Title: AI from Concrete to Abstract: Demystifying Artificial Intelligence for the General Public

arXiv:2006.04013v6 Announce Type: cross

Abstract

As Artificial Intelligence (AI) permeates a vast array of sectors, there is an urgent necessity to equip the general public with a fundamental comprehension of what AI entails. This paper introduces a novel methodology titled "AI from concrete to abstract" (AIcon2abs), which integrates visual programming with WiSARD weightless artificial neural networks. This approach is designed to help laypeople, including children, achieve this essential understanding.

The core strategy of AIcon2abs focuses on demystifying AI through hands-on activities centered on building learning machines and observing their learning processes. By engaging in these practical exercises, individuals can acquire the skills necessary to become informed participants in discussions and decisions regarding the integration of AI mechanisms.

Unlike current educational approaches that typically treat machine intelligence as an external module or component—training it separately before coupling it to the primary application created by learners—AIcon2abs takes a different path. In this new methodology, both training and classification tasks are implemented as integral blocks within the main program, functioning similarly to other standard programming constructs.

A significant side benefit of AIcon2abs is that it clearly highlights the distinction between conventional computer programs and those capable of learning from data. Furthermore, the inherent simplicity of the WiSARD weightless artificial neural network model facilitates an easy visualization and understanding of how training and classification tasks are internally realized.

Source: arXiv Generated at: 2026-06-04 00:00:00 UTC