Re:Symbolic mathematics finally yields to neural networks
by logicchains on Wednesday May 27, 2020 @12:23AM
(#23319375)
Have you considered attempting to develop a "differentiable programming language" so that it's possible to do gradient descent directly on the space of programs? There was some work on building a "differentiable lambda calculus", https://core.ac.uk/download/pdf/82396223.pdf, but I haven't heard of anyone trying to use something like that in modern program synthesis.
Re:Symbolic mathematics finally yields to neural networks
by m3at on Thursday May 28, 2020 @05:39AM
(#23335332)
Late on the thread but you might be interested in this: https://www.tensorflow.org/swift
Re:Symbolic mathematics finally yields to neural networks
by logicchains on Saturday May 30, 2020 @03:28AM
(#23359225)
That's something different: it allows doing gradient descent on the space of paramaters for programs, not on the space of programs.
Re:Symbolic mathematics finally yields to neural networks
by yazr on Tuesday May 26, 2020 @04:05PM
(#23315811)
Love to chat. Sent u an email (former HFT and kernel guy).
One topic which is feel is neglected is a good GCN (or any GNN) to operate on existing code trees. Most approaches seem to prefer seq or at most tree inputs.
Is this simply not finding yet a good network architecture, or is it a performance issue ?
Re:Symbolic mathematics finally yields to neural networks
by siv_ran7611 on Tuesday May 26, 2020 @11:46PM
(#23319179)
I dont see any email listed on your profile. Would like to have a chat to work/learn on this stuff
We are working both on synthesizing programs from scratch (see https://arxiv.org/abs/2002.09030 for example) and on understanding computer programs using machine learning (see e.g. https://arxiv.org/abs/1911.01205).
I'm always happy to correspond with people about these topics.