tag:blogger.com,1999:blog-13073795.post116103032215450540..comments2023-03-10T08:31:14.101-05:00Comments on Scheming: ``Automatic Differentiation'' in OCamlWill Farrhttp://www.blogger.com/profile/11756898910041903896noreply@blogger.comBlogger3125tag:blogger.com,1999:blog-13073795.post-60545459923920613952007-09-21T13:30:00.000-04:002007-09-21T13:30:00.000-04:00I think it's great that automatic differentiation ...I think it's great that automatic differentiation is also becoming popular in programming languages like ocaml.<BR/><BR/>The first AD tools have been mostly written in Fortran and C and I believe the most efficient AD tools are still in Fortran and C.<BR/><BR/>Lateley, with operator overloading techniques possible in Fortran90 and C++ these tools have been becoming more user-friendly while Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-13073795.post-10739383869202131602007-07-19T14:03:00.000-04:002007-07-19T14:03:00.000-04:00Here is my example of the automatic differentiatio...Here is my example of the <A HREF="http://www.parallelpython.com/content/view/17/31/#AUTO_DIFF" REL="nofollow">automatic differentiation with Python</A>Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-13073795.post-1167932457853331322007-01-04T12:40:00.000-05:002007-01-04T12:40:00.000-05:00Andrew, your python code is completely different t...Andrew, your python code is completely different to this ocaml code. You're using symbolic differentiation whereas the ocaml code is using a quite different algorithm 'automatic differentiation'. With automatic differentiation you can differentiate things like entire fluid dynamic simulations. You couldn't begin to do this with symbolic differentiation. <BR/><BR/>For my own take on automatic sigfpehttps://www.blogger.com/profile/08096190433222340957noreply@blogger.com