Next:	General Intelligence and Seed AI		Bookmark
Up:	General Intelligence and Seed AI		Monolithic
Prev:	Version History

Appendix A: Glossary

NOTE:

If a referenced item does not appear in this glossary, it may be defined in Creating Friendly AI.

• affector
• API
• atto
• Bayesian binding
• Bayesian Probability Theorem
• cache
• CFAI
• codelet
• computational horizon
• computational temperature
• Consensus
• continuous
• Copycat
• counterfactual
• crystalline
• cytoarchitecture
• declarative
• Deep Blue
• discrete
• e.g.
• Eurisko
• exa
• femto
• GEB
• gender-neutral pronouns
• giga
• GISAI
• granularity horizon
• hertz
• heuristic
• holism
• horizon
• i.e.
• iff
• instantiate
• intelligence
• intentionality
• Kasparov
• kilo
• Lakoff and Johnson
• latency
• Law of Pragmatism
• Life
• LISP tokens
• Marr
• mega
• micro
• microworld
• milli
• mindstuff
• MITECS
• nano
• Necessary, But Not Sufficient
• ontology
• ontotechnology
• orthogonal
• past light cone
• peta
• Physicist's Paradigm
• pico
• predictive horizon
• procedural
• Q.E.D.
• qualia
• qualitative
• quantitative
• reductholism
• reductionism
• reflection
• relevance horizon
• RNUI
• salience
• scalar
• search trees
• seed AI
• sensory modality
• SIAI
• space of simultaneity
• SPDM
• stochastic
• structural
• subjunctive
• tera
• three
• time
• Turing-computability
• ve
• ver
• verself
• vis
• world-model

• affector: 
      To be defined in Unimplemented section: Causality.  An "affector" is a factor, something that affects things, something that has effects.  Subtly different from describing something as a cause or as a factor; somewhere between the two.  The term is useful, in that I often find no other term has the exact connotation I want to use.  To say that A causes B is to say that A completely accounts for B.  A can affect B without completely accounting for B.  Also, to describe something as a "cause" is generally to describe it as an intermediate point on a causal chain; that is, a cause is a combination of effect and affector; to describe something as an "affector" is to look only forward, rather than looking both forward and backward.  "Cause" is often overloaded with other meanings and does not have the precision of "affector".  Finally, "cause" is a noun and a verb where "affector" is clearly a noun, a difference of terminology which subtly affects the way we think about causality.  So please excuse the jargon.
• API: 
      Application Programming Interface.  The membrane that separates libraries from the programmer.  A set of functions, objects, methods, properties, formats, and data structures that a programmer can use to communicate with an operating system, a commercial code library, an Internet browser, et cetera.  Essentially, an API is the structure of the inputs and outputs of a system.
• atto: 
      10^-18.  One-quintillionth; a thousandth of a "femto".  See the Hacker's Dictionary for the full list of quantifiers.
• Bayesian binding: 
      The strength of the binding between a piece of information and the conclusion derived from it, under the Bayesian Probability Theorem.  The more improbable a given pattern is, the less likely that the pattern was produced by a pure coincidence.  Suppose that you flip coin A thirty times, and then flip coin B thirty times, and the pattern of heads and tails matches exactly.  Since this is a billion-to-one improbability for fair, independent coins, there is a very strong Bayesian binding between the two patterns, leading an observer to conclude that there is almost certainly some cause tying the two coins together.
      In discussions of meta-rationality, the Bayesian Probability Theorem is used to estimate the strength of the binding between your beliefs and reality - i.e., the extent to which the fact that you believe X licenses you to conclude that X is true.  Suppose that I observe myself to say, "The sky is green."  If I know that I believe in the greenness of the sky so strongly that I would declare the sky green even though it were purest sapphire blue, then my observing myself to say "The sky is green" says nothing - according to the BPT - about the actual likelihood that the sky is green.  This effect - that I will see myself believing "The sky is green" - is predicted in both the groups where the sky is green and the groups where the sky is blue; thus the observation does nothing to indicate which group the actual sky falls into.  If, on the other hand, I don't care much about the color of the sky, then I am only likely to say that the sky is green if it's actually green, and my observing myself to make this statement is strong evidence in favor of the greenness of the sky.
      What was that about faith?
• Bayesian Probability Theorem: 
      The Bayesian Probability Theorem relates observed effects to the a priori probabilities of those effects in order to estimate the probabilities of underlying causes.  For example, suppose you know the following:  1% of the population has cancer.  The probability of a false negative, on a cancer test, is 2%.  The probability of a false positive, on a cancer test, is 10%.  Your test comes up positive.  What is the probability that you have cancer?
      The instinctive human reaction is terror.  After all, the probability of a false positive is only 10%; isn't the probability that you have cancer therefore 90%?
      The Bayesian Probability Theorem demonstrates why this reasoning is flawed.  In a group of 10,000 people, 100 will have cancer and 9,900 will not have cancer.  If cancer tests are administered to the 10,000 people, four groups will result.  First, a group of 8,910 people who do not have cancer and who have a negative test result.  Second, a group of 990 who do not have cancer and who have a positive test result.  Third, a group of 2 who have cancer and who have a negative test result.  Fourth, a group of 98 who have cancer and who have a positive test result.
      Before you take the test, you might belong to any of the four groups; the Bayesian Probability Theorem says that your probability of having cancer is equal to (2 + 98)/(8,910 + 990 + 2 + 98), 1/100 or 1%.  If your test comes up positive, it is now known that you belong to either group 2 or group 4.  Your probability of having cancer is (98)/(990 + 98), 49/544 or approximately 9%.  If your test comes up negative, it is known that you belong to either group 1 or group 3; your probability of having cancer is 2/8,912 or around .02%.
