This paper from @jenseisert and colleagues sounds interesting!

"The incorporation of automatic differentiation in tensor networks algorithms has ultimately enabled a new, flexible way for variational simulation of ground states and excited states. In this work, we review the state of the art of the variational iPEPS framework. We present and explain the functioning of an efficient, comprehensive and general tensor network library for the simulation of infinite two-dimensional systems using iPEPS, with support for flexible unit cells and different lattice geometries."

https://scirate.com/arxiv/2308.12358

#quantum #TensorNetwork #computational #physics #AutomaticDifferentiation #iPEPS

I really enjoyed the talk by Manuel Drehwald at #RustSciComp23 who drew the lines of an exciting future for #AutomaticDifferentiation in #Rust with #LLVM #Enzyme , which should be directly integrated into the compiler at an horizon of a couple of months.

If I understood correctly, the idea is to differentiate code at the LLVM IR level, *after optimization* (and to do another pass of optimization after that). This can produce faster code than the AD engines that operate at the source code level.

Check out the new user manual for the ice-sheet model #SICOPOLIS. A great shout-out to Shreyas Gaikwad for having set up the framework for it and writing the part pertinent to #AutomaticDifferentiation!
https://sicopolis.readthedocs.io

not to brag or anything but my ad-delcont library has been an inspiration to this :)

https://github.com/konn/ad-delcont-primop

this is a line of work that uses delimited continuations to implement reverse-mode #AutomaticDifferentiation , rather than reifing the program into a graph. As such, it enables a nice purely functional API and this latest incarnation performs pretty well too

#machinelearning #functionalprogramming #haskell

Call for Talk Proposals for the Enzyme (#AutomaticDifferentiation in #LLVM) Conference.

https://enzyme.mit.edu/conference

#JuliaLang #AD

Check out our #matterlab @uoft work on inverse design for the Hückel method using automatic differentiation.

https://arxiv.org/abs/2211.16763

Led by Rodrigo Vargas and Kjell Jorner

#ChemiVerse #machinelearning #AutomaticDifferentiation #ai #chemistry #toronto

📺 I started a new video series on primitive rules for #automaticdifferentiation :https://www.youtube.com/watch?v=PwSaD50jTv8&list=PLISXH-iEM4Jn3SEi07q8MJmDD6BaMWlJE
starting at scalar rules, continuing with vector/array rules, and finally some results from using the implicit function theorem.

Primitive rules build the basis for automatically differentiating through arbitrary computer programs.

New video to be released every three days :)

A consistently solid #JuliaLang YouTube Channel: https://youtube.com/c/MachineLearningSimulation

It mainly covers some advanced topics in one of the strongest areas of #Julia, #AutomaticDifferentiation, especially when applied to the scientific computing domain. For example: https://youtu.be/e4O6Z9o_D0k

Most topics also have a video covering it using #Jax or one of the specialized #PyTorch or #TensorFlow extensions (e.g., TensorFlow Distributions).

Found a bug in #zoomasm's 360 projection (the distance estimate scaling was wrong).

Trying to do the maths by hand is too hard, so I copied my #GLSL #DualNumber implementation (for #AutomaticDifferentiation) from my fragm-examples repository, minus the #CPreProcessor macro hell, plus some quaternion-to-rotation-matrix code ported from Python that I found online.

Now it looks okish in the #equirectangular view, need to render some tests at various orientations and inject spatial metadata for viewing in VLC to be more sure I got it right...