AI Concepts
Inference is the process that allows humans to discover new information from existing knowledge. Its functions include deduction, induction and abduction. Deduction results in inferences that are certain. Induction results in inferences that are probable. Abduction results in inferences that are possible. These three processes are the basis of reasoning.
Abductive reasoning is used to generate possible causes for some observed conditions that can be used in the process of fault diagnosis. These possible causes can then be challenged using deduction to narrow the answers.
More about abductive reasoning at http://thinker.iwarp.com/.
How to determine the soundness of an argument (i.e., syllogism)
Form 1: [Since | because | for | given that | due to | insofar as | inasmuch as | in view of | as shown by | can be inferred from | on the ground that | blank] [all] Xs are Ys and A is an X, [thus | therefore | accordingly | it follows that | implies that | hence | consequently | so | we may infer that | we may conclude that | in conclusion] A is a Y.
Form 2: [we may infer that | we may conclude that | blank ] A is a Y, [Since | because | given that | due to | in view of | as shown by | can be inferred from | on the ground that ] [all] Xs are Ys and A is an X.
If form is correct and A = A, X = X, Y = Y, then the argument is “valid” otherwise the argument is “invalid”.
If the argument is “valid” and the premises are true, then the argument is “sound” otherwise the argument is “unsound”.
Non-standard forms should be converted to standard forms before evaluation.
If the terms are not equal, they may be resolved with synonyms.
If the truth of any premise is unknown, it may be resolved with multi-level deduction.
Of the 512 types of syllogism 19 are valid.