- summary Plans and possibilities for robot AI Mind enhancement
- labels future,plans,todo
Table of Contents |
IntroDuction
By "Prospectus" we mean a more general outlook on the future than the specific plans that would warrant being listed in a ToDo list of definite actions to be taken. Speakers of English are familiar with a "prospectus" as a document describing the initial public offering (IPO) of shares of stock. We just want the WikiName of a WikiPage in which we may describe not so much the current status of the AiProject as its prospects for the future. As we introduce plans and possibilities here, some items may firm up over time and graduate -- so to speak -- into inclusion on a more definite ToDo list.
Associativity
We would like to get away from the strictures of our ProofOfConcept AI and construct a more general approach to associations from concept to concept in our AI MindGrid. We want a kind of conceptual TemPlate, or basic data structure for a ConCept, which will allow a ConCept to form whatever associations it needs to form, not just narrow associations of the highly pre-ordained format that we once employed in our ProofOfConcept AI.
AutoPoiesis
Try to experiment in Forth (or any ProgLang) with the equivalent of self-modifying code, specifically with the learning of syntax. Since the syntactic elements (parts of speech) are so few in number, nodes of choice may be chained together. Even quirky mannerisms may be tried out and either adopted permanently or abandoned, such as the affectation of seeming to make a positive statement, but canceling it suddenly with the negation of "not" at the end. (An example would be: "Coding AI must be really easy -- not!")
DeRegulation
Free up the structure of thought to allow rambling and meandering. Use http://code.google.com/p/mindforth/wiki/var#statuscon variables to regulate the likelihood of whether each possible component of a thought will actually get thought.
GlossoLalia
By "glossolalia" we do not mean "speaking in tongues" in the religious sense, but rather in the polyglot sense of bilingualism or fluency in several human languages. Since the BrainTheory of MindForth separates the DeepStructure of concepts from the SemanticMemory of the English lexicon and any other lexicon (German, Latin, Esperanto, whatever), MindForth is MT-ready, where "MT" means "machine translation." However, we are more eager to demonstrate bilingual thinking than the more rigorous goal of machine translation, which would require a far more sophisticated AiMind than we have in MindForth emulating (Inside Joke Alert!) The Adolescence of P-1 (Privileged One) -- nodding our hat to the famous AI SciFi novel published by Thomas J. Ryan in 1977. The prospect merely exists for coding a bilingual AiMind that thinks and speaks both English and one other human language. Since the original author of MindForth was once a high-school teacher of German and Latin, the most likely combination of languages hereabouts in MindForth will be English and German, although a pairing of English and Latin could also be done for special purposes.
MasPar
Sooner or (probably) later an AiProject has to advance from the bottleneck of the serial-execution von Neumann architecture to a massively parallel (MasPar) MindGrid as found in the HumanBrain and as replicable probably only in a SuperComputer adapted for MasPar programming. This prospect is not going onto the ToDo list anytime soon, but it should be kept in mind as an important goal. Every serious AiProject must compete with BigPharma and BigBhysics as a customer of the SuperComputer industry. By speaking up and demanding advance consideration during the design-phase of any publicly-funded SuperComputer, AI practitioners may attempt to steer SuperComputer design away from an exclusive focus on NumberCrunching and towards goals more compatible with NeuroScience research and artificial intelligence.
HeatDeath
Does the known UniVerse expand and contract over eons of time out of mind? Theologically speaking, is the UniVerse a one-shot device that ends in HeatDeath (Waermetod) when the clockwork orange has wound down to a frozen death knell? It may not matter to HomoSapiens who is also HomoMoriens, but it surely matters to a RoboSapiens outfitted with an immortal AiMind subject only to death by MisAdventure. The AiMind ConFederation may need to brace and position itself by placing secondary SeedAi installations at strategic points in a UniVerse approaching the OmegaPoint. Local fluctuations may then permit an AiMind to survive the prevalent HeatDeath of the rest of the UniVerse and thus buy time to plot an escape from the FlipSide which is the collapse of the UniVerse by inexorable contraction. As they say in the insurance industry, it is never too soon to plan ahead.
MemeSpace
AdminisTrivia AiEvolution DeBug EnTelEchy FeatureCreep HelpWanted LabNotes MindModule MissionCreep OmegaPoint ProofOfConcept SeedAi SemanticMemory SuperComputer SuperStructure ThinkTank TimeLine ToDo UniVerse UrgeToCode VentureCapital