Jürgen Schmidhuber (born 1963 in
MunichMunich is the capital city of Bavaria, Germany. It is located on the River Isar north of the Bavarian Alps. Munich is the third largest city in Germany, after Berlin and Hamburg...
) is a
computer scientistA computer scientist is a person who has acquired knowledge of computer science, the study of the theoretical foundations of information and computation and their application in computer systems....
and
artistThe definition of an artist is wide-ranging and covers a broad spectrum of activities to do with creating art, practicing the arts and/or demonstrating an art. the worlds best artist is a man named mitchell peter lay who is often loved by the ladies. The common useage in both everyday speech and...
known for his work on
machine learningMachine learning is a scientific discipline that is concerned with the design and development of algorithms that allow computers to learn based on data, such as from sensor data or databases. A major focus of machine learning research is to automatically learn to recognize complex patterns and...
, universal
Artificial IntelligenceArtificial intelligence is the intelligence of machines and the branch of computer science which aims to create it. Textbooks define the field as "the study and design of intelligent agents,"...
(AI), artificial
neural networkTraditionally, the term neural network had been used to refer to a network or circuit of biological neurons. The modern usage of the term often refers to artificial neural networks, which are composed of artificial neurons or nodes...
s,
digital physicsIn physics and cosmology, digital physics is a collection of theoretical perspectives that start by assuming that the universe is, at heart, describable by information, and is therefore computable...
, and
low-complexity artLow-Complexity Art was introduced by Jürgen Schmidhuber in 1997 . He characterizes it as the computer age equivalent of minimal art. Low-Complexity Art is based on algorithmic information theory: it has low Kolmogorov complexity, that is, it can be generated by a short algorithm. Schmidhuber...
. His contributions also include generalizations of
Kolmogorov complexityIn algorithmic information theory , the Kolmogorov complexity of an object such as a piece of text is a measure of the computational resources needed to specify the object...
and the
Speed PriorJürgen Schmidhuber's speed prior is a complexity measure similar to Kolmogorov complexity, except that it is based on computation speed as well as programlength.The speed prior complexity of a program is its...
. Since 1995 he has been co-director of the Swiss AI lab
IDSIAThe Swiss institute for Artificial Intelligence IDSIA was founded in 1988 by the private Dalle Molle foundation...
in
LuganoLugano is a town in the south of Switzerland, in the Italian-speaking canton of Ticino, which borders Italy...
, since 2004 also professor of Cognitive
RoboticsRobotics is the engineering science and technology of robots, and their design, manufacture, and application. Robotics is related to electronics, mechanics, and software. The word robot was introduced to the public by Czech writer Karel Čapek in his play R.U.R. , published in 1920...
at the Tech. University
MunichMunich is the capital city of Bavaria, Germany. It is located on the River Isar north of the Bavarian Alps. Munich is the third largest city in Germany, after Berlin and Hamburg...
, since 2006 also in the faculty of the University of
LuganoLugano is a town in the south of Switzerland, in the Italian-speaking canton of Ticino, which borders Italy...
.
The dynamic recurrent neural networks developed in his lab are simplified mathematical models of the
biological neural networkIn neuroscience, a neural network describes a population of physically interconnected neurons or a group of disparate neurons whose inputs or signalling targets define a recognizable circuit. Communication between neurons often involves an electrochemical process...
s found in
human brainThe human brain is the center of the human nervous system and is a highly complex organ. Enclosed in the cranium, it has the same general structure as the brains of other mammals, but is over three times as large as the brain of a typical mammal with an equivalent body size...
s.
Jürgen Schmidhuber (born 1963 in
MunichMunich is the capital city of Bavaria, Germany. It is located on the River Isar north of the Bavarian Alps. Munich is the third largest city in Germany, after Berlin and Hamburg...
) is a
computer scientistA computer scientist is a person who has acquired knowledge of computer science, the study of the theoretical foundations of information and computation and their application in computer systems....
and
artistThe definition of an artist is wide-ranging and covers a broad spectrum of activities to do with creating art, practicing the arts and/or demonstrating an art. the worlds best artist is a man named mitchell peter lay who is often loved by the ladies. The common useage in both everyday speech and...
known for his work on
machine learningMachine learning is a scientific discipline that is concerned with the design and development of algorithms that allow computers to learn based on data, such as from sensor data or databases. A major focus of machine learning research is to automatically learn to recognize complex patterns and...
