From OpenCog
Jump to: navigation, search

Definition of "Problem"

A problem is a situation which is experienced by an agent as different from the situation which the agent ideally would like to be in. A problem is solved by a sequence of actions that reduce the difference between the initial situation and the goal

A problem can be analysed into more specific components. First of all it consists of the two situations, the present one which we will call the initial state, and the desired one, which we can call the goal state. The agent's task is to get from the initial state to the goal state by means of series of actions that change the state. The problem is solved if such a series of actions has been found, and the goal has been reached. In general, problem-solving is a component of goal-directed action or control.

(from http://pespmc1.vub.ac.be/PROBSOLV.html)

Note that, defined as such, a problem may be one that demands General Intelligence for its solution (e.g. survival in a complex environment), or it may be one that only demands Specialized Intelligence for its solution (e.g. playing chess). In the latter case problems are often addressed by the application of individual Learning Algorithms.

Problem Types

The Mind Ontology Links section of this page contains a list of some common "machine learning problem" types. Many of these problem types are very general and have subclasses that are AGI hard and other subclasses that are solvable via Narrow AI techniques.

Mind Ontology Links