User:Thepigdog/Origins of thought

This article attempts to makes sense of thought processes, by looking for their origins.

From the data

Prediction

The basis for prediction is inductive probabilities. Inductive probabilities are built on Occam's Razor. The simplest explanation, that explains all the facts, is the most probable to be the correct explanation. "The simplest" means the shortest description of the facts using the language and knowledge that we have.

"Simplest" can only be defined in terms of the language used to describe events. This language encodes descriptions using prior probabilities. These prior probabilities are the same as the priors in Bayes law.

Inductive reasoning only detects patterns in the world. There is no assignment of meaning. There is not even cause and effect. There only the patterns, which may be used to predict the future.

Relative probability

Most probabilities are relative probabilities between a limited number of alternatives. An alternative that explains the facts with 1 bit less of data than another alternative is twice as likely to be correct.

Absolute probabilities

An absolute or true probability may only be determined by considering every possibility, at least to the point of measuring the number of possibilities. For this reason measuring absolute probabilities is very difficult.

Absolute probabilities are formed from relative probabilities, by summing the relative probabilities of all alternatives, in order to normalize them. The sum of the probabilities of all possible outcomes is one.

Intelligent Agent

An intelligent agent acts to optimize some function called the agents goal.

To optimize this goal the agent must,

Acquire data.
Use prediction to determine the likely consequences of actions.
Choose actions that increase the acquisition of data.
Choose actions that optimize the agents goal function.

Sensory perception

Vision Systems

User:Thepigdog/Vision systems

A vision system may compare two images of the same or similar objects, by identifying a mapping of points from one image to the other image. The mapping identifies points in the two images that correspond to the same location in the 3 dimensional object.

To achieve this, the mapping vector is chosen so as minimize the information content of the total system (the two images, and the mapping). Minimizing the information means choosing the mapping so that,

Pixel properties between the two pixels mapped together are similar.
Mapping vectors close together are usually similar.

If the two images are stereoscopic images, the mapping vectors gives 3 dimensional information about the shape of the object. When comparing a past image with a new image the mapping provides a basis for recognition of the object, as well as information about the relative rotation and orientation of the objects in the two images.

The vision system allows the perception of the world as a collection of objects, classified by a measure of how similar they are.

Hearing

User:Thepigdog/Hearing

Hearing starts with the isolation of sound sources from a sound landscape. This may achieved using a phased array of microphones collectively called a beam forming microphone. By combining time delayed signals, sounds from any direction may be isolated.

After conversion of an isolated signal to the frequency domain, a time series of sounds may be compared with sounds in memory by creating a mapping vector from the recorded sound to the input sound. The size of the mapping information may be used as a measure of the similarity of the sounds.

Object Model

Objects

Humans see the world in terms of objects. An object is a recognizable pattern of data that persists in time. Objects have properties. Objects have identity.

Two identical patterns are not necessarily the same object. Identity is maintained by tracking objects.

Objects as a construction of the human mind

There is no law that the data from the universe must be interpreted as objects. Objects are an interpretation of the data that humans have evolved to use.

Logic

User:Thepigdog/Origins of logic

Logic is the description of the world based on propositions that are either true or false.

Reasoning

Reasoning of the use of logic to predict consequences.

Meta Values

User:Thepigdog/Meta values

Quotation marks distinguish between English as logic to be directly interpreted, and English as data to be considered. This has been used in computer languages to represent a string of characters, without structure.

Because of the ambiguities, and previous use of quotations, the symbol "meta" is used to represent the boundary between logic and data. The meta level is a number determining the status of the representation. Zero means that logic is represented. One means data that may be considered as logic. Two means data which, in the logic at level one, may be considered as data.

The change in meta level is the first parameter to meta. So,

meta(-1(meta(1, 5)) = 5

The use of meta-levels is essential to the practical implementation of intelligence in computers, even though it is somewhat ignored in mathematics.

Value container

User:Thepigdog/Value container

A value container contains a set of values, each tagged to identify which world the value came from. A value container is a single value, that in a controlled way, manages the multiple values that may appear as the inverse to a function.

Value containers allow the set of values to be constrained, as the evaluation proceeds to give a solution set from multiple constraint conditions.