The logicist manifesto: At long last let logic-based artificial intelligence become a field unto itself☆

This paper is a persistent view that a logic-based AI that is completely out of the paradigm still included in AI's "inclusive" paradigm such as connectionionism and continuous system law should be an independent field . This document contains an independent summary of logical-based artificial intelligence and counterarguments to many objections that would inevitably be opposed to the independent declaration described here.

I discussed a more creative version of this declaration (we may begin to see how they mistaken through the dialectics of this article) and some vigorous non-logicism specific to the earlier draft of this article Thank you very much. , Logically familiar comments, how many sympathetic people, not so simple a soul mate. I am also very grateful to three anonymous reviewers for their insightful suggestions and suggestions; everything they said has been continuously reflected by myself. Furthermore, unless at the Turner Eden meeting of the CAP meeting on the computer science philosophy, these meetings link logic, technology philosophy, computer science, computational cognitive science and artificial intelligence, and this declaration will never appear. . Finally, my colleague Konstantine Arkoudas has a big debt, his logical view and the use of AI and CogSci are exciting.

These two methods are quite different. For example, Selmer Bringsjord's "The Logicist Manifesto: Finally, making logic based artificial intelligence your own domain" and the way Monica Anderson is using. The term "model-less approach" insists that "machine learning is the only artificial intelligence." In the logic world, explicit formulas and algorithms written in mathematical programming languages ​​express intelligence according to the rules of formal logic. These models are expressed in taxonomy and ontology, and clearly capture the characteristics and relationships of things that make up things, and the actions that make up the world.

This is the most exciting development in the field of artificial intelligence. But do not try to understand the complexity of the field - it may be a series of continuous and extensive articles themselves - see some major developments in deep learning history (and extension, machine learning, AI) Let's. It goes a long way in a relatively short time. Logicologist Walter Pitts and neuroscientist Warren McCulloch solved this problem when creating the first neural network mathematical model in 1943. They announced their combination of mathematics and algorithms in their innovative "logical logic of intrinsic neural activity" aimed at mimicking the human mind process.