, and can make short term decisions while optimizing for a longer term goal. https://aws.amazon.com/deepracer/

14. AWS DeepRacer - Training https://www.youtube.com/watch?v=-PeGCyBTzVc

15. The following parameters are used to attain a solution: Set of actions (A) Set of states (S) Reward (R) Policy ( π ) Value (V) Markov Decision Process The mathematical approach for mapping a solution in reinforcement learning is called Markov Decision Process(MDP)

16. Understanding Q-Learning Place an agent in any one of the rooms(0,1,2,3,4) and the goal is to reach outside the building(room 5) 5 rooms in a building connected by doors. Each room is numbered 0 through 4 The outside of the building can be thought of as one big room(5) Door 1 & 4 lead into the building from room 5(outside)

17. Understanding Q-Learning(Graph Representation) Let's represent the rooms on a graph, each room as a node, and each door as a link

18. Q-Learning Example: Selected Path 1 -> 5 If we iterate the loop to select path from 1 to 5 then The matrix Q get's updated-

19. Q-Learning Example: Selected Path 2 -> 3 -> 4 -> 5 If we iterate the loop to select path from 2 to 5 then The matrix Q get's updated-

20. Thank You Thank you very much for the opportunity to take part in this knowledge sharing session!

21. A Q & A “The important thing is not to stop questioning.” - Albert Einstein & Q