[[Ashley 2019 - Learning to Select Mates]] does a simple but interesting experiment to evaluate the effect of doing mate selection by looking at the genome (control network parameters). The [[Ashley 2019 - Learning to Select Mates in Evolving Non-playable Characters]] is a fuller treatment of the same idea. [[Werner 1997 - Too many love songs: Sexual selection and the evolution of communication]] - if females prefer novelty (“being surprised”), then evolved variety and rate of innovation are greatest - this suggests a mechanism for novelty search - Citations - [[Panhuis 2001 - Sexual selection and speciation]] - Good reason to believe speciation is driven by sexual selection (it makes sense), but evidence is difficult to gather. - Idea: Compare sexual selection driving diversity through speciation to a hyper parameter search - more biologically plausible of a mechanism - Is giving up the reins controlling speciation ever useful vs just enforcing speciation via external means? - Does it depend on how “smart” the preferences are? Does co-evolving them work? - Idea: Compare sexual selection with a preference for “surprise” to novelty search - a distributed model— each female has its own understanding of commonality - Seems likely to get a fashion cycle (ABABABA) if preferences are not innate -