One definition of artificial intelligence (or AI) is "the study and design of intelligent agents where an intelligent agent is a system that perceives its environment and takes actions which maximizes its chances of success.
One aspect similar to learning in humans is abstracting the classification of an unknown sample from samples of known classification. This is the essence of Statistical Pattern Recognition
In order to create artificial intelligence scientists have studied the nature of intelligence. Some of the principles learned apply to artificial as well as natural intelligence. This study includes many fields such as: physics, psychology, biology, information theory, computational models, and statistical pattern recognition. Antibodies and Histones are examples of molecules specialized to recognize patterns.
Like all algorithms, pattern recognition algorithms are composed of data and operations to manipulate that data.
In pattern recognition applications, the objects in the universe are defined by Feature Vectors; lists of attributes that describe real world objects. These quantifiable variables could include such data as height, weight, color, age, etc. These arrays or vectors can describe cars, people, or molecules.
Histones and Pattern Recognition
The Histone molecules are in some sense finite state devices with inputs from the reversibly modifiable sites on the histone tails. The situation is analagous to a processor with an n-bit bus, with n being the number of modifiable sites on the tails. This is a powerfull computational model even for a small number of interacting tetrahedral atoms.
See Smart molecules to observe the possibilities for executing “Relations” such as “containment” and “divisibility.” The cayley table supplies a possible template for organized “operations” similar abstractly to addition or multiplication.
Pattern recognition is a two-step process:
- Deriving the decision rule
- Using the decision rule
Decision Theory involves the following topics:
Probability densities (Parzen estimators)
Learning algorithms whether used by humans or molecules can be classified as: