PHE YEONG KIANG A152076. Introduction For this course LMCK1531 KEPIMPINAN & KREATIVITI, I will talk about what I've try to do in the Robot Soccer's Club.

Slides:

Advertisements

Similar presentations

Know your film terms…. Take a look at the following screens, and try to names the shot or camera angle you can see by writing them down. You may work.

Advertisements

Effective Reinforcement Learning for Mobile Robots Smart, D.L and Kaelbing, L.P.

1 Monte Carlo Methods Week #5. 2 Introduction Monte Carlo (MC) Methods –do not assume complete knowledge of environment (unlike DP methods which assume.

1 Section 14.1 Computability Some problems cannot be solved by any machine/algorithm. To prove such statements we need to effectively describe all possible.

Lehrstuhl für Informatik 2 Gabriella Kókai: Maschine Learning Reinforcement Learning.

Università di Milano-Bicocca Laurea Magistrale in Informatica Corso di APPRENDIMENTO E APPROSSIMAZIONE Lezione 6 - Reinforcement Learning Prof. Giancarlo.

RL at Last! Q- learning and buddies. Administrivia R3 due today Class discussion Project proposals back (mostly) Only if you gave me paper; e-copies yet.

Model-Free vs. Model- Based RL: Q, SARSA, & E 3. Administrivia Reminder: Office hours tomorrow truncated 9:00-10:15 AM Can schedule other times if necessary.

Reinforcement Learning

Q. The policy iteration alg. Function: policy_iteration Input: MDP M = 〈 S, A,T,R 〉  discount  Output: optimal policy π* ; opt. value func. V* Initialization:

Reinforcement Learning

Robot Learning Jeremy Wyatt School of Computer Science University of Birmingham.

Policies and exploration and eligibility, oh my!.

Cooperative Q-Learning Lars Blackmore and Steve Block Expertness Based Cooperative Q-learning Ahmadabadi, M.N.; Asadpour, M IEEE Transactions on Systems,

1 Hybrid Agent-Based Modeling: Architectures,Analyses and Applications (Stage One) Li, Hailin.

Machine Learning Lecture 11: Reinforcement Learning

1 Kunstmatige Intelligentie / RuG KI Reinforcement Learning Johan Everts.

1 MACHINE LEARNING TECHNIQUES IN IMAGE PROCESSING By Kaan Tariman M.S. in Computer Science CSCI 8810 Course Project.

Reinforcement Learning Presented by: Kyle Feuz.

Learning: Reinforcement Learning Russell and Norvig: ch 21 CMSC421 – Fall 2005.

Reinforcement Learning (1)

Q. Administrivia Final project proposals back today (w/ comments) Evaluated on 4 axes: W&C == Writing & Clarity M&P == Motivation & Problem statement.

Policies and exploration and eligibility, oh my!.

Kunstmatige Intelligentie / RuG KI Reinforcement Learning Sander van Dijk.

By Jessika. School is back! It is so boring! Now I need so much help that my parent’s don’t understand! Wait! I can use Open Study last year it past me.

CS Reinforcement Learning1 Reinforcement Learning Variation on Supervised Learning Exact target outputs are not given Some variation of reward is.

Topics on Final Perceptrons SVMs Precision/Recall/ROC Decision Trees Naive Bayes Bayesian networks Adaboost Genetic algorithms Q learning Not on the final:

Reinforcement Learning

Introduction Many decision making problems in real life

Advances in Computing Chapter 13 Multimedia, Artificial Intelligence, and Intelligent Agents © The McGraw-Hill Companies, Inc., 2000.

Encoding Robotic Sensor States for Q-Learning using the Self-Organizing Map Gabriel J. Ferrer Department of Computer Science Hendrix College.

Reinforcement Learning (II.) Exercise Solutions Ata Kaban School of Computer Science University of Birmingham.

Ensemble of ensemble of tree and neural network Louis Duclos-Gosselin.

Machine Learning.

Balancing Exploration and Exploitation Ratio in Reinforcement Learning Ozkan Ozcan (1stLT/ TuAF)

Reinforcement Learning

Q-learning Watkins, C. J. C. H., and Dayan, P., Q learning,

Introduction to Reinforcement Learning Dr Kathryn Merrick 2008 Spring School on Optimisation, Learning and Complexity Friday 7 th.

Overview of Machine Learning RPI Robotics Lab Spring 2011 Kane Hadley.

Reinforcement Learning Ata Kaban School of Computer Science University of Birmingham.

Curiosity-Driven Exploration with Planning Trajectories Tyler Streeter PhD Student, Human Computer Interaction Iowa State University

Cooperative Q-Learning Lars Blackmore and Steve Block Multi-Agent Reinforcement Learning: Independent vs. Cooperative Agents Tan, M Proceedings of the.

PHE YEONG KIANG A Introduction For this courses LMCK2711 TANGGUNGJAWAB ALAM SEKITAR, I will talk about how Robot Soccer's Club make environmental.

1 Introduction to Reinforcement Learning Freek Stulp.

Accurate Robot Positioning using Corrective Learning Ram Subramanian ECE 539 Course Project Fall 2003.

Course Overview  What is AI?  What are the Major Challenges?  What are the Main Techniques?  Where are we failing, and why?  Step back and look at.

