9 Jun 2009 mentifex   » (Master)

Meandering Chain of Thought

MeanderingChain #summary A moving wave of activation wanders across the conscious mind

MileStones RoadMap UserManual


=== Synopsis === 


After the QuIckening of your AGI software and its first GenerationOfThought leading to a CognitiveChainReaction, a meandering chain of thought is not so simple to implement as you may think. A lack of RoBot EmBodiment will prevent your AI Mind from taking its cue for thought from events being perceived through a real-world sensorium. Your AGI can think only about its own memories and about input from you the human user.

== Sensory Deprivation ==

Not having a rich panoply of sensory inputs to think about, your primitive AGI will follow the pathways of SpreadingActivation. As the AGI thinks about each available concept, a chain of thought will snake its way across the MindGrid. You may program the AGI in such a way that it asks a question whenever it tries to think a thought without sufficient information available to complete the idea. By asking a question of the human user or searching the Web, your AGI will learn new information for its knowledge base.

== Machine Learning ==

For traditional AI researches in academia, it has been a Holy Grail to achieve the MachineLearning that will come easy to your AGI. All a machine has to do in order to learn is to ask questions, but ah, there's the rub. What is a machine, that a machine may ask a question, and what is a question, that a mind may answer it?

== Moving Wave ==

You as an AGI programmer, or your AGI underlings as programmers paid to play AGI catch-up with the rest of the world, will be at pains to make sure that only one dominant concept at a time is most active in the AGI mindswirl. Why? Why engineer what would probably happen anyway? Isn't one concept or another always most active simply by definition? Maybe so, but the Moving Wave Algorithm demands it not by accident, but by design. As if a baton were being passed, a sound AGI, with or without a sound RoBot body, will think of one thing at a time and will follow the meadering chain of association from each cresting concept to the next cresting concept. If the AGI engineers are not careful to have the summit of AGI thinking let go of each sparkling concept as it begins to dim and fade away, then the act of thinking will not move forward. A unitary mind must pay unitary attention to a unitary concept. And what else is a unitary mind but a conscious mind?

== ConSciousness ==

A meandering chain of thought presupposes a mechanism of emerging ConSciousness in a machine. Sometimes it is called [http://en.wikipedia.org/wiki/Artificial_consciousness artificial consciousness] as on WikiPedia but self-awareness in man or machine is simply ConSciousness, artifical or not. For a machine, the final tipping point into consciousness is perhaps thoughts of self-reference, thoughts about "I" and "me". A chain of meandering thought must eventually stumble upon the fact of the existence of the thinker, Monsieur Rodin, and so the attainment of SelfReferentialThought will be another milestone on the long march to AGI Minds.

=== OutReach === http://agi-roadmap.org/Meandering_chain_of_thought is an open, collaborative page where you may contribute your own ideas and use the associated Talk page for discussion.

=== AI For You ===

Click to run

AI For You


=== Memespace ===

AdminisTrivia AiEvolution AiHasBeenSolved AiTree BrainTheory CognitiveArchitecture ComPutationalization ConSciousness DiaSpora DisAmbiguation EmBodiment EnTelEchy ForthMindTextFile GenerationOfThought HardTakeoff HumanLevel ImMortality IndustrialEspionage InFerence JavaScript JointStewardship KbSearch KbTraversal KnowledgeBase MachineLearning MasPar MeanderingChain MetEmPsychosis MileStones MindModule MovingWave NaturalLanguageProcessing OutReach PermanentInstallation PortingOfCode ProliFeration ProsperityEngine QuIckening RecursiveSelfEnhancement ReJuvenate RoadMap RoBot RumorMill ScienceMuseums SeedAi SelfReferentialThought SemanticMemory SeTi SpreadingActivation SubConscious SuperComputer SuperIntelligence TechnologicalSingularity TelePresence TimeLine UserManual VpAi WikiPedia

labels: milestone roadmap

Latest blog entries     Older blog entries

New Advogato Features

New HTML Parser: The long-awaited libxml2 based HTML parser code is live. It needs further work but already handles most markup better than the original parser.

Keep up with the latest Advogato features by reading the Advogato status blog.

If you're a C programmer with some spare time, take a look at the mod_virgule project page and help us with one of the tasks on the ToDo list!