Programme

Monday 28 September 2026

Time	Session
19:00 -	Informal meetup

The informal meetup will be at the Maypole pub, Portugal Place, Cambridge, CB5 8AF.

Tuesday 29 September 2026

Registration will be in the open area as you enter the Centre for Mathematical Sciences. The main workshop will be in meeting room 4 (i.e., MR4) downstairs.

Time	Session
09:00 - 09:20	Registration
09:20 - 09:30	Welcome
09:30 - 10:00	Parallel performance of vector AD, Troels Henriksen
10:00 - 10:30	diffblas and difflapack: derivatives of BLAS and LAPACK, Sri Hari Krishna Narayanan
10:30 - 11:00	Well-Posed but Ill-Conditioned: Picking an AD Abstraction Level, Valentin Churavy
11:00 - 11:30	Refreshment break
11:30 - 12:00	asdex: Sparse AD in JAX, Adrian Hill
12:00 - 12:30	ARTIST, an open-source, PyTorch-based differentiable ray tracer for solar thermal power plants, Max Pargmann
12:30 - 13:00	Reverse AD of pythonOCC for gradient-based shape optimization, Mladen Banovic
13:00 - 14:00	Lunch break with posters
14:00 - 14:30	Advancements in Fast Derivatives for Beam Optics, Bernardo Abreu Figueiredo
14:30 - 15:00	(Fast) CHAD, Tom Smeding
15:00 - 15:30	Algorithmic Differential Inversion for Regularized Training of Neural Differential Equations, Gero Kauerauf
15:30 - 16:00	Refreshment break
16:00 - 16:30	Differentiable systems over framework boundaries: AD interoperability in practice, Jonathan Brodrick
16:30 - 17:00	Multithread handling in CoDiPack, Max Sagebaum
17:00 - 17:30	Extending C++ Reflection for Compile-Time Automatic Differentiation, Thomas Prideaux-Ghee, Polina Vyshnevska, Lucas Rey
17:30 - 19:00	Free time
19:00 - 22:00	Workshop dinner

Wednesday 30 September 2026

Registration will be in the open area as you enter the Centre for Mathematical Sciences. The main workshop will be in meeting room 4 (i.e., MR4) downstairs.

Time	Session
09:00 - 09:30	DJ4Earth: new opportunities in automatic differentiation for Earth system modeling, Patrick Heimbach
09:30 - 10:00	Towards a differentiable Earth system model in Julia: Progress and perspectives, Brian Groenke
10:00 - 10:30	SpeedyWeather.jl: Differentiable, GPU-accelerated and machine-learned climate modelling, Milan Klöwer
10:30 - 11:00	Refreshment break
11:00 - 11:30	Convex Relaxations of Fixed-point Iterations and their Subgradients, Neil Kichler
11:30 - 12:00	Star Coloring Graphs Induced by Stencils, Paul Hovland
12:00 - 12:30	Taylor Mode AD in Practice: Experiments with Mixed-Partial Computation, Tim Siebert
12:30 - 13:30	Lunch break with posters
13:30 - 14:00	Work in progress of forward differentiation of fixed point loops, Jean-Luc Bouchot
14:00 - 15:15	Breakout session
15:15 - 15:30	Close