Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

Ai minecraft

68 views

Published on

The words ‘artificial intelligence’ are widely used these days. However, courses on AI usually are quite boring. What about a crash course using the most popular game of the year? Even if it might seem as a ‘game for children’, Minecraft AI bots are currently being used for research by big, serious companies, besides all level students.

Published in: Technology
  • Be the first to comment

  • Be the first to like this

Ai minecraft

  1. 1. NOEMI FERRERA Esta foto de Autor desconocido está bajo licencia CC BY Music by C418: https://minecraft.gamepedia.com/C418
  2. 2. Who am I? @TheTestLynx (Minecraft!) noemiferrera.com @TheTestLynx
  3. 3. Agenda 1. Definitions & purpose 2. How to program in Minecraft I. Minecraft Challenges II.Minecraft Makecode 3. How AI works in Minecraft I. Rules and goals of Minecraft II.Machine learning paradigms III.Project Malmo
  4. 4. What’s Minecraft? @TheTestLynxEsta foto de Autor desconocido está bajo licencia CC BY-SA-NC
  5. 5. What’s Minecraft?  Sandbox construction videogame  2011 Mojang – 2014 Microsoft  Best selling videogame of all time  Collect, assemble & destroy blocks in 3D @TheTestLynx
  6. 6. What’s a Minecraft Mod? Independent, user-made modifications to the game that extend or modify the original @TheTestLynx https://www.planetminecraft.com/resources/mods/
  7. 7. What’s AI? 1. a branch of computer science dealing with the simulation of intelligent behavior in computers 2. the capability of a machine to imitate intelligent human behavior @TheTestLynx https://www.merriam-webster.com/dictionary/artificial%20intelligence
  8. 8. Photo by Alex Knight from Pexels
  9. 9. Why do I care?  Trending in IT Software  Trending in testing  Automate the unthinkable  The future  Visual based testing  Automatic inference @TheTestLynx
  10. 10. Example @TheTestLynx This picture of unknown author is under license CC BY-NC Thia picture of unknown author is under license CC BY-NC
  11. 11. Why Minecraft?  Easy to understand  Fun  Goal based  Widely used – great dataset  Backed up: Microsoft, Carnegie Mellon University, Facebook @TheTestLynx
  12. 12. Minecraft challenges @TheTestLynx
  13. 13. Minecraft challenges @TheTestLynx Elements not retrievables Precondition for all tests Could not be the first screen
  14. 14. Minecraft challenges @TheTestLynx Multiple users for functions like online gaming and chat Written using different technologies
  15. 15. Automating Minecraft @TheTestLynx This picture of unknown author is under license CC BY-SA-NC
  16. 16. Preparation 1. Install Minecraft (30€ / 72 BYN) 2. Install Code Connection 3. Create a Minecraft world with cheat mode turned on 4. Start Code Connection @TheTestLynx Minecraft.makecode.com
  17. 17. Example : Agent digs @TheTestLynx
  18. 18. Example : Agent digs - JS @TheTestLynx
  19. 19. Example : Agent digs - Python @TheTestLynx
  20. 20. Example : Agent digs @TheTestLynx
  21. 21. Example : Agent digs @TheTestLynx
  22. 22. Example : Agent digs @TheTestLynx
  23. 23. AI with Minecraft @TheTestLynx This picture of unknown author is under license CC BY-NC-ND
  24. 24. Building manual https://minecraft.gamepedia.com/Tutorials/Beginner%27s_guide @TheTestLynx
  25. 25. Rules for survival @TheTestLynx 1. Fall damage can kill you 2. Lava can kill you –> never dig straight down 3. Monsters can kill you 4. Monsters appear at night or in caves 5. Being underwater can kill you 6. If you don´t sleep 4 nights in a row, flying monsters attack you 7. Hunger can kill you 8. You get hungry with time and movement 9. Some creatures can become hostile 10. Careful with food and potions …
  26. 26. Rules for starting @TheTestLynx 1. A hole in the ground could be a shelter 2. Collect wood 3. Make a pickaxe 4. Collect wool 5. Make a bed … goals?
  27. 27. Achievements guide - goals https://minecraft.gamepedia.com/Tutorials/Achievement_guide @TheTestLynx
  28. 28. Rules based system @TheTestLynx […]Set of "if-then" statements that uses a set of assertions, to which rules on how to act upon those assertions are created. In software development, rule-based systems can be used to create software that will provide an answer to a problem in place of a human expert.[…] https://www.webopedia.com/TERM/R/rule_based_system.html
  29. 29. Machine learning Arthur samuel @TheTestLynx Field of study that gives computers the ability to learn without being explicitly programmed. https://www.ibm.com/developerworks/community/blogs/jfp/entry/What_Is_Machine_Learning?lang=en Field of study that gives computers the ability to learn without being explicitly programmed.
  30. 30. Machine learning paradigms @TheTestLynx Supervised learning Unsupervised learning Reinforcement learning
  31. 31. Supervised learning @TheTestLynx These pictures are from an unknown author under license CC BY-SA-NC  Training set
  32. 32. Unsupervised learning @TheTestLynx These pictures are from an unknown author under license CC BY-SA-NC
  33. 33. What’s an intelligent agent? @TheTestLynx This picture of unknown author it’s under license CC BY-SA-NC An autonomous entity which acts, directing its activity towards achieving goals (i.e. it is an agent), upon an environment using observation through sensors and consequent actuators (i.e. it is intelligent) https://en.wikipedia.org/wiki/Intelligent_agent autonomous entity goals achieving using observation
  34. 34. Reinforcement learning @TheTestLynx https://en.wikipedia.org/wiki/Reinforcement_learning Area of machine learning concerned with how software agents ought to take actions in an environment in order to maximize some notion of cumulative reward. This picture of unknown author it’s under license CC BY-SA-NC Area of machine learning concerned with how software agents ought to take actions in an environment in order to maximize some notion of cumulative reward. Area of machine learning concerned with how software agents ought to take actions in an environment in order to maximize some notion of cumulative reward. Area of machine learning concerned with how software agents ought to take actions in an environment in order to maximize some notion of cumulative reward.
  35. 35. Project Malmo @TheTestLynx Java8 JDK, python3, git https://gym.openai.com/ NumPy, Pillow, lxml AI experimentation & research built on Minecraft https://github.com/Microsoft/malmo
  36. 36. Project Malmo @TheTestLynx Sending your code to the gym Observation Reward Done Info This picture of unknown author is under license CC BY-SA-NC
  37. 37. Project Malmo @TheTestLynx https://youtu.be/iOWbeMvml3M
  38. 38. Project Malmo - challenges @TheTestLynx 2017 - The Malmo Collaborative AI Challenge https://www.crowdai.org/ - Crowdsourcing AI to solve real-world problems MARLO 2018 - Multi-Agent Reinforcement Learning in Minecraft NeurIPS 2019 : MineRL Competition
  39. 39. Recap What’s AI What are Minecraft mods Tool to learn/teach code with Minecraft What makes a system good fit for AI What’s Machine learning ML paradigms Reinforcement learning & Intelligent agents Where to find AI projects to join @TheTestLynx
  40. 40. Thank you! This picture of unknown author is under license CC BY-SA-NC @TheTestLynx

×