      Colloquially, the good Reverend Bayes is invoked wherever prior probabilities have a major influence on the outcome of a question.  However, the Bayesian Probability Theorem has a much wider range of application - in normative reasoning, the BPT controls the binding between all sensory information and all beliefs.  Normative reasoners are often called "Bayesian reasoners" for this reason.  The Bayesian Probability Theorem is so ubiquitous and so intricate that I would cite it as one of the very, very few counterexamples to the Law of Pragmatism - in CFAI 3.1.4: Bayesian reinforcement, for example, I discuss how some of the functionality of pain and pleasure, though implemented in a separate hardwired system in humans, could emerge directly from the BPT in normative reasoners (i.e., in Friendly AIs).
• cache: 
      To "cache" a result is to store it for later use.  For example, your Web browser has a "cache folder" containing files that you have already downloaded from the Internet; when your browser encounters a URL that it's already followed, it can retrieve the file from the cache folder instead of downloading it again.
• CFAI: 
      An abbreviation used for Creating Friendly AI, a publication of the Singularity Institute for Artificial Intelligence.  Located at http://singinst.org/CFAI/index.html
• codelet: 
      A free-standing piece of code - a piece of code that can be detached or reattached, or a piece of code that can float independently.  Hofstadter and Mitchell's Copycat uses a codelet-based architecture to implement perception; bond-forming codelets are dumped into the Workspace to form bonds.  If the computational temperature rises, bond-breaker codelets are dumped into the Workspace.
      The concept of a codelet doesn't necessarily imply the free-floating status of a "daemon", or the programmatic individualism of an "agent" (two (flawed) concepts from traditional AI).  A codelet is to a function what an object is to a data structure.  A codelet is a function regarded as a thing in its own right, with its own properties and characteristics and behaviors, with the freedom to move around or be moved around.
• computational horizon: 
      The scope of a problem.  The amount of computing power that needs to be devoted to a task.  Outside the "horizon" lie all the facts that are not relevant to the task, all the details that are too fine to be processed with the available computing power, all the consequences that are too hazy or too unimportant to predict, and so on.  Deciding where the computational horizon lies often has a major impact on the quality and speed of a cognitive task.  See predictive horizon, granularity horizon, and relevance horizon.
• computational temperature: 
      A technique used in Hofstadter and Mitchell's Copycat (an AI that solves analogy problems) to control the degree of randomness and flexibility in the system.  When the computational temperature is high, bonds break and form easily; avenues of exploration are selected on a more random basis.  When the computational temperature is low, bonds don't break easily and only the better possibilities are explored.  In Copycat, the computational temperature is linked to the goodness of the current cognitive structures.  Elegant correspondences lower the temperature; conflicts raise the computational temperature.
      Thus, for example, Copycat may begin by seeing a plausible set of bonds.  This drops the computational temperature.  More perceptions are built up, and the computational temperature keeps dropping, until a conflict is detected - for example, an extra correspondence, or lack of correspondence, that breaks a one-to-one mapping.  This raises the computational temperature; the offending perceptual structures - bonds, groups, correspondences - dissolve; and a new set of structures, hopefully more elegant, is given the opportunity to form.
      An interesting side effect - besides the fact that Copycat can successfully solve analogy problems - is that the final computational temperature provides a good measure of the elegance of Copycat's answer.  Low computational temperatures in Copycat correspond fairly well to the analogy-problem answers that human observers would regard as elegant.
      In some ways, Copycat's computational temperature may be the first true artificial emotion.
• Consensus:  Defined in Interlude: The Consensus and the Veil of Maya.
      A term for the perceptions that are shared by all of humanity but which are not identical with external reality.  The color purple, the bands in a rainbow, and social/political perceptions are extreme examples, but none of our perceptions are precisely identical with external reality.  Nobody actually lives in external reality, and we couldn't understand it if we did; too many quarks flying around.  When we walk down a hall, watch the floor and walls and ceiling moving past us, we're actually walking around inside our visual cortex.  Despite this, the Consensus usually has an extremely tight sensory, predictive, and manipulative binding to external reality, so the rules are just as strict.
      If an object or description is part of the Consensus rather than external reality, that doesn't mean the description is arbitrary.  It does mean that you're likely have trouble defining the description mathematically, or coming up with a firm philosophical grounding, or otherwise inventing a definition that works 100% of the time.  99.99% is almost always good enough.
• continuous: 
      Can be divided and subdivided indefinitely.  Contrast to discrete.
• Copycat:  Defined in 2.3: Concepts.
      An AI, written by Melanie Mitchell and conceived by Douglas R. Hofstadter, which tries to solve analogy problems in the microdomain of letter-strings.  For example:
      If 'abc' goes to 'abd', what does 'bcd' go to?
      If 'abc' goes to 'abd', what does 'pqrs' go to?
      If 'abc' goes to 'abd', what does 'vuts' go to?
      If 'abc' goes to 'bcd', what does 'pqrs' go to?
      If 'abc' goes to 'bcd', what does 'ace' go to?
      If 'abc' goes to 'bcd', what does 'lwmb' go to?
      If 'abc' goes to 'abd', what does 'xyz' go to?  (Bear in mind that Copycat has no concept for "circularity".)
      Copycat solves these problems through a perceptual architecture.  It mentally builds up a structure of bonds, groups, correspondences, concept-mappings, until it has a "rule" for the first transition that can be applied to the second transition.  If the rule can't apply, or can't apply well, then new cognitive pressures come into play, breaking down previously perceived perceptions and allowing new groups and correspondences and rules to form.