, universal
Artificial IntelligenceArtificial intelligence is the intelligence of machines and the branch of computer science which aims to create it. Textbooks define the field as "the study and design of intelligent agents,"...
(AI), artificial
neural networkTraditionally, the term neural network had been used to refer to a network or circuit of biological neurons. The modern usage of the term often refers to artificial neural networks, which are composed of artificial neurons or nodes...
s,
digital physicsIn physics and cosmology, digital physics is a collection of theoretical perspectives that start by assuming that the universe is, at heart, describable by information, and is therefore computable...
, and
low-complexity artLow-Complexity Art was introduced by Jürgen Schmidhuber in 1997 . He characterizes it as the computer age equivalent of minimal art. Low-Complexity Art is based on algorithmic information theory: it has low Kolmogorov complexity, that is, it can be generated by a short algorithm. Schmidhuber...
. His contributions also include generalizations of
Kolmogorov complexityIn algorithmic information theory , the Kolmogorov complexity of an object such as a piece of text is a measure of the computational resources needed to specify the object...
and the
Speed PriorJürgen Schmidhuber's speed prior is a complexity measure similar to Kolmogorov complexity, except that it is based on computation speed as well as programlength.The speed prior complexity of a program is its...
. Since 1995 he has been co-director of the Swiss AI lab
IDSIAThe Swiss institute for Artificial Intelligence IDSIA was founded in 1988 by the private Dalle Molle foundation...
in
LuganoLugano is a town in the south of Switzerland, in the Italian-speaking canton of Ticino, which borders Italy...
, since 2004 also professor of Cognitive
RoboticsRobotics is the engineering science and technology of robots, and their design, manufacture, and application. Robotics is related to electronics, mechanics, and software. The word robot was introduced to the public by Czech writer Karel Čapek in his play R.U.R. , published in 1920...
at the Tech. University
MunichMunich is the capital city of Bavaria, Germany. It is located on the River Isar north of the Bavarian Alps. Munich is the third largest city in Germany, after Berlin and Hamburg...
, since 2006 also in the faculty of the University of
LuganoLugano is a town in the south of Switzerland, in the Italian-speaking canton of Ticino, which borders Italy...
.
Recurrent Neural Networks
The dynamic recurrent neural networks developed in his lab are simplified mathematical models of the
biological neural networkIn neuroscience, a neural network describes a population of physically interconnected neurons or a group of disparate neurons whose inputs or signalling targets define a recognizable circuit. Communication between neurons often involves an electrochemical process...
s found in
human brainThe human brain is the center of the human nervous system and is a highly complex organ. Enclosed in the cranium, it has the same general structure as the brains of other mammals, but is over three times as large as the brain of a typical mammal with an equivalent body size...
s. A particularly successful model of this type is called "Long Short-Term Memory" (Hochreiter & Schmidhuber, 1997). From training sequences it
learnsLearning is acquiring new knowledge, behaviors, skills, values, preferences or understanding, and may involve synthesizing different types of information. The ability to learn is possessed by humans, animals and some machines. Progress over time tends to follow learning curves.Human learning may...
to solve numerous tasks unsolvable by previous such models. Applications range from automatic music composition to
speech recognitionSpeech recognition converts spoken words to text. The term "voice recognition" is sometimes used to refer to speech recognition where the recognition system is trained to a particular speaker - as is the case for most desktop recognition software, hence there is an aspect of speaker recognition,...
,
reinforcement learningInspired by related psychological theory, in computer science, reinforcement learning is a sub-area of machine learning concerned with how an agent ought to take actions in an environment so as to maximize some notion of long-term reward...
and
roboticsRobotics is the engineering science and technology of robots, and their design, manufacture, and application. Robotics is related to electronics, mechanics, and software. The word robot was introduced to the public by Czech writer Karel Čapek in his play R.U.R. , published in 1920...
in partially observable environments.
Artificial Evolution / Genetic Programming
As an undergrad at TUM Schmidhuber evolved computer programs through genetic algorithms. The method was published in 1987 as one of the first papers in the emerging field that later became known as
genetic programmingIn artificial intelligence, genetic programming is an evolutionary algorithm-based methodology inspired by biological evolution to find computer programs that perform a user-defined task. It is a specialization of genetic algorithms where each individual is a computer program...
