Output details

Evolvability via the Fourier transform

<22>Valiant’s 2007 radical idea to formalise evolution as a learning process suggested that variants of that model could inform what could be reasonably evolved. This work was the first to undertake the challenge to extend the model and obtain the first positive results for real valued hypotheses. Preprints circulated since 2007 (see [10]) generated rapid impact by explicating the possibilities of exploring further those extensions (see Section 5), leading to follow-up work (http://dx.doi.org/10.1145/2090236.2090248 and [j14,j10,c15,c10,c9] in http://www.informatik.uni-trier.de/~ley/pers/hd/f/Feldman:Vitaly) greatly strengthening the results herein. This work is cited in Valiant’s 2011 ACM Turing Award Lecture (http://amturing.acm.org/vp/valiant_2612174.cfm) and his 2013 book ISBN9780465032716.