#automaticdifferentiation

Hacker NewsAn illustrated guide to automatic sparse differentiation<a href="https://iclr-blogposts.github.io/2025/blog/sparse-autodiff/" rel="nofollow noopener noreferrer" translate="no" target="_blank">https://iclr-blogposts.github.io/2025/blog/sparse-autodiff/</a><a href="https://mastodon.social/tags/HackerNews" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#HackerNews</a> <a href="https://mastodon.social/tags/automaticdifferentiation" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#automaticdifferentiation</a> <a href="https://mastodon.social/tags/sparseautodiff" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#sparseautodiff</a> <a href="https://mastodon.social/tags/illustration" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#illustration</a> <a href="https://mastodon.social/tags/machinelearning" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#machinelearning</a> <a href="https://mastodon.social/tags/ICLR2025" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#ICLR2025</a>

LavX NewsUnlocking the Power of Reverse Mode Automatic Differentiation in Machine LearningAutomatic Differentiation (AD) is revolutionizing how developers compute derivatives within complex machine learning models. This article dives deep into reverse mode AD, its implementation, and its c...<a href="https://news.lavx.hu/article/unlocking-the-power-of-reverse-mode-automatic-differentiation-in-machine-learning" rel="nofollow noopener noreferrer" target="_blank">https://news.lavx.hu/article/unlocking-the-power-of-reverse-mode-automatic-differentiation-in-machine-learning</a><a href="https://mastodon.cloud/tags/news" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#news</a> <a href="https://mastodon.cloud/tags/tech" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#tech</a> <a href="https://mastodon.cloud/tags/MachineLearning" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#MachineLearning</a> <a href="https://mastodon.cloud/tags/AutomaticDifferentiation" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#AutomaticDifferentiation</a> <a href="https://mastodon.cloud/tags/ReverseModeAD" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#ReverseModeAD</a>

AndreasHave you ever thought 💡 of using JAX as 🧮 <a href="https://mathstodon.xyz/tags/automaticdifferentiation" class="mention hashtag" rel="tag">#automaticdifferentiation</a> engine in 💻 finite element simulations? Boost the performance 🏇 of computationally-expensive hyperelastic material models with <a href="https://mathstodon.xyz/tags/jit" class="mention hashtag" rel="tag">#jit</a> in 🔍 FElupe! 🚀 🚀<a href="https://github.com/adtzlr/felupe" target="_blank" rel="nofollow noopener noreferrer" translate="no">https://github.com/adtzlr/felupe</a><a href="https://mathstodon.xyz/tags/python" class="mention hashtag" rel="tag">#python</a> <a href="https://mathstodon.xyz/tags/jax" class="mention hashtag" rel="tag">#jax</a> <a href="https://mathstodon.xyz/tags/finiteelementmethod" class="mention hashtag" rel="tag">#finiteelementmethod</a> <a href="https://mathstodon.xyz/tags/scientificcomputing" class="mention hashtag" rel="tag">#scientificcomputing</a> <a href="https://mathstodon.xyz/tags/computationalmechanics" class="mention hashtag" rel="tag">#computationalmechanics</a> <a href="https://mathstodon.xyz/tags/fea" class="mention hashtag" rel="tag">#fea</a> <a href="https://mathstodon.xyz/tags/fem" class="mention hashtag" rel="tag">#fem</a> <a href="https://mathstodon.xyz/tags/hyperelasticity" class="mention hashtag" rel="tag">#hyperelasticity</a>

Dr. P. M. SecularThis paper from <a href="https://qubit-social.xyz/@jenseisert" class="u-url mention">@jenseisert</a> 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."<a href="https://scirate.com/arxiv/2308.12358" target="_blank" rel="nofollow noopener noreferrer" translate="no">https://scirate.com/arxiv/2308.12358</a><a href="https://mathstodon.xyz/tags/quantum" class="mention hashtag" rel="tag">#quantum</a> <a href="https://mathstodon.xyz/tags/TensorNetwork" class="mention hashtag" rel="tag">#TensorNetwork</a> <a href="https://mathstodon.xyz/tags/computational" class="mention hashtag" rel="tag">#computational</a> <a href="https://mathstodon.xyz/tags/physics" class="mention hashtag" rel="tag">#physics</a> <a href="https://mathstodon.xyz/tags/AutomaticDifferentiation" class="mention hashtag" rel="tag">#AutomaticDifferentiation</a> <a href="https://mathstodon.xyz/tags/iPEPS" class="mention hashtag" rel="tag">#iPEPS</a>

