Work

Unsupervised Learning and Effective Complexity: Introducing JPG and Neural Sophistication

Year: 2024

Type: conference-paper

Abstract: Measuring the complexity of arbitrary data has been of interest to many scientific domains, including machine learning and particularly unsupervised learning. In this paper, we cover relevant concepts including Kolmogorov complexity, entropy and minimum description length. We argue that these measures alone are failing to distinguish noise from meaningful complexity. We push for the concept sophistication which measures the complexity of the structured part of the data, ignoring unstructured noise. This concept is reified in two manners: using image compression algorithms and using autoencoders.

Source: Proceedings - International Conference on Tools with Artificial Intelligence, TAI

DOI: 10.1109/ictai62512.2024.00040

Authors: Erick Gomez Soto, Rémi Emonet, Marc Sebban

Institutions: Centre National de la Recherche Scientifique, Institut national de recherche en sciences et technologies du numérique, Université Jean Monnet, Université Jean Monnet

Language: English

Topic: Computability, Logic, AI Algorithms

Subfield: Computational Theory and Mathematics

Field: Computer Science

Domain: Physical Sciences

Open Access status: green

Funder: Agence Nationale de la Recherche

Award: MANUTECH Surfaces Light EngineerinG Health & Society