Harry Desmond

Priors for symbolic regression

2023 · arXiv (Cornell University)

When choosing between competing symbolic models for a data set, a human will naturally prefer the "simpler" expression or the one which more closely resembles equations previously seen in a similar context. This suggests a …