Marcus Hutter

Starting in 2000, Hutter developed and published a mathematical theory of artificial general intelligence, AIXI, based on idealised intelligent agents and reward-motivated reinforcement learning.{{cite book |author=Marcus Hutter |url=https://archive.org/details/arxiv-cs0004001 |title=A Theory of Universal Artificial Intelligence based on Algorithmic Complexity |year=2000 |bibcode=2000cs........4001H |arxiv=cs.AI/0004001}}{{Cite web |last=Hutter |first=Marcus |date=2013-11-28 |title=To create a super-intelligent machine, start with an equation |url=https://theconversation.com/to-create-a-super-intelligent-machine-start-with-an-equation-20756 |access-date=2024-08-25 |website=The Conversation |language=en-US}} His first book Universal Artificial Intelligence: Sequential Decisions Based on Algorithmic Probability was published in 2005 by Springer.Marcus Hutter (2005). Universal Artificial Intelligence: Sequential Decisions Based on Algorithmic Probability. Berlin; Heidelberg; New York: Springer. {{ISBN|9783540221395}}. Also in 2005, Hutter published with his doctoral student Shane Legg an intelligence test for artificial intelligence devices.{{Cite web |last=Graham-Rowe |first=Duncan |date=12 August 2005 |title=IQ test for AI devices gets experts thinking |url=https://www.newscientist.com/article/dn7842-iq-test-for-ai-devices-gets-experts-thinking/ |access-date=2024-08-25 |website=New Scientist |language=en-US}} In 2009, Hutter developed and published the theory of feature reinforcement learning.{{cite journal |author=Marcus Hutter |year=2009 |title=Feature Reinforcement Learning: Part I. Unstructured MDPs |url=http://www.hutter1.net/ai/phimdpx.pdf |journal=Journal of Artificial General Intelligence |volume=1 |issue=1 |page=3 |doi=10.2478/v10229-011-0002-8 |arxiv=0906.1713 |bibcode=2009JAGI....1....3H |issn=1946-0163}} In 2014, Lattimore and Hutter published an asymptotically optimal extension of the AIXI agent.{{cite conference |author=Tor Lattimore and Marcus Hutter |title=Algorithmic Learning Theory |chapter=Bayesian Reinforcement Learning with Exploration |series=Lecture Notes in Computer Science |conference=Proc. 25th International Conf. on Algorithmic Learning Theory ({ALT'14}) |year=2014|volume=8776 |pages=170–184 |chapter-url=http://www.hutter1.net/publ/pacbayes.pdf | doi=10.1007/978-3-319-11662-4_13 |hdl=1885/14709 |isbn=978-3-319-11661-7 |hdl-access=free }}

An accessible podcast with Lex Fridman about his theory of Universal AI appeared in 2021{{Cite web |last=Fridman |first=Lex |date=26 February 2020 |title=Marcus Hutter: Universal Artificial Intelligence, AIXI, and AGI |url=https://www.youtube.com/watch?v=E1AxVXt2Gv4 |via=YouTube |language=en-US}} and a more technical follow-up with Tim Nguyen in 2024 in the Cartesian Cafe.{{Cite web |last=Nguyen |first=Tim |date=10 May 2024 |title=Universal Artificial Intelligence and Solomonoff Induction |url=https://www.youtube.com/watch?v=7TgOwMW_rnk |via=YouTube |language=en-US}} His new (2024) book{{cite book |author=Marcus Hutter |title=An Introduction to Universal Artificial Intelligence |url=https://www.routledge.com/An-Introduction-to-Universal-Artificial-Intelligence/Hutter-Catt-Quarel/p/book/9781032607023 |date=2024 |publisher=Taylor & Francis |isbn=9781032607023}} also gives a more accessible introduction to Universal AI and progress in the 20 years since his first book, including a chapter on ASI safety, which featured as a keynote at the inaugural workshop on AI safety in Sydney.{{Cite web |last=Hutter |first=Marcus |date=9 December 2024 |title=ASI Safety via AIXI |url=https://www.youtube.com/watch?v=qHqv3GvWBTM |via=YouTube |language=en-US}}

