Skip to content
Last Dictionary
Plane
AGI & SuperintelligenceMathematics & StatisticsPhilosophy & Mind

universal intelligence

nounAlso: Legg–Hutter intelligence, universal intelligence measure, Legg–Hutter definition of intelligence

In plain wordsA mathematical definition of intelligence as an agent's average ability to achieve goals across every possible environment, with simpler environments counting more.

  1. A formal measure of an agent's general intelligence, defined as its expected performance across all computable reward-giving environments, each weighted by its simplicity. An agent's score is the sum, over environments μ, of 2 raised to minus the Kolmogorov complexityKolmogorov complexity, n.: How briefly something can be described by a computer program that recreates it exactly: a million zeros needs only a tiny program, while a truly random string needs one about as long as itself. of μ, multiplied by the expected total reward the agent obtains in μ; performance in simple environments counts most, but no environment is ignored. The measure formalizes the informal definition its authors distilled from some seventy existing ones: 'Intelligence measures an agent's ability to achieve goals in a wide range of environments.' It is maximized by AIXIAIXI, prop. n.: A mathematical description of a perfect learning agent that weighs every possible explanation of its experience and acts to get the most future reward. It cannot actually be built, but it defines an ideal. and is incomputable, though approximate tests have been proposed. Criticisms include its dependence on the choice of reference machine, its treatment of intelligence as reward-seeking, and the absence of any weighting toward environments that matter to humans.

    Under universal intelligence, a system that excels only at chess scores barely above one that does nothing.

Υ(π):=∑μ∈E2−K(μ) Vμπ\Upsilon(\pi) := \sum_{\mu \in E} 2^{-K(\mu)} \, V^{\pi}_{\mu}

Origin

Proposed by Shane Legg and Marcus Hutter in 'Universal Intelligence: A Definition of Machine Intelligence' (Minds and Machines, 2007), after their survey of definitions of intelligence; Legg developed it in his 2008 doctoral thesis, Machine Super Intelligence.

Related

See also
  • AIXIAIXI, prop. n.: A mathematical description of a perfect learning agent that weighs every possible explanation of its experience and acts to get the most future reward. It cannot actually be built, but it defines an ideal.
  • artificial general intelligenceartificial general intelligence, n.: AI that could learn and do almost any mental task a capable adult can, rather than excelling at one narrow job. Experts disagree on the exact bar and on when, or whether, it will arrive.
  • Kolmogorov complexityKolmogorov complexity, n.: How briefly something can be described by a computer program that recreates it exactly: a million zeros needs only a tiny program, while a truly random string needs one about as long as itself.
  • Solomonoff inductionSolomonoff induction, n.: A mathematical recipe for ideal prediction: consider every computer program that could have produced the data so far, trust shorter programs more, and predict what they would output next.
  • intelligenceintelligence, n.: The ability to learn, understand, reason and solve problems — especially new ones — in order to reach goals. Experts disagree about how to define it precisely, for people and even more for machines.
  • Shane LeggShane Legg, prop. n.: A New Zealand-born AI researcher who co-founded DeepMind and helped make 'artificial general intelligence' a mainstream term and goal.

Sources

  1. Legg, S. & Hutter, M. (2007). Universal Intelligence: A Definition of Machine Intelligence. Minds and Machines 17(4).
  2. Legg, S. & Hutter, M. (2007). A Collection of Definitions of Intelligence. In Advances in Artificial General Intelligence. IOS Press.

Mentioned in

artificial general intelligenceartificial general intelligence, n.: AI that could learn and do almost any mental task a capable adult can, rather than excelling at one narrow job. Experts disagree on the exact bar and on when, or whether, it will arrive., intelligenceintelligence, n.: The ability to learn, understand, reason and solve problems — especially new ones — in order to reach goals. Experts disagree about how to define it precisely, for people and even more for machines., Shane LeggShane Legg, prop. n.: A New Zealand-born AI researcher who co-founded DeepMind and helped make 'artificial general intelligence' a mainstream term and goal., AIXIAIXI, prop. n.: A mathematical description of a perfect learning agent that weighs every possible explanation of its experience and acts to get the most future reward. It cannot actually be built, but it defines an ideal., g factorg factor, n.: The common factor behind the finding that people who do well on one kind of mental test tend to do well on others. Researchers also look for a similar general factor across AI models., Marcus HutterMarcus Hutter, prop. n.: A German computer scientist who defined AIXI, a mathematical ideal of a perfectly intelligent agent, and who argues that compressing data well is a measure of intelligence.