Julia is a big winner in the ARPA E dIFFERENTIATE program. ARPA-E just announced that will fund 23 Projects to Accelerate the Energy Technology Design Process. Three of the projects will use Julia to build computational platforms. Two of the projects (including ours at LANL) will use machine learning based on differentiable programming in Julia.
All the winning projects are listed here:
It’s not clear from the descriptions which of the proposals will use Julia (except for the proposal from Julia Computing, of course). Could you list the 5 projects that are using Julia in this program?
Julia Computing, Inc. – Cambridge, MA
Accelerating Coupled HVAC-Building Simulation with a Neural Component Architecture
and
Carnegie Mellon University—Pittsburgh, PA
High-fidelity Accelerated Design of High-performance Electrochemical Systems
were based on methodologies like differentiable programming, neural differential equations, and GPU-accelerated DAEs in Julia.
Los Alamos National Laboratory – Los Alamos, New Mexico
Machine Learning based Well Design to Enhance Unconventional Energy Production—
is @monty’s using differentiable programming in Julia.
Can Julia can solve my office being ridiculous cold in the summer?
No. It’ll make it more energy efficient for it to be cold in the summer, incentivizing the building manager to make it colder.
Will this include work on the general underlying infrastructure (flux, zygote,cuarrays etc)?
Indeed
Yes; we (LANL project) are using differentiable programming; it is in the project abstract.
On a related note, we also have a geothermal extraction project which also focuses on machine learning. Again we use Julia. So we cover now all the energy solutions 
I am the PI of this project as well.
This is all great news !
Nice to see that the “Differentiable Programming in Julia” infrastructure is getting good funding.
I actually got a (much smaller!) funding for a 6 months Master internship on applying Julia+AD to microgrid sizing (~about the class of problem that the NREL spinoff program HOMER Pro is addressing).
If a Master student wants to visit France this Spring, please contact me!
Yes, this is already implemented. Just use Juila to run parallel calculations on a reasonably powerful server (preferably using all cores) for a couple of days.
If, for some reason, your calculations are too fast (which, unfortunately, is a drawback Julia users frequently encounter), you can always ask the community for extra workload — just provide login details.
We are here to help.
Here’s the CMU press release:
https://www.cmu.edu/energy/news-multimedia/2019/arpa-e-award.html
Who thought of this acronym? Design Intelligence Fostering Formidable Energy Reduction (and) Enabling Novel Totally Impactful Advanced Technology Enhancements (DIFFERENTIATE). Totally impactful 
I suppose it is a backronym.
That said, there should be a progressive tax on acronyms longer than 5 letters.