      Note that Copycat doesn't select the correct answer from a list of alternatives; it actually invents the answer, which is very impressive and very rare.  Copycat is a really fascinating AI, and you can read about it in Metamagical Themas, or read the source code (it's a good read, and available as plain text online - no decompression required).
• counterfactual: 
      A what-if scenario deliberately contrary to reality; e.g. "What if I hadn't dropped that glass of milk?"  See also subjunctive.
• crystalline:  Defined in 1.1: Seed AI.
      Loosely speaking, "crystalline" is the opposite of "rich" or "organic".  If vast loads of meaning rest on the shoulders of individual computational tokens, so that a single error can break the system, it's crystalline.  "Crystalline" systems are the opposite of "rich" or "organic" error-tolerant systems, such as biological neural networks or seed-AI mindstuff.  Error-tolerance leads to the ability to mutate; mutation leads to evolution; evolution leads to rich complexity - networks or mindstuff with lots of tentacles and connections, computational methods with multiple pathways to success.
• cytoarchitecture: 
      "Cytoarchitecture" refers to the general way neurons connect up in a given lump of neuroanatomy - many-to-many, many-to-one, and so on.
• declarative: 
      The distinction between "procedural" and "declarative" knowledge/skill/information is one of the hallowed dichotomies of traditional AI.  Although the boundary isn't as sharp as it's usually held to be - especially in seed AI - the distinction is often worth making.
      Your knowledge of how to walk is "procedural" and is stored in procedural form.  You don't say to yourself:  "Right leg, left leg, right leg, left leg."  All the knowledge about how to balance and maintain momentum isn't stored as conscious, abstract, declarative thought in your frontal lobes; it's stored as unconscious procedural thought in your cerebellum and spinal cord.  (Neurological stereotyping included for deliberate irony.)
      Inside source code, the procedural/declarative distinction is even sharper.  A piece of code that turns on the heat when the temperature reaches 72 and turns on the air conditioning when the temperature reaches 74, where 72 and 74 are hard-coded constants, has procedurally stored the "correct" temperature of 73.  The number "73" may not even appear in the program.
      A piece of code that looks up the "target temperature" (which happens to be 73) and twiddles heat or A/C to maintain that temperature has declaratively stored the number 73, but has still procedurally stored the method for maintaining a temperature.  It will be easy for the programmer - or for internally generated programs - to refer to, and modify, the "target temperature".  However, the program still doesn't necessarily know how it's maintaining the temperature.  It may not be able to predict that the heat will go on if the temperature reaches 72.  It may not even know that there's such a thing as "heat" or "air conditioning".  All that knowledge is stored in procedural form - as code which maintains a temperature.
      In general, procedural data is data that's opaque to the program, and declarative data is data that the program can focus on and reason about and modify.  Seed AIs blur the boundary by analyzing their own source code, but this doesn't change the basic programmatic truth that declarative=good and procedural=bad.
• Deep Blue: 
      The chess-playing device that finally beat the human champion, Kasparov.  A great, glorified search tree that beat Kasparov essentially through brute force, examining two billion moves per second.  Built by IBM.
• discrete: 
      Composed of a finite number of parts which cannot be further divided.  Contrast to continuous.
• e.g.: 
      Exempli gratia; Latin, "for example".
• Eurisko: 
      Eurisko was the first truly self-enhancing AI, created by Douglas B. Lenat.  Eurisko's mindstuff - in fact, most of the AI - was composed of heuristics.  Heuristics could modify heuristics, including the heuristics which modified heuristics.
      I've never been able to find a copy of Eurisko's source code, but, by grace of Jakob Mentor, I have obtained a copy of Lenat's original papers.  It turns out that Eurisko's "heuristics" were arbitrary pieces of LISP code.  Eurisko could modify heuristics because it possessed "heuristics" which acted by splicing, modifying, or composing - in short, mutating - pieces of LISP code.  Many times this would result in a new "heuristic" which caused a LISP exception, but Eurisko would simply discard the failed heuristic and continue.  In a sense, Eurisko was the first attempt at a seed AI - although it was far from truly self-swallowing, possessed no general intelligence, and was created from crystalline components.
      Engines of Creation (by K. Eric Drexler) contains some discussion of Eurisko's accomplishments.
• exa: 
      10^18.  One billion billion; a thousand "peta".  See the Hacker's Dictionary for the full list of quantifiers.
• femto: 
      10^-15.  One-quadrillionth; a thousandth of a "pico".  See the Hacker's Dictionary for the full list of quantifiers.
• GEB: 
      Gödel, Escher, Bach:  An Eternal Golden Braid by Douglas R. Hofstadter.  This book is mandatory reading for all members of the human species.
• gender-neutral pronouns: 
      Ve, vis, ver, verself.  I was forced to start using gender-neutral pronouns when referring to intelligent AIs, since to use "he" or "she" would imply cognitive hardware that such an AI would very specifically not have.
      I realize that these pronouns strike people as annoying the first time around.  I'm sorry for that, and I truly regret having to annoy my readers, but "it" is simply inadequate to refer to AIs.  Not only is "it" used as a pronoun for inanimate matter, but "it" is also a general anaphor, like "this" or "that".  "It" can refer to anything at all in a sentence, not just the AI, so complex sentences - especially ones that use "it" for other purposes - become impossible to parse syntactically.  Sometimes a sentence can be rewritten so that no pronoun is necessary, but for sentences with multiple pronoun references, this rapidly becomes either impossible, or too tangled.  I would rather use unusual words than tangled syntax.  At least "ve" gets eaiser to parse with time.
