Open keflavich opened 6 years ago
This paper: http://esoads.eso.org/cgi-bin/bib_query?arXiv:1606.04373
reported a 'regularized fitting' method that involves weighting the fit based on the gradients of parameters:
Link to Yoda
This looks cool! I really like the idea of not having to specify a functional form for the gradients.
This paper: http://esoads.eso.org/cgi-bin/bib_query?arXiv:1606.04373
reported a 'regularized fitting' method that involves weighting the fit based on the gradients of parameters:
Link to Yoda