Theoretical Preduction and Computational Scientific Discovery
- Concha Martínez Vidal (dir. congr.)
- José L. Falguera López (dir. congr.)
- José M. Sagüillo (dir. congr.)
- Víctor M. Verdejo Aparicio (dir. congr.)
- Martín Pereira Fariña (dir. congr.)
Editorial: Servicio de Publicaciones e Intercambio Científico ; Universidad de Santiago de Compostela
ISBN: 978-84-9887-939-1
Año de publicación: 2012
Páginas: 497-504
Congreso: Sociedad de Lógica, Metodología y Filosofía de la Ciencia en España. Congreso (7. 2012. Santiago de Compostela)
Tipo: Aportación congreso
Resumen
Can deductive reasoning be applied to the context of discovery of mathematical physics? I term "preduction" the method of reasoning that, starting from the available theoretical background as a whole, allows for the anticipation of previously unknown results, provided that the combination of already accepted results of different theories – those taken as premises of the deductive reasoning – is compatible with dimensional analysis. This is the method by which many hypotheses, laws, and theoretical models are introduced into physics. "Computational preduction" further extends the possibilities of machine learning, which in the past forty years has informed computational systems, implementing both the rediscovery of empirical laws and the automated discovery of equations in data bases. Automated preduction would also facilitate scientific creativity in theoretical physics.