Marcus Hutter (born 14 April 1967 in Munich) is a German computer scientist and professor specializing in the mathematical foundations of artificial general intelligence. He is best known for developing AIXI, a theoretical framework for an optimal intelligent agent, and for establishing the Hutter Prize for lossless compression of human knowledge. As of 2024, he works as a senior researcher at DeepMind.
Hutter studied physics and computer science at the Technical University of Munich. In 2000, he joined Jürgen Schmidhuber's group at the Dalle Molle Institute for Artificial Intelligence Research in Manno, Switzerland, where he began his foundational work on universal AI. He later served as a professor at the College of Engineering, Computing and Cybernetics at the Australian National University in Canberra, Australia.
AIXI and Universal Artificial Intelligence
Starting in 2000, Hutter developed and published a mathematical theory of artificial general intelligence named AIXI. The theory combines algorithmic probability with reinforcement learning to define an idealised agent that maximises expected reward over all possible futures. AIXI is not computable in practice but serves as a theoretical benchmark for intelligent behaviour. His first book, Universal Artificial Intelligence: Sequential Decisions Based on Algorithmic Probability, was published in 2005 by Springer.
In 2005, Hutter and his doctoral student Shane Legg published an intelligence test for artificial intelligence devices, proposing a formal measure of intelligence based on performance across a wide range of environments. In 2009, Hutter developed the theory of feature reinforcement learning, which extends AIXI to handle complex state spaces more efficiently. In 2014, Lattimore and Hutter published an asymptotically optimal extension of the AIXI agent, addressing some of its computational limitations.
Hutter Prize
In 2006, Hutter announced the Hutter Prize for Lossless Compression of Human Knowledge, offering a total of €50,000 in prize money. The prize incentivises participants to compress a fixed 100 MB English text file, with the goal of encouraging progress toward artificial general intelligence. In 2020, Hutter raised the prize money to €500,000, reflecting the ongoing importance of compression as a proxy for intelligence.
Public Engagement and Recent Work
Hutter has engaged with broader audiences through interviews and publications. In 2021, he appeared on the Lex Fridman podcast to discuss his theory of Universal AI. A more technical follow-up with Tim Nguyen was released in 2024 on the Cartesian Cafe. His new book, published in 2024, provides a more accessible introduction to Universal AI and reviews progress over the two decades since his first book, including a chapter on artificial superintelligence (ASI) safety. This work was featured as a keynote at the inaugural workshop on AI safety in Sydney.
Legacy and Impact
Hutter's work has influenced theoretical research in machine learning and reinforcement learning, particularly in the area of universal induction and decision-making. His formalisation of AIXI has inspired subsequent research on optimal agents and has been cited extensively in the field of artificial general intelligence. The Hutter Prize continues to attract participants and has contributed to advances in data compression algorithms.
See Also
- Solomonoff induction