Mike Taks Bram van de Klundert. About Published 2005 Cited 286 times Kenneth O. Stanley Associate Professor at University of Central Florida Risto Miikkulainen.

Slides:



Advertisements
Similar presentations
© Negnevitsky, Pearson Education, Lecture 12 Hybrid intelligent systems: Evolutionary neural networks and fuzzy evolutionary systems Introduction.
Advertisements

Yuri R. Tsoy, Vladimir G. Spitsyn, Department of Computer Engineering
Tetris – Genetic Algorithm Presented by, Jeethan & Jun.
Constructing Complex NPC Behavior via Multi- Objective Neuroevolution Jacob Schrum – Risto Miikkulainen –
Tetris and Genetic Algorithms Math Club 5/30/2011.
On the Genetic Evolution of a Perfect Tic-Tac-Toe Strategy
Biologically Inspired AI (mostly GAs). Some Examples of Biologically Inspired Computation Neural networks Evolutionary computation (e.g., genetic algorithms)
Evolving New Strategies The Evolution of Strategies in the Iterated Prisoner’s Dilemma 01 / 25.
Genetic Algorithms, Part 2. Evolving (and co-evolving) one-dimensional cellular automata to perform a computation.
Neuro-Evolution of Augmenting Topologies Ben Trewhella.
Evolving Neural Network Agents in the NERO Video Game Author : Kenneth O. Stanley, Bobby D. Bryant, and Risto Miikkulainen Presented by Yi Cheng Lin.
Evolutionary Algorithms Simon M. Lucas. The basic idea Initialise a random population of individuals repeat { evaluate select vary (e.g. mutate or crossover)
Design of Autonomous Navigation Controllers for Unmanned Aerial Vehicles using Multi-objective Genetic Programming Gregory J. Barlow March 19, 2004.
1. 2 overview Background on video games Background on video games Neural networks Neural networks NE NE NEAT NEAT rtNEAT rtNEAT NERO NERO.
Evolutionary Computation Application Peter Andras peter.andras/lectures.
Nicholas Mifsud.  Behaviour Trees (BT) proposed as an improvement over Finite State Machines (FSM)  BTs are simple to design, implement, easily scalable,
Marcus Gallagher and Mark Ledwich School of Information Technology and Electrical Engineering University of Queensland, Australia Sumaira Saeed Evolving.
Computational Intelligence in Games: An Overview Zahid Halim Faculty of Computer Science and Engineering Ghulam Ishaq Khan Institute of Engineering Sciences.
Evolving Multi-modal Behavior in NPCs Jacob Schrum – Risto Miikkulainen –
Genetic Algorithm.
Evolutionary Robotics NEAT / HyperNEAT Stanley, K.O., Miikkulainen (2001) Evolving Neural Networks through Augmenting Topologies. Competing Conventions:
林偉楷 Taiwan Evolutionary Intelligence Laboratory.
More precisely called Branch of AI behind it.
Slides are based on Negnevitsky, Pearson Education, Lecture 12 Hybrid intelligent systems: Evolutionary neural networks and fuzzy evolutionary systems.
Soft Computing Lecture 18 Foundations of genetic algorithms (GA). Using of GA.
Evolutionary Computation. Evolutionary Complexification Two major goals in intelligent systems are the discovery and improvement of solutions to complex.
CAP6938 Neuroevolution and Developmental Encoding Working with NEAT Dr. Kenneth Stanley September 27, 2006.
More on coevolution and learning Jing Xiao April, 2008.
Boltzmann Machine (BM) (§6.4) Hopfield model + hidden nodes + simulated annealing BM Architecture –a set of visible nodes: nodes can be accessed from outside.
Optimal resource assignment to maximize multistate network reliability for a computer network Yi-Kuei Lin, Cheng-Ta Yeh Advisor : Professor Frank Y. S.
Computer Go : A Go player Rohit Gurjar CS365 Project Presentation, IIT Kanpur Guided By – Prof. Amitabha Mukerjee.
CAP6938 Neuroevolution and Developmental Encoding Real-time NEAT Dr. Kenneth Stanley October 18, 2006.
Artificial Life/Agents Creatures: Artificial Life Autonomous Software Agents for Home Entertainment Stephen Grand, 1997 Learning Human-like Opponent Behaviour.
2005MEE Software Engineering Lecture 11 – Optimisation Techniques.
ELeaRNT: Evolutionary Learning of Rich Neural Network Topologies Authors: Slobodan Miletic 3078/2010 Nikola Jovanovic 3077/2010
科技英文作業 黃方世 陳竹軒. Introduction Talking about video games. Training agents for what. Focus on training Non-player-character (NPC).
EGEE-III INFSO-RI Enabling Grids for E-sciencE EGEE and gLite are registered trademarks Grid-enabled parameter initialization for.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology 1 Evolving Reactive NPCs for the Real-Time Simulation Game.
Genetic Algorithms CSCI-2300 Introduction to Algorithms
Pac-Man AI using GA. Why Machine Learning in Video Games? Better player experience Agents can adapt to player Increased variety of agent behaviors Ever-changing.
Artificial Intelligence Research in Video Games By Jacob Schrum
CAP6938 Neuroevolution and Artificial Embryogeny Competitive Coevolution Dr. Kenneth Stanley February 20, 2006.
Evolving Multimodal Networks for Multitask Games
CAP6938 Neuroevolution and Developmental Encoding Evolving Adaptive Neural Networks Dr. Kenneth Stanley October 23, 2006.
CAP6938 Neuroevolution and Developmental Encoding Intro to Neuroevolution Dr. Kenneth Stanley September 18, 2006.
Riza Erdem Jappie Klooster Dirk Meulenbelt EVOLVING MULTI-MODAL BEHAVIOR IN NPC S.
N- Queens Solution with Genetic Algorithm By Mohammad A. Ismael.
CAP6938 Neuroevolution and Artificial Embryogeny Approaches to Neuroevolution Dr. Kenneth Stanley February 1, 2006.
CAP6938 Neuroevolution and Artificial Embryogeny Evolutionary Computation Theory Dr. Kenneth Stanley January 25, 2006.
Artificial Intelligence By Mr. Ejaz CIIT Sahiwal Evolutionary Computation.
Overview Last two weeks we looked at evolutionary algorithms.
CAP6938 Neuroevolution and Artificial Embryogeny Evolutionary Comptation Dr. Kenneth Stanley January 23, 2006.
CAP6938 Neuroevolution and Artificial Embryogeny Real-time NEAT Dr. Kenneth Stanley February 22, 2006.
Evolutionary Computing Systems Lab (ECSL), University of Nevada, Reno 1 Authors : Siming Liu, Christopher Ballinger, Sushil Louis
An Evolutionary Algorithm for Neural Network Learning using Direct Encoding Paul Batchis Department of Computer Science Rutgers University.
March 1, 2016Introduction to Artificial Intelligence Lecture 11: Machine Evolution 1 Let’s look at… Machine Evolution.
 Negnevitsky, Pearson Education, Lecture 12 Hybrid intelligent systems: Evolutionary neural networks and fuzzy evolutionary systems n Introduction.
Evolutionary Computation Evolving Neural Network Topologies.
AI in Space Group 3 Raquel Cohen Yesenia De La Cruz John Gratton.
CAP6938 Neuroevolution and Artificial Embryogeny Real-time NEAT Dr. Kenneth Stanley February 22, 2006.
Neuro-evolving Maintain-Station Behavior for Realistically Simulated Boats Nathan A. Penrod David Carr Sushil J. Louis Bobby D. Bryant Evolutionary Computing.
TORCS WORKS Jang Su-Hyung.
Dr. Kenneth Stanley January 30, 2006
HyperNetworks Engın denız usta
Dr. Kenneth Stanley September 25, 2006
Evolving Multimodal Networks for Multitask Games
UT^2: Human-like Behavior via Neuroevolution of Combat Behavior and Replay of Human Traces Jacob Schrum, Igor Karpov, and Risto Miikkulainen
Dr. Kenneth Stanley September 20, 2006
Dr. Kenneth Stanley February 6, 2006
Boltzmann Machine (BM) (§6.4)
Presentation transcript:

Mike Taks Bram van de Klundert

About Published 2005 Cited 286 times Kenneth O. Stanley Associate Professor at University of Central Florida Risto Miikkulainen Professor at the University of Texas at Austin Bobby D. Bryant Assistant Professor at university of Nevada

Contents Introduction NEAT rtNEAT NERO

Introduction real-time NeuroEvolution of Augmenting Topologies (rtNEAT) Adaption of the NEAT algorithm Create new genre of games requiring learning Black and white Tamagotchi

NEAT Neuro Evolution of Augmenting Topologies Growing neural network

Representation List of connection genes Innovation number Global counter In node Out node Weight Enabled

Initial population Uniform population of simple networks No hidden nodes Random weights

Mutation Weight mutation Structure mutation Add a connection between two nodes Replace a connection by a node Connection not removed only disabled Out connection inherits the value

Crossover terms Disjoint: gene is only in one network Excess: disjoint and outside of the range of innovations

Crossover shared genes: Uniform crossover Blend crossover Disjoint and excess genes Taken from most fit parent

Crossover example Equal fitness 9, 10 excess 6, 7, 8 disjoint

Speciation

Species assignment Check if there is a species close enough to the individual If not, create new species

Fitness

Selection

Trailer NERO

rtNEAT Differences Selection and replacement Removing agent

Differences Work real time Originally NEAT evaluates one complete generation of individuals, generates offspring “en masse”

Differences During a game, performance statistics are being recorded Replacing agents Perform actions every n game-ticks

Selection and replacement Calculate fitness Remove worst agent of sufficient age Choose parents among the best Create offspring, Reassign all agents to species

Removing agent Remove worst agent based fitness adjusted for species size New agents are continuously born, life time individually kept track of Possibility to just replace the neural network of an agent

NERO Player is a trainer Set up exercises Save and load neural networks

Training mode Place objects on the field (static enemies, turrets, rovers, flags,...) Adjust fitness rewards by sliders

Training mode Agents spawn in the “factory”

Sensors Radar to track enemy location Rangefinder Line-of-fire...

Evolved topology

Battle mode Assemble team of 20 agents Ends if one team is empty

Experiments Slightly nondeterministic game engine The same game is thus never played twice

Behaviors Different behaviors are trained Seek and fire by placing a single static enemy on the training field Firing and hitting a target was to slow to evolve. Aiming script was used

Behaviors Avoidance trained by controlling an agent manually Agent runs backwards facing the enemy and shooting at it

Behaviors Train agents to avoid turret fire

More complex behaviors Let agents attack enemy behind a wall Train agents to avoid hazardous corridors

More complex behaviors Train agents versus targets that are standing against a wall

More complex behaviors Incrementally add walls, agents will be able to navigate

Battling Paper rock scissors Seek vs avoidance

Battling

Questions?