igomath
“These ideas, that are driving AlphaGo, are gonna drive our future. This is it folks!” - AlphaGo movie trailer
Igo Math - Natural and Artificial Intelligence and the Game of Go
This course is offered as MAT230 at Akita International University, launched in 2019. It is a course designed for
- explaining Artificial Intelligence in depth by focusing on a single application, the game of Go;
- improving metacognition and decision making skills by drawing parallels between artificial and natural intelligence;
- enhancing emotional intelligence and mental resilience by playing the game.
For semesters when the course is not offered, Igo Hiroba is an opportunity to learn and play the game.
FOR STUDENTS:
- syllabus,
- lecture notes preliminary and partial notes, constantly updated,
- the course’s youTube channel
- game related sample test questions
- example essays for writing assignments Igo and Me, AIs and Me, My learning method (early 2019),My learning method (late 2019),
- Tsumego (Go puzzle) collections, when in doubt, start with the elementary collection
- Game record sheet for 9x9,
- Michael Redmond’s beginner 9x9 video series
- 81 Little Lions: An Introduction to the 9x9 Board for Advanced Beginners
FOR EDUCATORS AND RESEARCHERS:
- PAPER with Antti Törmänen The Game Is Not over Yet—Go in the Post-AlphaGo Era
- course proposal,
- draft paper,
- research poster.
- 2018 Australian Go Congress, Academic Symposium, slides and re-recorded talk.
FOR EVERYONE:
- The rules of Go, short summary of the Japanese rule set, (a foldable flyer version)
- The Surrounding Game movie
- AlphaGo movie
- Mindscape Podcast Episode 52: Frank Lantz on the Logic and Emotion of Games
- Handcrafted AI self-plays
Recommended Software
Web
- online-go.com Online Go Server (OGS), the most international server for playing with humans and bots. For online delivery, this is the main tool.
- goproblems.com Tsumego practice.
Desktop, laptop
- Sabaki - Go board and SGF editor This game record editor will be used in the class and it is recommended for working on game review and tsumego assignments. By itself, it cannot play the game, but it can connect to many engines, some of the listed below. (Mac/Linux/Windows)
- GnuGo A classic engine often used for benchmarking. Windows version
- Pachi Another now classic engine using Monte-Carlo tree search.
- Leela Zero superhuman engine modelled after AlphaGo Zero. It is a good tool for AI analysis, either using Sabaki, or Lizzie.
Tablets, smartphones
There are several great commercial apps on both platforms. Here we list only the open-source and/or free available ones. Beware of ‘lite’ versions of paid apps with ads. They are not suitable for serious play and the ads do work subliminally.
Android
- GobanDroid Board editor, good for recording a game.
- GoDroid Board editor and AI engine (GnuGo v3.8).
- ElyGo Lite A ‘lite’ version of a commercial app, but it is ad-free and it includes GnuGo.
- Pachi for Android An AI stronger than GnuGo.
iOS
- Little Go Board editor, and the Fuego AI engine.
Go Teaching Videos (external sources)
Beginner Videos by Stephanie Yin from the New York Institute of Go
Sunday Go Lessons
- 9x9 tutoring game Part 1
- 9x9 tutoring game Part 2
- Life and death problems
- 10 minute Go series for total beginners starting on the 9x9 board.
Paul Lockhart
Michael Redmond 9p
- AlphaGo versus AlphaGo Game reviews of AlphaGo self-plays.
Crazy Nines
Math, Computing & AI videos (external sources)
- Elo rating by James Grime
- Minimax by Computerphile Intuitive explanation of adversarial tree search, illustrated with minimalistic and real game examples.
- Neural Networks by 3Blue1Brown Elementary introduction, however the fourth episode requires familiarity with the chain rule in (multivariate) calculus
- AlphaZero - Mastering Games Without Human Knowledge by DeepMind developer David Silver
- Life Scientific BBC podcast with Demis Hassabis