AI generated podcast
"Advancing Artificial Intelligence with Kant"
"To know anything in space (for instance, a line) I must draw it" Kant, I. , Critique of Pure Reason, B 137/138 |
Listen on Academia.edu to an AI generated podcast in English of my German paper of 1991
MIT KANT FORTSCHREITEN IN DER KÜNSTLICHEN INTELLIGENZ
[paper published in KANT YEARBOOK (Kantovski sbornik), 16, 75-84 (1991)]
Click here to access the Academia page with my paper and podcast
then click there on the field "Listen to podcast summary"
ABSTRACT
Through "knowledge processing," AI (artificial intelligence) intends to realize the functions of human thought (and not just its results) in computers. "Knowledge representation" is the main problem today, for which many individual detailed solutions have been found in practice, but the fundamental questions remain unanswered.
With the help of Kant's analysis of the conceptual faculty in the "Critique of Pure Reason," a new methodology—called "critical methodology"—can be developed for basic research in this field.
Previous AI models neglect the conceptual faculty, limit themselves to the concepts themselves (symbolic dimension, formal logic), and one-sidedly favor an analytical (passive) object reference (AI dogmatism).
Critical methodology emphasizes the importance—for knowledge processing—of the functions from which concepts arise, introduces the additional object dimension (two-dimensional, transcendental logic), and adds a synthetic (productive) object reference.
The goal is both an "operational knowledge representation" in which knowledge of every kind is represented as a structure of synthetic operations, and an integrated model of knowledge (symbol+object, synthetic+analytic).