      At one point I was using "ve" to refer to a human of indefinite gender, but I have since realized that this is just as inaccurate as referring to an AI as "he" or "she".  I now keep a coin near my computer that I flip to decide whether a human is "he" or "she".  (No, I haven't fallen into the bottomless pit of political correctness.  Everyone has the right to use whatever language they like, and I can flip a coin if I want to.  Your right to use "he" implies my right to flip a coin.  Right?)
• giga: 
      10^9.  One billion; a thousand "mega".  See the Hacker's Dictionary for the full list of quantifiers.
• GISAI: 
      "General Intelligence and Seed AI".  An abbreviation for this document.  Permanent location: http://singinst.org/GISAI/index.html
• granularity horizon: 
      The fineness of the detail that needs to be modeled.  How much reductionism needs to be applied to capture all the relevant details.  The tradeoff between expenditure of computing power, and the benefits to be gained from finer modeling.
      Every time your eye moves, the amount of processing power being devoted to each part of the visual field changes dramatically.  ("The cortical magnification factor in primates is approximately inversely linear, at least for the central twenty degrees of field."  "It has been estimated that a constant-resolution version of visual cortex, were it to retain the full human visual field and maximal human visual resolution, would require roughly 10^4 as many cells as our actual cortex (and would weigh, by inference, roughly 15,000 pounds)."  MITECS, "Computational Neuroanatomy".)  And yet we can watch an object rotating, so that different parts move all over the visual cortex, and it doesn't appear to distort.
      Every time you change scale or the level of detail at the modality level of representation, the data may wind up going into essentially a different format - at least, from the perspective of someone trying to detect identity by bitwise comparision.  Even adding or subtracting pieces of the puzzle, without changing scale, can be a problem if the AI has painstakingly built up a perceptual tower that doesn't take well to tampering with the foundations.  Cognitive methods need to be able to take these kinds of random pushes and shoves.  "Error-tolerance" isn't just important because of actual errors, but because all kinds of little flaws naturally build up in a cognitive task, as the result of cognition.
      See computational horizon.
• hertz: 
      A measure of frequency, equal to 1 cycle per second.  A neuron that fires 200 times per second is operating at 200 Hz.  CPU clock speeds are currently measured in megahertz (MHz) or gigahertz (GHz).
• heuristic: 
      I use the term to refer to any piece of knowledge which provides a rule of thumb - anything from "Don't rest your hand on a hot stove" to "Try to control the center of the chessboard".
      Some other definitions:  Douglas Lenat once wrote an AI called Eurisko in which the mindstuff - in fact, practically the entire AI - was composed of "heuristics" which could modify other heuristics, including the heuristics doing the modifyng.  For example, "investigate extreme cases" was modified by a heuristic to yield "investigate cases close to extremes".  (Douglas Lenat went on to state that "Heuristics are compiled hindsight; they are judgemental rules which, if only we'd had them earlier, would have enabled us to reach our present state of achievement more rapidly.")  In classical AI,  a "heuristic" is usually a function used to prune search trees by indicating branches which are likely or unlikely to be desirable.
• holism: 
      Holism:  The attitude that the whole is greater than the sum of the parts.  To take the holistic view is to look upward, focus on the high-level properties.  See reductionism and reductholism.  See also Gödel, Escher, Bach:  An Eternal Golden Braid, particularly the dialogues "Prelude" and "Ant Fugue".
      "No one in his right mind could deny holism." -- GEB
• horizon: 
      See computational horizon.
• i.e.: 
      Id est; Latin for "that is"; usually used to mean "in other words".
• iff: 
      "Iff" is shorthand for "if-and-only-if".
• instantiate: 
      Loosely, program A "instantiates" program B if it can perfectly simulate program B.  An "instantiation" of a program is a running copy of that program.
      This issue actually gets waaay more complicated, but I'm not going to inflict that on you now.  Maybe later.  Nobody has ever come up with a mathematical definition of "instantiation", but it's a useful concept.
• intelligence:  Defined in 2.1: World-model.
      What is intelligence?  In the case of humans, intelligence is a brain with around 40 billion neurons, and 104 cytoarchitecturally distinct areas in the cerebral cortex alone.  What intelligence is is the subject of this whole web page.
      The cause of intelligence can be more succinctly described:  Evolution is the cause of intelligence, and intelligence is an evolutionary advantage because it enables us to model, predict, and manipulate reality.  Or rather, it enables us to model, predict, and manipulate regularities in reality.
• intentionality:  Defined in 3.1: Time and Linearity.
      Behaving in such a way as to give rise to the appearance of deliberate, goal-oriented behavior.  The classic example is the corrective action of a thermostat; by switching on the heat when the temperature drops below a certain point, or switching on the air-conditioning when the temperature rises above a certain point, a thermostat gives the appearance of "wanting" the temperature to stay in a certain range.  We can take the design stance and talk about cause and effect in sensors and circuits, or take the physical stance and talk about the underlying atoms, but it's usually most convenient to take the intentional stance and say that the thermostat maintains a certain temperature.
• Kasparov: 
      Gary Kasparov, the human who finally lost the world chess championship to Deep Blue.
• kilo: 
      10^3.  One thousand.  See the Hacker's Dictionary for the full list of quantifiers.
• Lakoff and Johnson: 
      George Lakoff and Mark Johnson, coauthors of Metaphors We Live By and Philosophy in the Flesh.  George Lakoff is also the author of Women, Fire, and Dangerous Things, a book about cognitive categories.
