0
$\begingroup$

I would like to know which journal is an appropriate outlet for the results described below.

I recently came across a particular neural-network training algorithm. The algorithm is based on a result from matrix calculus. I found a simple proof of the result although existing proofs are very tedious, though not very difficult.

I have to say that the result/proof is not very mathematically important, but it might be of interest to the machine-learning community because it is used in several algorithms.

I'd like to submit (and hopefully publish) the proof. Since the proof is quite short, I'd believe that note/letter journals would be preferable. Please let me know which journals would be appropriate outlets.

Thank you.

$\endgroup$
2
  • $\begingroup$ Welcome to Data Science! What are the journals with the articles you cite? What are the journals you read? // Much of new machine learning content is disseminated in conference presentations rather than journal articles, so you might be interested in looking up some of the lower-tier but still reputable machine learning conferences (maybe even a top-tier venue like ICML or NeuroIPS if you think you have something good). $\endgroup$
    – Dave
    Oct 10, 2022 at 4:05
  • $\begingroup$ Thank you. The papers using the math formula appear in LCML and ICLR. I understand they are top conferences. I think the formula itself will not interest the ML community. That being said, people who read the LCML/LCLR papers may wonder "why does the formula hold?" For those, I'd like to write/publish the short proof. So, I still think a note/letter journal (rather than a conference) may be the right outlet. If you have any such journals in mind, please let me know, whether top, middle, or low tiers. (As is obvious from my question, I'm not familiar with the ML community.) I appreciate it. $\endgroup$
    – Pierre
    Oct 10, 2022 at 5:31

1 Answer 1

0
$\begingroup$

As mentioned in the comments, most machine learning research is disseminated via conferences, the top three being NeurIPS, ICML, ICLR (this is a subjective opinion).

But of course, journals do exist and might be more suitable for some type of publications. In my (again subjective) opinion, good options are JMLR and JAIR.

For your particular case, I'd maybe recommend another journal which I do not have experience with, but which seems to be suitable for the type of publication you have in mind: TMLR. It focusses on short articles and technical correctness, while still having high quality standards and a proper, rapid, and even double-blind peer review process.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.