Abstract

A widespread objection to research on scientific discovery is that there has been a noticeable dearth of significant novel findings in domain sciences contributed by machine discovery programs. The implication is that the essential parts of the discovery process are not captured by these programs. The aim of this note is to document for the AI audience a novel finding in particle physics that was enabled by the machine discovery program PAULI reported previously. This finding consists of a theorem that expresses the minimum number of conservation laws that are needed, mathematically speaking, to account for any consistent experimental data on particle reactions. This note also reports how a puzzle raised by the theorem - its conflict with physics practice - is resolved.

full paper