/* Busy Coder */
©2017 Jiri Jelinek


* Paranormal *

AI Glossary

The purpose of this page is not to present details, but give a general idea.

A | B | C | D | E | F | G | H | I | J | K | L | M | N | O | P | Q | R | S | T | U | V | W | X | Y | Z


A search algorithm to find the shortest path through a search space using a heuristic.

Anything that can be viewed as perceiving its environment through sensors and acting upon that environment through effectors.
[Russel, Norvig 1995]

Alpha-Beta Pruning
A method of limiting search in the MiniMax algorithm. If done optimally, it reduces the branching factor from B to the square root of B.

Artificial Intelligence (AI)
The science and engineering of making intelligent machines.
[John McCarthy, who coined the term in 1956]

Intelligence implemented by an Intelligent System.
[Jiri Jelinek 2000]

One for fun:
AI is the science of how to get machines to do the things they do in the movies.
See AI in fiction.

AI effect
People's inclination to discount advances in AI after they've been accomplished, as an apparent consequence of coming to understand how that aspect of intelligence works. This discounting is often expressed by redefining the requirements for or description of "intelligent" behaviour. The idea is expressed by Tesler's Theorem: "AI is whatever machines haven't done yet." (which is often being misquoted).
[sources: Wikipedia and Lawrence G. Tesler's resume]

Artificial General Intelligence (AGI)
General intelligence implemented by an Intelligent System.
[Jiri Jelinek 2000]

A statement that is assumed to be true and can later be used along with theorems to prove other theorems.
[Gary William Flake, The Computational Beauty of Nature]


Bayesian Network (or a belief network)
A probabilistic graphical model that represents a set of variables and their probabilistic independencies.
[Wikipedia 2007]

Brain-Generated Music (BGM)
A music technology pioneered by NeuroSonics, Inc., that creates music in response to the listener's brain waves. This brain-wave biofeedback system appears to evoke the Relaxation Response by encouraging the generation of alpha waves in the brain.
[source: 1999 R. Kurzweil]


Chinese Room
An experiment wherein a man who knows no Chinese blindly replies to Chinese questions according to certain rules. The man can reply based on rules, but still has no understanding on what he is saying. Similarly, symbol manipulation (because it is simply according to rules) does not necessarily imply understanding.

Note: Different level of concept understanding doesn't necessarily prevent successful problem solving.

An approach to learning that seeks to place objects into meaningful groups automatically based on their similarity.
[source: 1999 R. Kurzweil]

A collection of processes by which sensory input is transformed, reduced, elaborated, stored, recovered, and used.
[Neisser, 1967]

An approach to studying intelligence and to creating intelligent solutions to problems. Connectionism is based on storing problem-solving knowledge as a pattern of connections among a very large number of simple processing units operating in parallel.
[source: 1999 R. Kurzweil]

One who breaks security on a system. Coined (in 1985) by hackers in defense against the journalistic misuse of the term hacker.


Data Mining
Data Mining - also known as "Knowledge Discovery in Databases" (KDD) or simply "Knowledge Discovery" has been defined as "The nontrivial extraction of implicit, previously unknown, and potentially useful information from data."
[Frawley, Piatetsky, Shapiro]


The part of a system that can translate messages into actions that can manipulate a system or an environment.
[Gary William Flake, The Computational Beauty of Nature]

An approach to AI that maintains that the only way to create general intelligence is to use programs with 'bodies' in the real world (i.e. robots).
[First and most strongly put forth by Rod Brooks at MIT]

A qualia based mental reaction which limits the attention allocation scope. The outcome can be positive as well as negative, depending on what gets excluded and how relevant it is at the time. The scope limitation can greatly improve performance (which is sometimes critical), but it can also temporarily impact subject's value system and cause decisions [potentially highly] incompatible with other [possibly more typical] mental states of the same subject - which can cause very undesirable [long-term] consequences. Human emotions are nature's attempt to implement heuristics, but not done very well. Phenomenons like emotion-induced blindness (which can be triggered by a roadside billboard etc.) can kill you under certain circumstances. When working on AI, we can do better.
[Jiri Jelinek 2007]

If that which is under study is a system, then the rest of the world is the environment.
[Gary William Flake, The Computational Beauty of Nature]

Expert System
A system capable of giving advice in a particular knowledge domain, by virtue of the fact that it contains knowledge provided by a human expert in this domain.


Formal System
A mathematical formalism in which statements can be constructed and manipulated with logical rules.
[Gary William Flake, The Computational Beauty of Nature]

Note: Some formal systems are built around a few basic axioms (such as Euclidean geometry) and can be expanded with theorems that can be deduced through proofs.

