The State of Julia for Scientific Machine Learning A brief survey. Considers why Julia isn’t more popular vs. Python & Jax. Finds calling Julia from other languages difficult.

I just skimmed a bit through, in my view it’s quite subjective: It’s quite surprising to talk about scientific machine learning in Julia but not mention ChainRules.jl, DifferentiationInterface.jl, Enzyme.jl, … In present year, Julia users have mostly moved away from improving the Julia language a…

[image] roflmaostc: In present year, Julia users have mostly moved away from improving the Julia language […] I think this is not even true, there is so many thing being improved in the core language still, like Memory types now and TTFX. Ok, I did not move away from the core Julia languag…

[image] roflmaostc: how is the Base language connected to state of machine learning? Neither Jax, PyTorch, etc. are base language Python features but instead independent libraries which appeared decades later. This is really the primary flaw of the paper. They frame all the shortcomings as a …

While this paper is a bit sloppy in terms of distinguishing the language from the ecosystem, I think it make a number of valid points. The Julia language itself is in general more useful than Python for numerical computation (let’s hope this is true because Python isn’t intended to be a numerical …

Quoting the paper: “… Julia’s language-level issues that are preventing further adoption.” What a load of crock. There are no bona fide language-level issues even mentioned in the paper as problems. Libraries? Perhaps. Language? I think the authors misundertand what a “programming language” is.

Totally agree. On the other hand, this is accepted for the NeurIPS Machine Learning and the Physical Sciences Workshop (ML4PS), Perspectives Track . It’s not a Neurips paper but still something internationally well known and the criticism here is fair and constructive. But I heard from one of the…

[image] mbauman: an undergraduate conference proceedings. Let’s not pile on too hard at them. I think this applies equally to anybody who made an effort in good faith. There’s a world of difference between analysis and trolling, and the former deserves the same respect we’d give to any collea…

Maybe, but I used to be an academic and I can’t tell from reading the text what the authors’ highest educational achievement is (nor am I sure why it would matter as long as they are able to make a cogent argument). And the discussion here has been very fair to point out the problems in this paper, …

Sure, it’s just easy to forget that there are real people behind some work, especially if they aren’t speaking up here. My post was aimed to be more preventative than scolding — just a nudge to keep the criticism constructive — but I can see it may have misfired a bit. Knowing a bit about the autho…

arXiv: "The State of Julia for Scientific Machine Learning" by Berman & Ginesin

Specific Domains Machine Learning

mbauman October 22, 2024, 7:55pm 9

Y’all, this looks like an undergraduate conference proceedings. Let’s not pile on too hard at them.

Topic		Replies	Views
State of machine learning in Julia Machine Learning	60	66894	August 26, 2022
Where does Julia provide the biggest benefits over other ML frameworks for research? Machine Learning	34	10799	September 16, 2019
Gradual Julia-ization of Python libraries General Usage package , python , pythoncall	36	3968	July 18, 2025
On Machine Learning and Programming Languages Machine Learning	48	9011	January 25, 2018
Is PyJulia org, e.g. juliacall (or pyjulia) Python packages maintained? New to Julia	36	3225	February 3, 2024

arXiv: "The State of Julia for Scientific Machine Learning" by Berman & Ginesin

Related topics