• latency: 
      Latency describes delays, specifically delays introduced by communication rather than local processing, and irreducible delays rather than delays caused by sending a large amount of data.  Most commonly used in discussion of computer networks and hardware systems.  The latency on a motherboard is, for example, the time it takes a message from the CPU to reach a video card.  The latency between nodes of a network is the time it takes for a message from node A to reach node B.  For a fine explanation of why "latency" is entirely distinct from "bandwidth" - adding an identical second channel doubles bandwidth but does not affect latency at all - see "It's the Latency, Stupid".
      Note that our Universe specifies that the minimum latency between two nodes will be at least one second for every 186,000 miles.  The latency between two nodes 300 microns apart must be at least one picosecond.
      Anders Sandberg, in "The Physics of Information Processing Superobjects: Daily Life Among the Jupiter Brains", suggests the measure S of the "diameter" of a single mind, where S is the ratio of the clock speed to the latency between computing elements.  (S = distance / (clock speed * message speed)).  Anders goes on to note that the human brain has S ~ 1 (1).  An oft-voiced conjecture is that the subjective "feel" of having a single, unified mind may require S <= 1.  (As far as I know, this conjecture is only applied to superintelligences, and nobody has suggested that S played a role in shaping human neurology.)
• Law of Pragmatism:  Defined in 1.2: Thinking About AI.
      Any form of cognition which can be mathematically formalized, or which has a provably correct implementation, is too simple to contribute materially to intelligence.
• Life: 
      So, you want to know the meaning of life, eh?
      When capitalized, "Life" usually refers to Conway's Game of Life, a two-dimensional cellular automaton.  Cells are laid out in a square grid, and cells can either be alive or dead.  Each cell is affected only by the eight cells around it.  With each tick, these rules are applied to each cell:
          1:  A cell with fewer than two living partners becomes or remains dead.
          2:  A cell with two living parters maintains its current state.
          3:  A cell with three living partners becomes or remains alive.
          4:  A cell with four or more living partners becomes or remains dead.
      These rules are enough to generate almost endless variance.
      As for the age-old controversy among biologists about how to define "life", I suggest the following:  "Life is anything designed primarily by evolution, plus anything that counts as a person."  Note that this definition includes mules, excludes computer viruses, includes biological viruses, and (by special clause) includes designed minds smart enough to count as people.
• LISP tokens: 
      LISP is a programming language, the traditional language of AI.
      "LISP" stands for "List Processor".  In LISP, everything is made from lists, including the code.  For example, a piece of code that adds 2 and 2 would be (plus 2 2).  This code is a list composed of three tokens:  plus, 2, and 2.  If "num1" was a LISP token that contained the value of 2, the code could be written (plus num1 2) and would return 4.
      When I say that classical AI is built from suggestively-named LISP tokens, I mean that the classical AI contains a data structure reading ((is(food hamburger)) (is(eater human)) (can-eat (eater food))); the classical AI then deduces that a "human" can "eat" a "hamburger", and this is supposed to be actual knowledge about hamburgers and eating.  What the AI really knows is that a G0122 can H8911 a G8733.  Drew McDermott pointed this out in a famous article called "Artificial Intelligence Meets Natural Stupidity".
      (LISP does have one real property which is important to AI, however; the code and the data structures follow the same format, making LISP the single premier language for self-modifying code.  A true AI would probably read C++ as easily as LISP, since the amount of complexity needed to parse code is comparatively trivial relative to the amount of cognitive complexity needed to understand code.  Even so, using a language well-suited to self-modification may simplify the initial stages of the AI where self-improvement is mostly blind.  Since LISP is getting ancient as programming languages go, and linked lists are awkward by today's standards, I've proposed a replacement for LISP called "Flare", which (among many other improvements) would use XML instead of linked lists.  Even so, of course, putting faith in the token level of Flare would be no better than putting faith in the token level of LISP.  At most, Flare might be well-suited to programming sensory modalities and mindstuff.  It would be nice to have such a language, since none of the existing languages are really suited to AI, but it's more likely that we'll just hack something up out of Python - during the initial stages, at least.  For more about Flare, see the obsolete document The Plan to Singularity.)
• Marr: 
      David Marr pioneered the field of computational neurology - in particular, the computational theory of vision.  It is almost impossible to convey the magnitude of Marr's contribution to AI.  Marr was the first person who did the work and wrote the code necessary to embody a piece of cognition.  He was the first person to propose a theory of AI that wasn't oversimplified to the point of caricature - the first theory to correctly identify the token level of processing.  Reading a document like General Intelligence and Seed AI, which is essentially built on top of David Marr's paradigms, it's hard to convey what the field was like before him.
      David Marr died of leukemia in 1980, at the age of thirty-five.
• mega: 
      10^6.  One million; a thousand "kilo".  See the Hacker's Dictionary for the full list of quantifiers.
• micro: 
      10^-6.  One-millionth; a thousandth of a "milli".  See the Hacker's Dictionary for the full list of quantifiers.
• microworld:  Defined in 2.1: World-model.
      A microworld is a virtual environment in which the AI can learn and grow.  Programmatically, such a world would consist of a simulated external environment, plus an AI watching the simulated external environment through a simulated camera or some other kind of simulated sensory system, and possibly altering the simulated external environment through some kind of simulated fingers, simulated cue sticks (for a billiard-ball world), or other simulated manipulators.  Henceforth I will leave out the word "simulated" - the AI is living in a very real external environment, just one that happens to be implemented on transistors instead of quarks.  (Actually, "implemented on transistors as well as quarks" might be a better way of putting it.)  The environment, to us, is "inside the computer", but it is still outside the AI.
• milli: 
      10^-3.  One-thousandth.  See the Hacker's Dictionary for the full list of quantifiers.