Fuzzy Logic
A logic based on "degrees of truth."
(unlike absolute Boolean True/False logic)


Game Theory
A mathematical formalism used to study human games, economics, military conflicts, and biology. The goal of game theory is to find the optimal strategy for one player to use when his opponent also plays optimally.
[Gary William Flake, The Computational Beauty of Nature]

General Intelligence
A cross-domain ability to gain knowledge in one context and correctly apply it in another.
[Jiri Jelinek 2007]

The ability to achieve complex goals in complex environments.
[Benjamin Goertzel]

Godel's incompleteness theorem
A theorem postulated by Kurt Godel, a Czech mathematician, that states that in a mathematical system powerful enough to generate the natural numbers, there inevitably exist propositions that can be neither proved nor disproved.
[source: 1999 R. Kurzweil]

Good Old-Fashioned Artificial Intelligence (GOFAI)
A branch of AI which mainly deals with symbolic problems and representations.


Hack mode
Zen-like state of total focus on The Problem. This is why every good human-hacker is part mystic. Some aspects of hacker etiquette will appear quite odd to an observer unaware of the high value placed on hack mode. Being yanked out of 'deep hack mode' may be experienced as a physical shock.

Hacker (not to be confused with a Script Kiddie)
An entity that enjoys exploring the details of programmable systems and how to stretch their capabilities, as opposed to most users, who prefer to learn only the minimum necessary.

One of the most important ethical duties of hackers is to share their expertise by writing open-source and facilitating access to information and computing resources wherever possible.
Many people think that being a hacker also means being someone who breaks a law. This is a misunderstanding. Many hackers do illegal things and many do not. Lots of hackers share feelings which are expressed in the famous Hacker Manifesto.

Note: Originally, someone who makes furniture with an axe.


Ability to act in order to achieve a goal.
[Jiri Jelinek 2004]

Note: There are many different definitions of intelligence (some collected here). Some researchers (including myself) consider a thermostat to be an intelligent system (very limited one of course), some argue against it. Some believe that the term "intelligence" should not be used at all and that the AI field should have been named differently to avoid confusion and prevent the endless arguing about how to define it. Since there is no official agreement on the "I" definition, you are free to pick one which makes sense from your AI R&D perspective. Whatever you go with, expect a crowd of opponents. Let's keep in mind that the development of smart machines is a lot more important than arguing about the "I" definition.

Intelligent Agent (IA)
See Intelligent System.

Intelligent System (IS)
A system capable of intelligent behavior.

Note: An IS can be as simple as a reflex machine (e.g. thermostat) but also very complex (e.g. a team of scientists or companies working together towards a goal.



Kolmogorov Complexity
(also known as descriptive complexity, Kolmogorov-Chaitin complexity, stochastic complexity, algorithmic entropy, or program-size complexity) of an object (such as a piece of text) is a measure of the computational resources needed to specify the object.


Lambda Calculus
A model of computation that is capable of universal computation.
[Gary William Flake, The Computational Beauty of Nature]

Note: The Lisp programming language was inspired by Lambda calculus.

Language Acquisition
A sub-branch of AI; traditionally computational linguists tried to make computers understand human language by giving the computer grammar rules. Language acquisition is a technique for the computer to generate the grammar rules itself.

A system that performs a subjective behavior.
[Jiri Jelinek 2000]

Note: Biologists define it as a complex interaction of nucleic acids and proteins.

A science that deals with the principles and criteria of validity of inference.
Logic says which forms of inference are valid and which are not.

Note: There are many systems of logic, e.g. Aristotelian, Formal, Mathematical, Philosophical, Predicate, Computability, Multi-valued and more.


Menu-based Natural Language Interface. Users select words and phrases from menus to construct sentences the system will understand - this takes the guess work out of NLI technology.

An algorithm for game playing in games with perfect information. See Alpha-Beta Pruning.

Modus Ponens
An inference rule that says: if you know x and you know that 'If x is true then y is true' then you can conclude y.

Moore's Law
First postulated by former Intel CEO Gordon Moore in the mid-1960s, Moore's Law is the prediction that the size of each transistor on an integrated circuit chip will be reduced by 50 percent every twenty-four months. The result is the exponentially growing power of integrated circuit-based computation over time. Moore's Law doubles the number of components on a chip as well as the speed of each component. Both of these aspects double the power of computing, for an effective quadrupling of the power of computation every twenty-four months.
[source: 1999 R. Kurzweil]

Moravec's paradox
A principle in artificial intelligence and robotics, stating that normal intuitions about which problems are "easy" or "hard" do not apply to machines, in fact, what is easiest for people is hardest for machines, and what is hardest for people may be easy for machines.
[The principle was first articulated by Hans Moravec]


Narrow AI
An AI displaying intelligence regarding specific tasks in relatively narrow domains.
[Benjamin Goertzel]

Natural Language Processing (NLP)
A subfield of AI and computational linguistics. It studies the problems of automated generation and understanding of natural human languages.


Objective Experience
The experience of an entity as observed by another entity, or measurement apparatus.
[source: 1999 R. Kurzweil]


A qualia-based perception that causes avoidance behavior.
[Jiri Jelinek 2007]

A subfield of AI concerned with systems that construct sequences of actions to achieve goals.

A sequence of statements in which each subsequent statement is derivable from one of the previous statements or from an axiom of a formal system.
[Gary William Flake, The Computational Beauty of Nature]

Note: The final statement of a proof is usually the theorem that one has set out to prove.


