[home] [search] An Introduction to Universal Artificial Intelligence [contact] [up]  

by Marcus Hutter and David Quarel and Elliot Catt

Order Information

 Title: An Introduction to Universal Artificial Intelligence
Authors: Marcus Hutter and David Quarel and Elliot Catt
Publisher: Chapman & Hall, ISBN: 9781032607023, DOI: TBA
Date: © May 2024, Pages: 500

The book can be ordered from routledge.com, or amazon.com / co.uk / de / or most other bookshops.

Abstract

 This book provides a gentle introduction to Universal Artificial Intelligence (UAI), a theory that provides a formal underpinning of what it means for an agent to act intelligently in an unknown environment. First presented in book (Hutter, 2004), UAI offers a framework in which virtually all AI problems can be formulated, and a theory of how to solve them. UAI unifies ideas from sequential decision theory, Bayesian inference, and algorithmic information theory to construct AIXI, an optimal reinforcement learning agent that learns to act optimally in unknown environments. AIXI is the theoretical gold standard for intelligent behavior.
     The book covers both the theoretical and practical aspects of UAI. Bayesian updating can be done efficiently with context tree weighting, and planning can be approximated by sampling with Monte Carlo tree search. It provides algorithms for the reader to implement, and experimental results to compare against. These algorithms are used to approximate AIXI. The book ends with a philosophical discussion of Artificial General Intelligence: Can super-intelligent agents even be constructed? Is it inevitable that they will be constructed, and what are the potential consequences?
     This text is suitable for late undergraduate students. It provides an extensive chapter to fill in the required mathematics, probability, information, and computability theory background.
     The book is not a second edition of Hutter (2004) UAI. It is a prequel (much more gentle introduction to the prerequisite topics) and a sequel (progress in the last 20 years). It is easier (more detailed less dense explanations) and harder (covers some advanced topics such a CTW and GoT). It includes various practical approximations, pseudo-code, and links to implementations of some (in Java and C).

Keywords: Artificial general intelligence; algorithmic information theory; Bayes mixture distributions; universal sequence prediction; context tree weighting; rational agents; sequential decision theory; universal intelligent agents; reinforcement learning; games and multi-agent systems; approximation/implementation/application; AGI-safety; philosophy of AI.

Short Table of Contents

Critics' Reviews

 “Is it possible to mathematically define and study artificial superintelligence? If that sounds like an interesting question, then this is definitely the book for you. Starting with probability theory, complexity theory and sequence prediction, it takes you right through to the safety of superintelligent machines.”
– Shane Legg, co-founder of DeepMind

“This is seminal work!”
– Roman Yampolskiy, Tenured Associate Professor at the University of Louisville, USA

“This is an important, timely, high-quality book by highly respected authors.”
– Jürgen Schmidhuber, Director of the AI Initiative at King Abdullah University of Science and Technology, Scientific Director at the Swiss AI Lab IDSIA, Co-Founder & Chief Scientist at NNAISENSE

“Clearly very strongly based on mathematical foundations. This offers a theoretical depth which will be of value in research, education (at an appropriate level), and for advanced practitioners.”
– Alan Dix, Director of the Computational Foundry at Swansea University and Professorial Fellow at Cardiff Metropolitan University

Introduction

 TBA.

Feedback

 Feedback of any kind is welcome. Please send your feedback to Marcus Hutter.
Your suggestions will be considered for the next edition/printing or a potential future online version.

Slides

 I gave courses based on (parts of) the book to graduate students. I have prepared nearly 400 slides with exercises, suitable for at least 30 hours of lectures. I taught versions of this course in 2012 at ETH Zürich, 2010-2019 at ANU Canberra, in 2006 at HeCSE Helsinki, and in 2003 at TU Munich. You can find recordings of talks on parts of this material here.

Code

 Many papers the sections of this book are based on have code somewhere publicly available. Links to code are provided at the corresponding paper reference (search for 'code =').

BibTeX Entry

 
@Book{Hutter:24uaibook2,
  author =       "Marcus Hutter and David Quarel and Elliot Catt",
  title =        "An Introduction to Universal Artificial Intelligence",
  year =         "2024",
  url =          "http://www.hutter1.net/ai/uaibook2.htm",
  keywords =     "Artificial general intelligence; algorithmic information theory;
                  Bayes mixture distributions; universal sequence prediction;
                  context tree weighting; rational agents; sequential decision theory;
                  universal intelligent agents; reinforcement learning;
                  games and multi-agent systems; approximation/implementation/application;
                  AGI-safety; philosophy of AI.",
  support =      "ARC grant DP150104590",
  for =          "010404(20%),080101(20%),080198(20%),080299(10%),080401(30%)",
}
 © 2000 by ... [home] [search] [projects] [publications] [personal] [contact] [sitemap] ... Marcus Hutter