Virgile AndreaniI really enjoyed the talk by Manuel Drehwald at <a href="https://fosstodon.org/tags/RustSciComp23" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#RustSciComp23</a> who drew the lines of an exciting future for <a href="https://fosstodon.org/tags/AutomaticDifferentiation" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#AutomaticDifferentiation</a> in <a href="https://fosstodon.org/tags/Rust" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#Rust</a> with <a href="https://fosstodon.org/tags/LLVM" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LLVM</a> <a href="https://fosstodon.org/tags/Enzyme" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#Enzyme</a> , 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.

marco<a href="https://sigmoid.social/tags/CFP" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#CFP</a> for `Differentiable Almost Everything: Differentiable Relaxations, Algorithms, Operators, and Simulators`, a workshop at <a href="https://sigmoid.social/tags/ICML" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#ICML</a> <a href="https://differentiable.xyz/" rel="nofollow noopener noreferrer" target="_blank">https://differentiable.xyz/</a><a href="https://twitter.com/FHKPetersen/status/1657283011459821569" rel="nofollow noopener noreferrer" target="_blank">https://twitter.com/FHKPetersen/status/1657283011459821569</a>Differentiable programming is a powerful tool, so I am quite interested in this workshop (especially as a <a href="https://sigmoid.social/tags/JuliaLang" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#JuliaLang</a> user, which has fantastic <a href="https://sigmoid.social/tags/AD" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#AD</a> support).<a href="https://sigmoid.social/tags/AutomaticDifferentiation" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#AutomaticDifferentiation</a> <a href="https://sigmoid.social/tags/ML" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#ML</a>

AndreasLots of repeated gradients/hessians to evaluate? Tensortrax is waiting for you. <a href="https://github.com/adtzlr/tensortrax" target="_blank" rel="nofollow noopener noreferrer" translate="no">https://github.com/adtzlr/tensortrax</a><a href="https://mathstodon.xyz/tags/python" class="mention hashtag" rel="tag">#python</a> <a href="https://mathstodon.xyz/tags/hyperdualnumbers" class="mention hashtag" rel="tag">#hyperdualnumbers</a> <a href="https://mathstodon.xyz/tags/automaticdifferentiation" class="mention hashtag" rel="tag">#automaticdifferentiation</a> <a href="https://mathstodon.xyz/tags/forwardmode" class="mention hashtag" rel="tag">#forwardmode</a>

Marco 🌳 Zoccanot to brag or anything but my ad-delcont library has been an inspiration to this :) <a href="https://github.com/konn/ad-delcont-primop" rel="nofollow noopener noreferrer" target="_blank">https://github.com/konn/ad-delcont-primop</a> this is a line of work that uses delimited continuations to implement reverse-mode <a href="https://sigmoid.social/tags/AutomaticDifferentiation" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#AutomaticDifferentiation</a> , 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<a href="https://sigmoid.social/tags/machinelearning" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#machinelearning</a> <a href="https://sigmoid.social/tags/functionalprogramming" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#functionalprogramming</a> <a href="https://sigmoid.social/tags/haskell" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#haskell</a>

Stephen De GabrielleUnderstanding and Implementing Automatic Differentiation 2022-12-04 :: racket, math, machine-learning, projects, tutorials By: Mike Delmonaco<a href="https://quasarbright.github.io/blog/2022/12/understanding-and-implementing-automatic-differentiation.html" rel="nofollow noopener noreferrer" target="_blank">https://quasarbright.github.io/blog/2022/12/understanding-and-implementing-automatic-differentiation.html</a><a href="https://types.pl/tags/Racket" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#Racket</a> <a href="https://types.pl/tags/RacketLang" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#RacketLang</a> <a href="https://types.pl/tags/RacketLanguage" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#RacketLanguage</a> <a href="https://types.pl/tags/AI" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#AI</a> #<a href="https://types.pl/tags/MachineLearning" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#MachineLearning</a> <a href="https://types.pl/tags/AutomaticDifferentiation" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#AutomaticDifferentiation</a> <a href="https://types.pl/tags/tutorial" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#tutorial</a>