Qualification problem
The problem with impossibility of listing all the preconditions required for a real-world action to have its intended effect.

A "quantum bit," used in quantum computing, that is both zero and one at the same time, until quantum decoherence (direct or indirect observation by a conscious observer) causes each quantum bit to disambiguate into a state of zero or one. One qu-bit stores two possible numbers (zero and one) at the same time. N qu-bits stores 2N possible numbers at the same time. Thus an N qu-bit quantum computer would try 2N possible solutions to a problem simultaneously, which gives the quantum computer its enormous potential power.
[source: 1999 R. Kurzweil]

Quantum Computing
A revolutionary method of computing, based on quantum physics, that uses the ability of particles such as electrons to exist in more than one state at the same time. See Qu-bit.
[source: 1999 R. Kurzweil]


A set of information perceived from senses.
[Jiri Jelinek 2000]

The faculty or process of drawing logical inferences.

Deductive reasoning = from the general to the particular.
Inductive reasoning = from the particular to the general.
Abductive reasoning = from the best available information to the best explanation.
Case-based reasoning (CBR) = technique whereby "cases" similar to the current problem are retrieved and their "solutions" modified to work on the current problem.

A mechanical or virtual, artificial agent.

Russell's Paradox
The ambiguity created by the following question: Does a set that is defined as "all sets that do not include themselves" include itself as a member? Russell's paradox motivated Bertrand Russell to create a new theory of sets.
[source: 1999 R. Kurzweil]


Script Kiddie
A Script Kiddie is typically a young male, usually not by any means a computer expert, who exploits weaknesses in security systems discovered by someone else. A script kiddie is a wannabe hacker who scans the Internet for compromised systems using freely available tools. Good examples of their often lacking knowledge are cases where script kiddies used a sophisticated hacking software to break into an operating system and then proceeded to attempt to run commands that weren't even compatible with the system.

The use of an internal representation (IR) of self, where the IR is linked to real features of the system.
[Jiri Jelinek 2004]

A sub-system which is able to 'perceive data from an information source' and/or 'recognize certain patterns in processed data' and provide the result to its parent system. The information source can be anything considered external to the sense which may include another sub-systems of the same parent system.
[Jiri Jelinek 2002]

Society of mind
A theory of the mind proposed by Marvin Minsky in which intelligence is seen to be the result of proper organization of a large number (a society) of other minds, which are in turn comprised of yet simpler minds. At the bottom of this hierarchy are simple mechanisms, each of which is by itself unintelligent.
[source: 1999 R. Kurzweil]

In a formal system, a string that is formed according to well-defined rules such that it is legal for the language that is the formal system. For example, in the formal system of arithmetic, the expression "5 + 3 * (2 - 4)" is a valid and well-formed statement, but "5 + )3 * * (2 (- 4)" is not.
[Gary William Flake, The Computational Beauty of Nature]

Any sequence of letters, numbers, digits, bits, or symbols.
[Gary William Flake, The Computational Beauty of Nature]

Strong AI (Hard AI)
It's described as an AI system which is able to think on a level (at least) equal to humans.

Ability to prefer.
[Jiri Jelinek 2000]

Part of the universe, with a limited extension in space and time.
[Glossary of Walter Fritz, 2000]


A statement in a formal system that has proof.
[Gary William Flake, The Computational Beauty of Nature]

A search for a way to achieve a goal.
[Jiri Jelinek 2000]

Turing Machine
A device for transforming one string of symbols into another string according to a predetermined set of rules. A typical Turing machine can be represented as a device that reads and writes symbols on an infinite tape and has a control unit that can take on a finite number of states. The control unit essentially consists of a table that tells the machine what to do. The first part of the instruction specifies what the machine should write, the second part determines whether the machine stays in the same state or shifts to another state (usually, a different set of instructions) and the third part specifies whether the printing head is to shift one frame to the left or to the right along the tape. Turing machines embody a method of mathematical reasoning. Given a large (but finite) amount of time, a Turing machine is capable of any computation that can be done by human (and any computer, no matter how powerful).

Turing Test
The Turing test requires (at least) 3 players: a human interrogator, a human respondent, and a computer respondent. The interrogator, who stays in a room apart from the two players, must determine which respondent is the computer. If the interrogator cannot correctly distinguish between the two respondents, the computer is deemed "intelligent." The test should take at least 5 minutes.
[proposed by Alan Turing, Computing Machinery and Intelligence 1950]

Note: As Alan Turing stated, the criterion provides a sufficient, but not necessary, condition for intelligent behavior. In other words, failing the test proves nothing.
Turing test is a variant of an older test known as "Imitation Game." In the original Imitation Game, the two respondents were a man and a woman; the job of the interrogator (a human of either sex) was to determine which was which.


Everything that exists.
[Oxford dictionary, 1995]



Weak AI (Soft AI)
An AI system whose capabilities are not intended to match or exceed the capabilities of human beings.