Software Multiagent Systems: CS543 Milind Tambe University of Southern California

Reinforcement Learning AI – Week 22 Sub-symbolic AI Two: An Introduction to Reinforcement Learning Lee McCluskey, room 3/10

PHE YEONG KIANG A Introduction For this course LMCK1621 Etika & Profesional, I will talk about what task already done in the Robot Soccer's Club.

CAT 3 – Analysing a skill Golf swing Joe Smith. Elite Performer’s Swing My Swing Prior to the commencement of the skill Follow through Ball release Initial.

Learning for Physically Diverse Robot Teams Robot Teams - Chapter 7 CS8803 Autonomous Multi-Robot Systems 10/3/02.

Possible actions: up, down, right, left Rewards: – 0.04 if non-terminal state Environment is observable (i.e., agent knows where it is) MDP = “Markov Decision.

Level Headed Drew Mcdermott Yale University, USA Artificial Intelligence 171(2007) 1183~1186 Sangyoon Yi Bi. Lab.

Reinforcement Learning Guest Lecturer: Chengxiang Zhai Machine Learning December 6, 2001.

Reinforcement Learning. Overview Supervised Learning: Immediate feedback (labels provided for every input). Unsupervised Learning: No feedback (no labels.

REINFORCEMENT LEARNING Unsupervised learning 1. 2 So far ….  Supervised machine learning: given a set of annotated istances and a set of categories,

Dinner for Two. Fuzzify Inputs Apply Fuzzy Operator.

Università di Milano-Bicocca Laurea Magistrale in Informatica Corso di APPRENDIMENTO AUTOMATICO Lezione 12 - Reinforcement Learning Prof. Giancarlo Mauri.

CS 5751 Machine Learning Chapter 13 Reinforcement Learning1 Reinforcement Learning Control learning Control polices that choose optimal actions Q learning.

Done Done Course Overview What is AI? What are the Major Challenges?

Reinforcement Learning

Accurate Robot Positioning using Corrective Learning

ALGORITHMS & FLOWCHARTING II

MACHINE LEARNING TECHNIQUES IN IMAGE PROCESSING

A paired-samples t-test compares the means of two related sets of data to see if they differ statistically. IQ Example We may want to compare the IQ scores.

MACHINE LEARNING TECHNIQUES IN IMAGE PROCESSING

Introduction to Reinforcement Learning and Q-Learning

My Animal Report Hook your reader: Tell an interesting fact.

THE ASSISTIVE SYSTEM SHIFALI KUMAR BISHWO GURUNG JAMES CHOU

Reinforcement Learning (2)

Presentation transcript:

PHE YEONG KIANG A152076

Introduction For this course LMCK1531 KEPIMPINAN & KREATIVITI, I will talk about what I've try to do in the Robot Soccer's Club.

What I doing in the Robot Soccer's Club? 1. Calculate the angle of robot to shoot the ball. 2. Study some machine learning knowledge.

Why I doing the task? 1. Because I want improve robot accuracy shooting the ball and reduce the process when robots shoot the ball. 2. Because I want improve robot learning technique.

How I doing the task? A.Improve robot accuracy shooting the ball. State1: When camera get State2: The computer will the position ball and robot. calculate the angle for the ball to find the location for the robot.

State3: The robot will change State4: The computer will the position to 0 degree.calculate the angle for robot to change the position to shoot the ball.

State5: Then the ball will be shot into the goal.

How I doing the task? B. Reduce the process when robots shoot the ball. State1: When camera get State2: The computer will the position ball and robot. calculate the angle for the ball to find the location for the robot.

State3: The computer will State4: Then the ball will be shot calculate the angle for robot into the goal. to direct change the position and shoot the ball, without change the position to 0 degree. I already successfully completed the part A, but part B still exploring.

Improve robot learning technique I using the Q- Learning to try improve robot learning technique. In fact, I have no any knowledge for Q-Learning, then I do a lot of investigation and to interrogate my seniors, finally I have some knowledge about the theory of Q- Learning. But in the meantime I was very upset, I asked the senior, senior said he did not know, because he never learned, so, I can only themselves to explore. Now I show what I learned in this Q- Learning.

This is the sample how the object get the state from start point to end point.

Algorithm State1: For each state-action pair (s, a), initialize the table entry Q i (s, a) to zero State2: Observe the current state s State3: Do forever: State3.1: Select an action a and execute it State3.2: Receive immediate reward r State3.3: Observe the new state s[new] State3.4: Update the table entry for Q[new](s, a)as follows: Q[old](s,a)=r + γ *max a * Q[new](s,a) s[old]=s[new]

Q(s2, a23) = r * max(Q(s3,a32),Q(s3,a36)) = *0=0, *0=0

S6 is FINAL STATE so Q(s3, a36) = r = 100

After the end point, the program will loop again until all the state get the value. Q(s2, a23) = r * max(Q(s3,a32), Q(s3,a36)) = * 0 = 0, * 100 = 50

The is the output after all state get the value. And the next part is using neural network to find the good way from start point to end point. But this part I also exploring.

I believe these tasks will improve the robot system. And I feel happy because can try to do this task, and I will continue to explore these tasks. Conclusion