• mindstuff:  Defined in Executive Summary and Introduction.
      Mindstuff is the basic substrate from which the AI's permanently stored cognitive objects (and particularly the AI's concepts) are constructed.  If a cognitive architecture is a structure of pipes, then mindstuff is the liquid flowing through the pipes.
      The mindstuff of classical AI is suggestively named LISP tokens.  The mindstuff of connectionist AI is neurons (neuroids, rather) plus the neuroid learning behaviors created by the training algorithms.  Of course, the comparision is deceptive, since neither classical nor connectionist AI have a lower modality layer or a higher thought layer.
      Insofar as seed AI has a "stuff that concepts are made of", it might be described as "reductholistic multilevel descriptions subject to conscious and autonomic manipulation", with the ultimate substrate probably being interpreted source code (in the early stages) or AI-coded assembly language (in the later stages).
• MITECS: 
      The MIT Encyclopedia of the Cognitive Sciences.  Wilson and Keil, 1999.  A truly excellent book containing 471 short articles about topics in the cognitive sciences.  See also my review in the Bookshelf.
• nano: 
      10^-9.  One-billionth; a thousandth of a "micro".  See the Hacker's Dictionary for the full list of quantifiers.
• Necessary, But Not Sufficient:  Defined in 1.2: Thinking About AI.
      A design feature may be Necessary for intelligence, but that does not mean it is Sufficient.  The latest fad AI may use "the same parallel architecture found in the human brain", but it will also use the same parallel architecture found in an earthworm's brain.
• ontology: 
      The basic level of reality.  Our ontology involves quarks, spacetime, and probability amplitudes.  The ontology of a Life game consists of dead cells, live cells, and the cellular-automaton rules.  The ontology of a Turing machine is the state transition diagram, the read/write head, and an infinitely long tape with ones and zeroes written on it.
• ontotechnology: 
      Ontotechnology is what you move on to when you're bored with nanotechnology and megascale spacetime engineering.  "Ontotechnology" is the art of meddling with underlying reality, and refers to technologies that change the laws of physics, create new Universes, or - this is the best part - change the kind of things that can be real.  Quarks are real, which we understand.  If you believe that qualia are objectively real - I do, but in my personal capacity, not my SIAI capacity - then thoughts and experiences and perceptions can also apparently be made objectively real.  If we can create qualia, why not make other things real too?  Why not turn the laws of physics into material substances that can be manipulated directly?  Heck, you could even tamper with the First Cause and make the whole of Reality go out like a candle!
      I invented the idea of ontotechnology.  Did you guess?
• orthogonal: 
      A mathematical term; in geometry, it means perpendicular.  Colloquially, two variables that can change independently of each other; not necessarily mutually irrelevant, but decoupled.  See also the entry in the Hacker's Dictionary.
• past light cone: 
      The set of all events in causal contact with a given spacetime point.  The past light cone is the space of all events from which a ray of light could have reached the current event.  The future light cone is the space of all events that can be reached by a ray of light from the current event.  Any event occurring outside our past light cone can have no causal impact on this moment.
• peta: 
      10^15.  One million billion; a thousand "tera".  See the Hacker's Dictionary for the full list of quantifiers.
• Physicist's Paradigm:  Defined in 1.2: Thinking About AI.
      The idea that advances in AI will be characterized by the discovery of a single bright idea that explains everything yet fits on a T-Shirt; e.g. "Physical Symbol Systems", "expert systems", "parallelism", "neural networks".  Confusing the task of a physicist, who "explains" a skyscraper by reference to molecular dynamics, with the task of the engineer who must design that skyscraper.
• pico: 
      10^-12.  One-trillionth; a thousandth of a "nano".  See the Hacker's Dictionary for the full list of quantifiers.
• predictive horizon: 
      How far into the future the consequences of an event need to be projected.  The amount of computing power devoted to projecting the consequences of an action.  In Friendly AI, the computational horizon for disaster-checking.
      Humans seem to do very well at recognizing the need to check for global consequences by perceiving local features of an action.  It remains to be seen whether this characteristic of 10^14 x 200hz synapses can be duplicated in N 2Ghz CPUs.  See CFAI 3.2.2: Layered mistake detection.
• procedural: 
      See declarative.
• Q.E.D.: 
      Quod erat demonstrandum; Latin for "So there!"
• qualia: 
      The substance of conscious experience.  "Qualia" is the technical term that describes the redness of red, the mysterious, indescribable, apparently irreducible quality of redness that exists above and beyond a particular frequency of light.  If a JPEG viewer stores a set of red pixels, pixels with color 0xFF0000, does it see red the way we do?  No.  Even if a program simulated all the feature-extraction of the human visual modality, would it actually see red?
      I first "got" the concept of qualia on reading the sentence "You are not the person who speaks your thoughts; you are the person who hears your thoughts."  (2).
      See "Facing Up to the Problem of Consciousness" by David Chalmers for a more extensive definition.
• qualitative:  Defined in 2.1: World-model.
      Qualitative properties are selected from a finite set; for example, the binary set of {on, off}, or the eighty-eight member set of "piano keys".  A qualitative match is when two qualities have identical values.  A qualitative binding is when the two qualities are hypothesized to be bound together - if, for example, the same item of the 88-member set "piano keys" occurs in two instances.  A qualitative binding is the weakest type of binding, since it can often occur by sheer coincidence.  However, the larger the set, the less likely a coincidence.  Small integers are usually qualitative properties; large integers should be treated as quantitative.  See also SPDM, quantitative binding, and structural binding.
• quantitative:  Defined in 2.1: World-model.