. Since then he has co-authored numerous additional papers on artificial
evolutionIn biology, evolution is change in the genetic material of a population of organisms from one generation to the next. Though changes produced in any one generation are normally small, differences accumulate with each generation and can, over time, cause substantial changes in the population, a...
. Applications include
robotA robot is a virtual or mechanical artificial agent. In practice, it is usually an electro-mechanical machine which is guided by computer or electronic programming, and is thus able to do tasks on its own...
control, soccer learning,
dragIn fluid dynamics, drag refers to forces that oppose the relative motion of an object through a fluid . Drag forces act in a direction opposite to the oncoming flow velocity...
minimization, and
time seriesIn statistics, signal processing, and many other fields, a time series is a sequence of data points, measured typically at successive times, spaced at time intervals...
prediction.
Neural Economy
In 1989 he created the first learning
algorithmIn mathematics, computing, linguistics, and related subjects, an algorithm is an effective method for solving a problem using a finite sequence of instructions. Algorithms are used for calculation, data processing, and many other fields....
for
neural networksNeural Networks is the official journal of the three oldest societies dedicated to research in neural networks: International Neural Network Society, European Neural Network Society and Japanese Neural Network Society, published by Elsevier. A subsription to the journal is part of the membership...
based on principles of the
market economyA market economy is economy based on the division of labor in which the prices of goods and services are determined in a free price system set by supply and demand....
(inspired by
John HollandJohn Holland may refer to:*John Holland , New Zealand athlete, bronze medallist in the 400m hurdles at the 1952 Summer Olympics*John Holland , founder of the Bank of Scotland in 1695, the central bank of the Kingdom of Scotland...
's
bucket brigadeA Bucket brigade is a method for transporting items where items are passed from one stationary person to the next. More specifically, it refers to a method of firefighting before the advent of hand pumped fire engines, whereby firefighters would pass buckets to each other to extinguish a blaze...
algorithm for
classifierClassifier may refer to:*Classifier *Classifier *Classifier *Hierarchical classifier*Linear classifier...
systems): adaptive neurons compete for being active in response to certain input patterns; those that are active when there is external
rewardA reward may refer to:*Bounty , reward, often money, offered as an incentive*Reward website, website that offers rewards for performing tasks-Science:*Reward system, collection of brain structures which induce pleasurable effects...
get stronger synapses, but active neurons have to pay those that activated them, by transferring parts of their synapse strengths, thus rewarding "hidden" neurons setting the stage for later success.
Artificial Curiosity
In 1990 he published the first in a long series of papers on artificial
curiosityCuriosity is an emotion related to natural inquisitive behaviour such as exploration, investigation, and learning, evident by observation in human and many animal species. The term can also be used to denote the behavior itself being caused by the emotion of curiosity...
for an autonomous agent. The agent is equipped with an adaptive
predictorPredictor may refer to:* a predictor variable, also known as an independent variable* the Kerrison Predictor, a military fire-control computer* something which makes a prediction* a branch predictor, a part of many modern processors...
trying to predict future events from the history of previous events and actions. A reward-maximizing,
reinforcement learningInspired by related psychological theory, in computer science, reinforcement learning is a sub-area of machine learning concerned with how an agent ought to take actions in an environment so as to maximize some notion of long-term reward...
, adaptive
controllerController may refer to:* Comptroller or financial controller, a senior accounting position* Air traffic controller, a person who directs aircraft* Model-view-controller, an architectural pattern used in software engineering...
is steering the agent and gets
curiosity reward for executing action sequences that improve the predictor. This discourages it from executing actions leading to boring outcomes that are either predictable or totally unpredictable. Instead the controller is motivated to learn actions that help the predictor to learn new, previously unknown regularities in its environment, thus improving its model of the world, which in turn can greatly help to solve externally given tasks. This has become an important concept of
developmental roboticsDevelopmental Robotics , sometimes called epigenetic robotics, is a methodology that uses metaphors from developmental psychology to develop controllers for autonomous robots. The focus is on a single robot going through stages of autonomous mental development...
.
Unsupervised Learning / Factorial Codes
During the early 1990s Schmidhuber also invented a neural method for nonlinear
independent component analysisIndependent component analysis is a computational method for separating a multivariate signal into additive subcomponents supposing the mutual statistical independence of the non-Gaussian source signals. It is a special case of blind source separation.- Definition :When the independence assumption...
