3 ms·
"working on compilers for imperative differentiable programming languages" what would be an example of such language?
by billconan 9y ago
"working on compilers for imperative differentiable programming languages"
what would be an example of such language?
- currymj 9y agoworking with dynamic neural net libraries that have autodiff (PyTorch, regular Torch, and I think also MXNet now) feels a lot like this. you just write normal functions, but after you execute them you can ask for gradients too.
- sedachv 9y agoCommon Lisp without any changes: https://people.eecs.berkeley.edu/~fateman/papers/ADIL.pdf https://people.eecs.berkeley.edu/~fateman/papers/ADIL.pdf Fortran with some changes required: http://www.ens.utulsa.edu/~diaz/cs8243/adifor.html http://www.ens.utulsa.edu/~diaz/cs8243/adifor.html C with some changes required: http://www.ens.utulsa.edu/~diaz/cs8243/adiff.html http://www.ens.utulsa.edu/~diaz/cs8243/adiff.html
- seanmcdirmid 9y agoBrainScript was one, though it seems deprecated ATM: https://docs.microsoft.com/en-us/cognitive-toolkit/brainscript-basic-concepts https://docs.microsoft.com/en-us/cognitive-toolkit/brainscri... Compilers might not be involved, interpreters can work just as well if most of the heavy lifting is in the solving.