#small-is-beautiful #automatic-differentiation library in Python for #neural-networks. “Easily extensible autograd implemented python with pytorch API. Uses numpy to do the heavy-lifting. Implementation is very similar to pytorch (graph-based reverse-mode autodiff).”
on 02026-01-04“Who Invented the Reverse Mode of #Automatic-Differentiation?” #PDF #paper by Andreas Griewank from 02012
on 02025-09-03a nice outline of #automatic-differentiation in #Rust, with working implementations for forward and reverse mode
on 02022-05-21Jeff Dean’s talk about #deep-learning in general and #TensorFlow on 2015-10-22, before its release. He explains how important #automatic-differentiation is to them and mentions Theano.
on 02016-01-11#automatic-differentiation of symbolic expressions provides #algorithms for linear-time rather than quadratic-time symbolic differentiation
on 02016-01-08“A basic demo of #automatic-differentiation” #smallisbeautiful
on 02016-01-06I think this is a Code Words article on #automatic-differentiation, but it doesn’t say and I haven’t finished reading it, yet so I’m not sure.
on 02015-11-16"TensorFlow" is a high-throughput #dataflow array computing library and IDE with built-in reverse-mode #automatic-differentiation for optimization, with Python and C++ APIs and #IPython integration. At this moment in history, the growth of computer power has made a bunch of important #DSP and statistical tasks just feasible, so we are seeing things like self-driving cars, superhuman image recognition, and so on. But it’s been very difficult to take advantage of the available computational power, because it’s in the form of GPUs and clusters. So this is designed to make it easy to do exactly these things, and to scale them with your available computing power, along with libraries of the latest tricks in neural networks, machine learning (which is pretty close to "statistics").
on 02015-11-09a #constraint solver and #optimizer for systems of #nonlinear equations using #Haskell to implement #gradient-descent, using #automatic-differentiation, compiled to #JS using Haste and drawn with #d3, equations rendered with MathJax. Some of the visualizations don’t work for me. Most of the examples are #kinematics. By the author of #Antimony!
on 02015-10-25#paper on “Picture”, a #probabilistic-programming language for #computer-vision. "Kulkarni 2015". Mentions a “Hamiltonian Monte Carlo” inverse-rendering method that “exploits the gradients of automatically differentiable renderers”, using #automatic-differentiation, I suppose. Doesn't exploit conditional independence. #toread
on 02015-08-19