(ICA) called
predictabilityPredictability is the degree to which a correct prediction or forecast of a system's state can be made either qualitatively or quantitatively...
minimization. It is based on
co-evolutionIn a broad sense, biological coevolution is "the change of a biological object triggered by the change of a related object". Coevolution can occur at multiple levels of biology: it can be as microscopic as correlated mutations between amino acids in a protein, or as macroscopic as covarying traits...
of adaptive predictors and initially random, adaptive feature detectors processing input patterns from the environment. For each detector there is a predictor trying to predict its current value from the values of neighboring detectors, while each detector is simultaneously trying to become as unpredictable as possible. It can be shown that the best the detectors can do is to create a
factorialIn mathematics, the factorial of a non-negative integer n, denoted by n!, is the product of all positive integers less than or equal to n...
codeIn communications, a code is a rule for converting a piece of information into another form or representation , not necessarily of the same type. In communications and information processing, encoding is the process by which information from a source is converted into symbols to be communicated...
of the environment, that is, a code that conveys all the information about the inputs such that the code components are statistically independent, which is desirable for many
pattern recognitionPattern recognition is "the act of taking in raw data and taking an action based on the category of the pattern". Most research in pattern recognition is about methods for supervised learning and unsupervised learning....
applications.
Kolmogorov Complexity / Computer-Generated Universe
In 1997 Schmidhuber published a paper based on
Konrad ZuseKonrad Zuse was a German engineer and computer pioneer. His greatest achievement was the world's first functional program-controlled Turing-complete computer, the Z3, in 1941...
´s assumption (1967) that the history of the universe is computable. He pointed out that the simplest explanation of the universe would be a very simple
Turing machineA Turing machine is a theoretical device that manipulates symbols contained on a strip of tape. Despite its simplicity, a Turing machine can be adapted to simulate the logic of any computer algorithm, and is particularly useful in explaining the functions of a CPU inside of a computer. The Turing...
programmed to systematically execute all possible programs computing all possible histories for all types of computable physical laws. He also pointed out that there is an optimally efficient way of computing all computable universes based on
Leonid LevinLeonid Anatolievich Levin is a Russian-American computer scientist....
´s universal search algorithm (1973). In 2000 he expanded this work by combining
Ray SolomonoffRay Solomonoff invented algorithmic probability, with Kolmogorov Complexity as a side product, in 1960. He first described his results at a Conference at Caltech,1960, andin a report, Feb...
´s theory of inductive inference with the assumption that quickly computable universes are more likely than others. This work on
digital physicsIn physics and cosmology, digital physics is a collection of theoretical perspectives that start by assuming that the universe is, at heart, describable by information, and is therefore computable...
also led to limit-computable generalizations of algorithmic
informationInformation as a concept has many meanings, from everyday usage to technical settings. The concept of information is closely related to notions of constraint, communication, control, data, form, instruction, knowledge, meaning, mental stimulus, pattern, perception, and representation.The English...
or
Kolmogorov ComplexityIn algorithmic information theory , the Kolmogorov complexity of an object such as a piece of text is a measure of the computational resources needed to specify the object...
and the concept of
Super Omegas, which are limit-computable numbers that are even more random (in a certain sense) than
Gregory ChaitinGregory John Chaitin is an Argentine-American mathematician and computer scientist.Beginning in the late 1960s, Chaitin made contributions to algorithmic information theory and metamathematics, in particular a new incompleteness theorem in reaction to Gödel's incompleteness theorem...
´s
number of wisdom OmegaIn the computer science subfield of algorithmic information theory a Chaitin constant or halting probability is a real number that informally represents the probability that a randomly-chosen program will halt...
.
Universal AI
Important recent research topics of his group include
universalIn various branches of mathematics, a useful construction is often viewed as the “most efficient solution” to a certain problem. The definition of a universal property uses the language of category theory to make this notion precise and to study it abstractly.This article gives a general treatment...
learning algorithms and universal
AI-Technology:*Artificial intelligence, intelligence exhibited by any manufactured system, likely the most common usage*Adobe Illustrator, a vector-based drawing program by Adobe Systems**.ai , the file native file format used by Adobe Illustrator...