marcoCall for Talk Proposals for the Enzyme (<a href="https://sigmoid.social/tags/AutomaticDifferentiation" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#AutomaticDifferentiation</a> in <a href="https://sigmoid.social/tags/LLVM" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LLVM</a>) Conference.<a href="https://enzyme.mit.edu/conference" rel="nofollow noopener noreferrer" target="_blank">https://enzyme.mit.edu/conference</a><a href="https://sigmoid.social/tags/JuliaLang" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#JuliaLang</a> <a href="https://sigmoid.social/tags/AD" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#AD</a>

aspuruCheck out our <a href="https://mastodon.social/tags/matterlab" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#matterlab</a> <a href="https://mstdn.social/@uoft" class="u-url mention" rel="nofollow noopener noreferrer" target="_blank">@uoft</a> work on inverse design for the Hückel method using automatic differentiation. <a href="https://arxiv.org/abs/2211.16763" rel="nofollow noopener noreferrer" target="_blank">https://arxiv.org/abs/2211.16763</a>Led by Rodrigo Vargas and Kjell Jorner <a href="https://mastodon.social/tags/ChemiVerse" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#ChemiVerse</a> <a href="https://mastodon.social/tags/machinelearning" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#machinelearning</a> <a href="https://mastodon.social/tags/AutomaticDifferentiation" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#AutomaticDifferentiation</a> <a href="https://mastodon.social/tags/ai" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#ai</a> <a href="https://mastodon.social/tags/chemistry" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#chemistry</a> <a href="https://mastodon.social/tags/toronto" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#toronto</a>

Felix Köhler📺 I started a new video series on primitive rules for <a href="https://fosstodon.org/tags/automaticdifferentiation" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#automaticdifferentiation</a> :<a href="https://www.youtube.com/watch?v=PwSaD50jTv8&list=PLISXH-iEM4Jn3SEi07q8MJmDD6BaMWlJE" rel="nofollow noopener noreferrer" target="_blank">https://www.youtube.com/watch?v=PwSaD50jTv8&list=PLISXH-iEM4Jn3SEi07q8MJmDD6BaMWlJE</a> 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 :)

marcoA consistently solid <a href="https://sigmoid.social/tags/JuliaLang" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#JuliaLang</a> YouTube Channel: <a href="https://youtube.com/c/MachineLearningSimulation" rel="nofollow noopener noreferrer" target="_blank">https://youtube.com/c/MachineLearningSimulation</a>It mainly covers some advanced topics in one of the strongest areas of <a href="https://sigmoid.social/tags/Julia" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#Julia</a>, <a href="https://sigmoid.social/tags/AutomaticDifferentiation" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#AutomaticDifferentiation</a>, especially when applied to the scientific computing domain. For example: <a href="https://youtu.be/e4O6Z9o_D0k" rel="nofollow noopener noreferrer" target="_blank">https://youtu.be/e4O6Z9o_D0k</a>Most topics also have a video covering it using <a href="https://sigmoid.social/tags/Jax" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#Jax</a> or one of the specialized <a href="https://sigmoid.social/tags/PyTorch" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#PyTorch</a> or <a href="https://sigmoid.social/tags/TensorFlow" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#TensorFlow</a> extensions (e.g., TensorFlow Distributions).

claudeFound a bug in <a href="https://post.lurk.org/tags/zoomasm" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#zoomasm</a>'s 360 projection (the distance estimate scaling was wrong).Trying to do the maths by hand is too hard, so I copied my <a href="https://post.lurk.org/tags/GLSL" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#GLSL</a> <a href="https://post.lurk.org/tags/DualNumber" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#DualNumber</a> implementation (for <a href="https://post.lurk.org/tags/AutomaticDifferentiation" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#AutomaticDifferentiation</a>) from my fragm-examples repository, minus the <a href="https://post.lurk.org/tags/CPreProcessor" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#CPreProcessor</a> macro hell, plus some quaternion-to-rotation-matrix code ported from Python that I found online.Now it looks okish in the <a href="https://post.lurk.org/tags/equirectangular" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#equirectangular</a> 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...

Recent searches

Search options

Administered by:

Server stats:

#AutomaticDifferentiation