General Unsupervised Adaptive Classification Algorithm for Modular Organization of Learning Evolution – Guacamole (Semeion©)
Guacamole is an unsupervised artificial system able to perform adaptive intelligent classifications more efficiently than classical supervised algorithms (learning machine). The prerequisites of such Artificial System has to be more effective than the classical supervised statistical algorithms are the following:
- It has to classify new Input Vectors with the same or with a better capability of the classical supervised algorithms;
- When it is stimulated with new Input Vectors, it has to perform as a dynamic memory, while the classical supervised algorithms work frozen in a one shot answer.
- It has to be able to generate by itself new Input Vectors for any classes.
- It has to be able to simulate the dynamic consequences of its classification performances.
Patents
Guacamole: A population of Unsupervised ANNs able to perform supervised pattern recognition. Applicant Semeion Research Center. Inventor: M Buscema. European Patent: 09425114.7 deposited 20/03/2009.
References
[1] P.M.Buscema, P.L.Sacco
Guacamole: A New Paradigm for Unsupervised Competitive Learning
Chapter 7, pp 211-230, in W.J. Tastle (ed.), Data Mining Applications Using Artificial Adaptive Systems, DOI 10.1007/978-1-4614-4223-3_1, Springer Science+Business Media New York 2013.