##AutomaticDifferentiation
LightNews LightNews
chrisrackauckas.bsky.social
#SciML fact of the day: automatic differentiation fails to give the correct derivative on a lot of very simple functions 😱 😱 😱 . #julialang #automaticdifferentiation

youtube.com/shorts/KTguZ...
Automatic differentiation is incorrect on very simple functions??? 😱 😱 😱
YouTube video by Chris Rackauckas
youtube.com
November 17, 2025 at 2:33 PM
juliacon.org
Alan Correa and Sarah Williamsom will host a minisymposium on "Differentiable Computational Models and their Applications" at JuliaCon 2026. Find out more at pretalx.com/juliacon-202... and submit your talk at juliacon.org/2026/cfp

#julialang #automaticdifferentiation
The adoption of differentiable modeling and simulation has gained significant traction in computational science and engineering (CSE) workflows, driving advancements in optimization, sensitivity analysis, uncertainty quantification, model discovery, and more. This minisymposium brings together scientists, researchers, practitioners, students, and software developers who are either utilizing or developing differentiable computational models. We also welcome those seeking inspiration to integrate these methodologies into their own work. Our goal is to provide a platform for knowledge exchange and learning from each other's experiences—highlighting both success stories and challenges encountered along the way. We invite contributions that share experiences in developing differentiable computational models—including specific challenges faced, innovative solutions found—and applications of these models across various domains. Attendees can look forward to learning from notable contributors within the Julia community who will present innovative use cases while fostering collaboration among participants. Through networking opportunities and shared insights, we aim to inspire both seasoned users and newcomers by demonstrating how differentiable modelling can enhance model-based CSE workflows!
February 2, 2026 at 10:06 AM
jenseisert.bsky.social
Variational optimization of projected entangled-pair states on the triangular lattice.

journals.aps.org/prb/abstract...

The “meta-message” is that the use of tools such as #automaticdifferentiation, and the co-design of the ansatz class and the physical system can improve #tensornetwork methods.
January 11, 2026 at 11:58 AM
actacrystd.iucr.org
M​ichael J. O'Brien et al.: CryoJAX: a cryo-electron microscopy image-simulation library in JAX ##CryoEM##ImageSimulation##AutomaticDifferentiation@harvard@unewhaven@Cornell...##IUCrhttps://journals.iucr.org/paper?S2059798326000550
CryoJAX: a cryo-electron microscopy image-simulation library in JAX
The authors have developed cryoJAX, a cryo-EM image-simulation library for developing data-analysis techniques across cryo-EM modalities. CryoJAX is built on JAX, an emerging scientific computing framework in Python well suited for cryo-EM data analysis.
journals.iucr.org
February 4, 2026 at 1:00 AM
actacrystd.iucr.org
CryoJAX is a cryo-EM image-simulation library for building computational data-analysis applications in JAX #CryoEM #ImageSimulation #AutomaticDifferentiation doi.org/10.1107/S205...
April 22, 2026 at 2:42 PM
paperswelove.org
Lightning Talk: Sophia Gold on An Intellectual History of Automatic Differentiation (2017)

https://www.youtube.com/watch?v=4KQ8w9PyR8U

#AutomaticDifferentiation #Derivatives #ComputerScience
July 9, 2026 at 12:15 PM
jsynchrotronrad.iucr.org
Xiangyu Yin et al.: MapsTorch: automatic differentiation for X-ray fluorescence data analysis ##XRayFluorescence##AutomaticDifferentiation##MachineLearning@advancedphoton...##IUCrhttps://journals.iucr.org/paper?S160057752501032X
MapsTorch: automatic differentiation for X-ray fluorescence data analysis
MapsTorch is a new open-source package that enables automatic differentiation for X-ray fluorescence (XRF). The robust performance of MapsTorch makes it possible to automate steps that are usually manual in XRF analysis, including initial spectrum fitting and elemental intensity refinement, and therefore can support automated, high-throughput workflows for synchrotron facilities.
journals.iucr.org
December 18, 2025 at 1:00 AM
katchwreck.mastodon.social.ap.brid.gy
`Various software efforts embrace the idea that object oriented programming enables a convenient implementation of the chain rule, facilitating so-called automatic differentiation via backpropagation. Such frameworks have no mechanism for simplifying the expressions (obtained via the chain rule) […]
Original post on mastodon.social
mastodon.social
April 23, 2026 at 5:50 AM
chrisrackauckas.bsky.social
Other work that we have done has highlighted that #automaticdifferentiation calculates incorrect derivatives on chaotic systems (frankschae.github.io/post/shadowi...).
Shadowing Methods for Forward and Adjoint Sensitivity Analysis of Chaotic Systems | FS
In this post, we dig into sensitivity analysis of chaotic systems. Chaotic systems are dynamical, deterministic systems that are extremely sensitive to small changes in the initial state or the system...
frankschae.github.io
July 9, 2025 at 10:29 AM