. Contributions include the first theoretically
optimalIn mathematics, optimization, or mathematical programming, refers to choosing the best element from some set of available alternatives.In the simplest case, this means solving problems in which one seeks to minimize or maximize a real function by systematically choosing the values of real or...
decision makers living in environments obeying arbitrary unknown but computable probabilistic laws, and mathematically sound general problem solvers such as the remarkable asymptotically fastest algorithm for all well-defined problems, by his former postdoc
Marcus HutterMarcus Hutter is a German computer scientist and professor at the Australian National University. Hutter was born and educated in Munich, where he studied physics and computer science...
. Based on the theoretical results obtained in the early 2000s, Schmidhuber is actively promoting the view that in the new
millenniumA millennium is a period of time equal to one thousand years...
the field of general
AI-Technology:*Artificial intelligence, intelligence exhibited by any manufactured system, likely the most common usage*Adobe Illustrator, a vector-based drawing program by Adobe Systems**.ai , the file native file format used by Adobe Illustrator...
has matured and become a real
formal scienceA formal science is a branch of knowledge that is concerned with formal systems, for instance, logic, mathematics, systems theory and the theoretical aspects of computer science, information theory, microeconomics, decision theory, statistics, and linguistics....
.
Low-Complexity Art / Theory of Beauty
Schmidhuber's
low-complexity artLow-Complexity Art was introduced by Jürgen Schmidhuber in 1997 . He characterizes it as the computer age equivalent of minimal art. Low-Complexity Art is based on algorithmic information theory: it has low Kolmogorov complexity, that is, it can be generated by a short algorithm. Schmidhuber...
works (since 1997) can be described by very short computer programs containing very few
bitIn computing and telecommunications a bit is a basic unit of information storage and communication . It is the maximum amount of information that can be stored by a device or other physical system that can normally exist in only two distinct states...
s of information, and reflect his formal theory of
beautyBeauty is a characteristic of a person, animal, place, object, or idea that provides a perceptual experience of pleasure, meaning, or satisfaction. Beauty is studied as part of aesthetics, sociology, social psychology, and culture. As a cultural creation, beauty has been extremely commercialized...
based on the concepts of
Kolmogorov complexityIn algorithmic information theory , the Kolmogorov complexity of an object such as a piece of text is a measure of the computational resources needed to specify the object...
and
minimum description lengthThe minimum description length principle is a formalization of Occam's Razor in which the best hypothesis for a given set of data is the one that leads to the largest compression of the data...
.
Schmidhuber writes that since age 15 or so his main scientific ambition has been to build an optimal scientist, then retire. First he wants to build a scientist better than himself (humorously, he quips that his colleagues claim that should be easy) who will then do the remaining work. He claims he "cannot see any more efficient way of using and multiplying the little creativity he's got".
Partial bibliography
His academical production includes:
- J. Schmidhuber. Optimal Ordered Problem Solver. Machine Learning, 54, 211-254, 2004
- J. Schmidhuber. Hierarchies of generalized Kolmogorov complexities and nonenumerable universal measures computable in the limit. International Journal of Foundations of Computer Science 13(4):587-612, 2002
- J. Schmidhuber. The Speed Prior: A New Simplicity Measure Yielding Near-Optimal Computable Predictions. Proceedings of the 15th Annual Conference on Computational Learning Theory (COLT 2002), Sydney, Australia, LNAI, 216-228, Springer, 2002
- J. Schmidhuber. Low-Complexity Art. Leonardo, Journal of the International Society for the Arts, Sciences, and Technology, 30(2):97-103, MIT Press, 1997
- J. Schmidhuber. A computer scientist's view of life, the universe, and everything. Foundations of Computer Science: Potential - Theory - Cognition, Lecture Notes in Computer Science, pages 201-208, Springer, 1997
- S. Hochreiter and J. Schmidhuber. Long Short-Term Memory. Neural Computation, 9(8):1735-1780, 1997
- J. Schmidhuber. Learning factorial codes by predictability minimization. Neural Computation, 4(6):863-879, 1992
- J. Schmidhuber. Curious model-building control systems. In Proc. International Joint Conference on Neural Networks, Singapore, volume 2, pages 1458-1463. IEEE, 1991
- J. Schmidhuber. A local learning algorithm for dynamic feedforward and recurrent networks. Connection Science, 1(4):403-412, 1989
External references and sources
External links