      Quantitative characteristics occupy a continuous range; they are selected from a range of real (i.e., floating-point) numbers.  A quantitative match is when two quantities are identical.  A quantitative binding occurs when two or more quantitative variables are equal to sufficient precision that coincidence is effectively impossible.  Quantitative bindings can also be established by covariance or other quantitative relations.  See also SPDM, qualitative, structural.
• reductholism: 
      A word appearing in Gödel, Escher, Bach.  "Reductholism" is a synthesis of reductionism and holism; I use it to indicate the general theory of systems with multiple levels, including both the holistic disciplines of looking up and the reductionist disciplines of looking down.  See also reductionism and holism.
• reductionism: 
      Reductionism:  The attitude that the whole is the sum of the parts.  To take the reductionist view is to look downward, focus on the low-level elements and the rules governing their interactions.  See holism and reductholism.  See also Gödel, Escher, Bach:  An Eternal Golden Braid, particularly the dialogues "Prelude" and "Ant Fugue".
      "No one in his left brain could deny reductionism." -- GEB
• reflection:  Defined in 2.4: Thoughts.
      The ability of a thinking system to think about itself.  A reflective mind is one that has an image of itself; a self-model.  In Friendly AI, a reflective goal system is one that can regard its own components and content as desirable or undesirable.  In mundane programming, a "reflective" programming language is one in which code can access information about code (for example, obtaining a list of all the methods or properties of an object).
• relevance horizon: 
      The Universe goes on forever, and what we can say about it goes on even longer.  Outside the "relevance horizon" lie all the knowledge and heuristics and skills that are not relevant to the task.  See computational horizon.
      Humans and AIs probably have very different relevance horizons.
• RNUI:  Defined in Interlude: Represent, Notice, Understand, Invent.
      Represent, Notice, Understand, Invent.  First your AI has to represent something, then it has to notice it, then it has to understand it, then invent it.  You can't take these items out of sequence.
      Representing means having the static data structures to hold the information.
      Noticing means being able to see simple relations, to perceive internal coherence; to tell the difference between a representation that makes sense, and a representation composed of random numbers.
      Understanding means being able to see goal-oriented properties, and how the thing understood fits into the larger structure of the Universe - the thing's functionality, the causes of that thing's characteristics, and so on.
      Inventing means being able to start with a high-level goal - "rapid transportation" - and design a bicycle.
• salience: 
      How much attention we're paying to something.  The salient event/object is the focus of attention (or a focus of attention); the salient event/object occupies the foreground of the mind, rather than the background.
      Deciding what to pay attention to is often a significant part of the problem.
• scalar: 
      A single number, as opposed to two or more numbers.  The speed of an airplane is a scalar quantity; it can be described by a single number.  The velocity of an airplane, which includes the direction as well as the speed, is a vector - it must be described by two numbers.  (Three numbers, if it's a three-dimensional airplane.)
      There's a timeworn joke about mosquitoes and mountain climbers that is usually mentioned at this point, but forget it.
• search trees: 
      One of the most venerable tools of AI.  In a game of tic-tac-toe, you can make any of nine possible moves, then I can make any of eight possible moves, then you can make any of seven possible moves...  The computational representation of the game - in a classical AI - would look like a tree; a single node representing the start of the game, with nine branches leading to nine first-move nodes; each first-move node would have eight branches leading to a total of seventy-two possible second-move nodes, and so on.  By searching through the entire tree, a classical AI could play a perfect game of tic-tac-toe.
      It is possible, even likely, that human cognition involves the use of similar (although much messier) search trees.  Or not.
• seed AI:  Defined in 1.1: Seed AI.
      An AI designed for self-understanding, self-improvement, and recursive self-improvement.  See 1.1: Seed AI, or the introductory article "What is Seed AI?" on the Singularity Institute's website.
• sensory modality:  Defined in 2.2: Sensory modalities.
      A sensory modality, in an AI, is a module analogous to the human visual cortex, the human auditory cortex, or some other chunk of neurology underlying one of the senses.  A modality contains the data structures needed to represent the target domain; the active processing which enables the perception of higher-level features and coherence in that domain; and the interface to the concept level which enables the abstraction of, and visualization of, patterns and objects in that domain.
• SIAI: 
      An abbrevation for "Singularity Institute for Artificial Intelligence".
• space of simultaneity: 
      A cross-section of the Universe consisting of every event that is happening "right now" according to your reference frame.  In Special Relativity, everyone has a different space of simultaneity depending on how fast they're going, and there is no "correct" space of simultaneity.  Turing machines and digital computers have a single, correct space of simultaneity.  (For a Turing machine, the space of simultaneity is the state of the tape during any given tick.)
• SPDM:  Defined in 2.1: World-model.
      Sensory, Predictive, Decisive, Manipulative.
      Intelligence is an evolutionary advantage because it enables us to model, predict, and manipulate reality.  This idea can be refined into describing four levels of binding between a model and reality.
      A sensory binding is simply a surface correspondence between data structures in the model and whatever high-level properties of reality are being modeled.
      A predictive binding is one that can be used to correctly predict future sensory inputs.
      A decisive binding is one that can be used to decide between limited sets of possible actions based on the utility of the predicted results.
      A manipulative binding is one that can be used to start from a specified result and plan a sequence of actions that will bring about the desired result.
      See also qualitative, quantitative, structural.
• stochastic: 
      Describing a statistical feature of a population.  Since stochastic processes use multiple parallel elements or averaged populations of elements to carry information, they are much more tolerant of errors than more crystalline constructs such as most modern-day software.
