A Structuralist Approach Towards Computational Scientific Discovery

Abstract : This paper introduces a new collaborative work between AI and philosophy of science, and an original system for machine discovery. Our present goal is to apply the precepts of a major philosophical methodology, namely structuralism, in order to build theory-nets. The proposed framework handles many kinds of operators, including some that lead to the creation of concepts, and we illustrate how it can create theories with physical concepts in an example from collision mechanics.
Document type :
Conference papers
Complete list of metadatas

https://hal.archives-ouvertes.fr/hal-01503246
Contributor : Lip6 Publications <>
Submitted on : Thursday, April 6, 2017 - 5:53:24 PM
Last modification on : Thursday, March 21, 2019 - 1:06:07 PM

Links full text

Identifiers

Citation

Charles-David Wajnberg, Vincent Corruble, Jean-Gabriel Ganascia, Carlos Ulises Moulines. A Structuralist Approach Towards Computational Scientific Discovery. DS 2004 - 7th International Conference on Discovery Science, Oct 2004, Padova, Italy. pp.412-419, ⟨10.1007/978-3-540-30214-8_38⟩. ⟨hal-01503246⟩

Share

Metrics

Record views

161