{{Main|Hutter Prize}}

In 2006, Hutter announced the Hutter Prize for Lossless Compression of Human Knowledge, with a total of €50,000 in prize money.{{Cite magazine |last=Chiang |first=Ted |date=2023-02-09 |title=ChatGPT Is a Blurry JPEG of the Web |url=https://www.newyorker.com/tech/annals-of-technology/chatgpt-is-a-blurry-jpeg-of-the-web |access-date=2024-08-25 |magazine=The New Yorker |language=en-US |issn=0028-792X}}{{cite web|url=http://prize.hutter1.net|title=50'000€ Prize for Compressing Human Knowledge|author=Marcus Hutter |work=hutter1.net|access-date=29 November 2016}} In 2020, Hutter raised the prize money for the Hutter Prize to €500,000.{{Cite web |last=Sagar |first=Ram |date=2020-04-07 |title=Compress Data And Win Hutter Prize Worth Half A Million Euros |url=https://analyticsindiamag.com/hutter-prize-data-compression/ |access-date=2024-03-07 |website=Analytics India Magazine |language=en-US}}

• Solomonoff induction

• {{cite journal |author=Marcus Hutter |title=The Fastest and Shortest Algorithm for All Well-Defined Problems |journal=International Journal of Foundations of Computer Science |url=https://doi.org/10.1142/S0129054102001199 |date=2002 |volume=13 |issue=3 |pages=431–443 |publisher=World Scientific | doi=10.1142/S0129054102001199 |arxiv=cs/0206022 |s2cid=5496821 }}
• {{cite book |author=Marcus Hutter |title=Universal Artificial Intelligence: Sequential Decisions Based on Algorithmic Probability |url=https://books.google.com/books?id=NP53iZGt4KUC |date=2005 |publisher=Springer |isbn=9783540221395 }}
• {{cite journal |author=Joel Veness, Kee Siong Ng, Marcus Hutter, William Uther and David Silver |title=A Monte-Carlo AIXI Approximation |journal=Journal of Artificial Intelligence Research |date=2011 |volume=40 |pages=95–142 |publisher=AAAI Press |url=http://dx.doi.org/10.1613/jair.3125 |doi=10.1613/jair.3125 |arxiv=0909.0801 |s2cid=206618 }}
• {{cite journal |author=Shane Legg and Marcus Hutter |title=Universal Intelligence: A Definition of Machine Intelligence|url=http://dx.doi.org/10.1007/s11023-007-9079-x |date=2007 |volume=17 |issue=4 |pages=391–444 |journal=Minds and Machines |doi=10.1007/s11023-007-9079-x |arxiv=0712.3329 }}
• {{cite journal |author=Marcus Hutter |title=A Complete Theory of Everything (will be subjective)|date=2010 |volume=3 |issue=4 |pages=329–350 |journal=Algorithms |doi=10.3390/a3040329 |doi-access=free |arxiv=0912.5434 }}
• {{cite journal |author=Samuel Rathmanner and Marcus Hutter |title=A Philosophical Treatise of Universal Induction |journal=Entropy |date=2011 |volume=13 |issue=6 |pages=1076–1136 |publisher=MDPI |doi=10.3390/e13061076 |doi-access=free |arxiv=1105.5721 }}
• {{cite journal |author=Marcus Hutter |title=Can Intelligence Explode? |url=https://www.ingentaconnect.com/content/imp/jcs/2012/00000019/F0020001/art00010 |date=2012 |volume=19 |issue=1–2 |pages=143–166 |journal=Journal of Consciousness Studies |arxiv=1202.6177 }}
• {{cite journal |author=Peter Sunehag and Marcus Hutter |title=Rationality, Optimism and Guarantees in General Reinforcement Learning |journal=Journal of Machine Learning Research |url=http://jmlr.org/papers/v16/sunehag15a.html |date=2015 |volume=16 |pages=1345–1390 }}
• {{cite journal |author=Reinhard Hutter and Marcus Hutter |title=Chances and Risks of Artificial Intelligence – A Concept of Developing and Exploiting Machine Intelligence for Future Societies |url=https://www.mdpi.com/2571-5577/4/2/37 |date=2021 |volume=4 |issue=2 |pages=1–19 |journal=Applied System Innovation }}
• {{cite book |author=Marcus Hutter |title=An Introduction to Universal Artificial Intelligence |url=https://www.routledge.com/An-Introduction-to-Universal-Artificial-Intelligence/Hutter-Catt-Quarel/p/book/9781032607023 |date=2024 |publisher=Taylor & Francis |isbn=9781032607023 }}

{{Reflist}}

{{Authority control}}

{{DEFAULTSORT:Hutter, Marcus}}

Category:Living people

Category:Machine learning researchers

Category:Artificial intelligence researchers

Category:German computer scientists

Category:Technical University of Munich alumni

Category:Academic staff of the Australian National University

Category:1967 births

Category:DeepMind people

Category:AI safety scientists