      Evolution tends to result in systems that are tolerant of errors and mutation; not necessarily because of a selection pressure for error-tolerance, but because every evolved system is ipso facto one that mutated without dying.  Systems with redundant low-level elements - stochastic processes - are the simplest way to get error-tolerance, and generally the first method that evolution hits on.  This is a reason why evolved organisms tend to use multiply layered stochastic systems for everything, even though each extra layer of stochastic abstraction considerably decreases efficiency.
• structural:  Defined in 2.1: World-model.
      Structural characteristics are made up of multiple qualitative or quantitative components.  A structural match is when two complex patterns are identical.  A structural binding occurs when two complex patterns are identical, or bound together to such a degree that coincidence is effectively impossible - only a pattern copy of some kind could have generated the identity.  A structural binding is usually the strongest form.  See also SPDM, qualitative, quantitative.
• subjunctive: 
      A what-if scenario, e.g. "What if you were to drop that glass of milk?"  Something imagined or visualized that isn't supposed to describe present or past reality, although a sufficiently attractive what-if scenario might later be turned into a plan, and thence into reality.  See also counterfactual.
• tera: 
      10^12.  One thousand billion; a thousand "giga".  See the Hacker's Dictionary for the full list of quantifiers.
• three:  Defined in 2.3: Concepts.
      The concept of "three" is decomposed in some detail in 2.3.3: The concept of "three".
• time:  Defined in 3.1: Time and Linearity.
      Time in a digital computer is discrete and has a single space of simultaneity, so anyone who's ever played Conway's Game of Life knows everything they need to know about the True Ultimate Nature of time in the AI.  With each tick of the clock, each frame is derived from the preceeding frame by the "laws of physics" of that ontology.  (Higher-level regularities in the sequence of frames form what we call causality; more about this in Unimplemented section: Causality.)
• Turing-computability: 
      If you really haven't heard the term "Turing-computable" before, the first thing you need to do is read Douglas R. Hofstadter's Gödel, Escher, Bach:  An Eternal Golden Braid.  Drop whatever you're doing, get the book, and read it.  It's no substitute, but there's also a nice definition of "Turing machine" in the Stanford Encyclopedia of Philosophy; also, I give a sample visualization of a Turing machine in Unimplemented section: Causality.
      Any modern digital computer can, in theory, be simulated by a Turing machine.  Any modern computer can also simulate a Turing machine, at least until it runs out of memory (Turing machines, as mathematical concepts, have an infinite amount of memory).  In essence, Turing demonstrated that a very wide class of computers, including modern Pentiums and PowerPC chips, are all fundamentally equivalent - they can all simulate each other, given enough time and memory.
      There is a task known as the halting problem - in essence, to determine whether Turing machine X, acting on input Y, will halt or continue forever.  Since the actions of a Turing machine are clearly defined and unambiguous, the halting problem obviously has a true, unique, mathematically correct yes-or-no answer for any specific question.  Turing, using a diagonalization argument, demonstrated that no Turing machine can solve the general halting problem.  Since any modern digital computer can be simulated by a Turing machine, it follows that no digital computer can solve the halting problem.  The halting problem is noncomputable.  (This doesn't necessarily demonstrate a fundamental "inferiority" of computers, since there's no reason to suppose that humans can solve the halting problem.  In fact, we can't do so simply by virtue of the fact that we have limited memories.)
      A controversial question is whether our Universe is Turing-computable - that is, can the laws of physics be simulated to arbitrary accuracy by a digital computer with sufficient speed and memory?  And if not, do the uncomputabilities carry over to (a) qualia and (b) intelligence?  I don't wish to speculate here about qualia; I don't understand them and neither do you.  However, I do understand intelligence, so I'm fairly sure that even if qualia are noncomputable, that noncomputability doesn't carry over into human general intelligence.
      I happen to believe that our physical Universe is noncomputable, mostly because I don't trust the Turing formalism.  The concept of causality involved strikes me as fundamentally subjective; there's no tail-end recursion explaining why anything exists in the first place; I've never seen a good mathematical definition of "instantiation"; plus a lot of other reasons that are waay beyond the scope of this document.  If the physical Universe is noncomputable, I believe that qualia are probably noncomputable as well.  However, this belief is strictly in my personal capacity and is not defended (or attacked, or discussed) in GISAI.  I am in the extreme minority in so believing, and in an even smaller minority (possibly a minority of one) in believing that qualia are noncomputable but that this says nothing about the impossibility, or even the difficulty, of achieving real intelligence on computable hardware.  I don't believe that semantics require special causal powers, that Godel's Theorem is at all difficult to explain to a computable intelligence, that human mathematical ability is noncomputable, that humans are superior to mere computers, or any of the other memes that customarily go along with the "noncomputability" meme.  Anyway, back to GISAI.
• ve: 
      A gender-neutral pronoun.  The equivalent of "he" or "she".
• ver: 
      A gender-neutral pronoun.  The equivalent of "him" or "her".
• verself: 
      A gender-neutral pronoun.  The equivalent of "himself" or "herself".
• vis: 
      A gender-neutral pronoun.  The equivalent of "his" or "her" (or "hers").
• world-model:  Defined in 2.1: World-model.
      The AI's mental image of the world.  Includes declaratively stored memories containing knowledge about the world, and the contents of sensory modalities - whether sensed, or imagined - constituting the mental workspace.  The latter includes subjunctive or even counterfactual mental imagery, since even a subjunctive chain of reasoning has internal consistency.

• 1:  200Hz neurons, 100m/s axons, and a 0.1m diameter
• 2:  I don't remember where I first heard this, but my guess is Raymond Smullyan.

Next:	General Intelligence and Seed AI
Up:	General Intelligence and Seed AI
Prev:	Version History