arXiv paper explainer.09 Aug 2020
arXiv.org is an open-access archive for over 1 million articles as of August 2020. There’s a lot of viewpoints about it. Here’s mine, in a question / answer format.
- Q1. Is an arXiv paper a paper?
- Q2. Keeping track.
- Q3. Citing arXiv papers.
- Q4. ArXiv papers in reviews.
- Q5. ArXiv papers on C.V.s.
- Q6. Why arXiv?
- Q7. I don’t want to arXiv.
- Q8. I want to arXiv!
Q1. Is an arXiv paper a paper?
A longer answer: an arXiv paper is like any other publication in that you must evaluate it in terms of its claims and relevance to the questions you are investigating. Some arXiv papers have gone through multiple rounds of review and revision. Some arXiv papers are accepted papers that have not yet been printed. There is a lot of variety.
Q2. Is it necessary to keep up with all of the arXiv papers in general, or even on my topic?
A2. No. You may want to run literature searches when preparing a new paper, though. Tools such as arxiv-sanity and paperswithcode help with organization somewhat, like organizing arXiv searches and linking arXiv papers to code.
This quote from a Q & A with Richard Hamming (1986) gives a longer answer to this question:
Question: How much effort should go into library work?
Hamming: It depends upon the field. I will say this about it. … If you read all the time what other people have done you will think the way they thought. If you want to think new thoughts that are different, then do what a lot of creative people do - get the problem reasonably clear and then refuse to look at any answers until you’ve thought the problem through carefully how you would do it, how you could slightly change the problem to be the correct one. So yes, you need to keep up. You need to keep up more to find out what the problems are than to read to find the solutions. The reading is necessary to know what is going on and what is possible. But reading to get the solutions does not seem to be the way to do great research. So I’ll give you two answers. You read; but it is not the amount, it is the way you read that counts.
Q3. Can I cite an arXiv paper?
A3. Yes. See Q1. You handle an arXiv citation the same way you would handle a conference or journal citation. For arXiv papers, if the paper has also been published, I prefer to cite the published version. Also, if your work uses ideas or results from an arXiv paper, you need to cite the originating paper, whatever its publication venue.
Q4. I am reviewing a paper and want to recommend a paper to the authors. The paper I want to recommend is on arXiv. Can I do this?
A4. Yes. See Q1, in which you would handle an arXiv recommendation the same way you would handle a conference or journal recommendation. A caveat, ensure that you are not recommending the authors’ paper to them. I have seen mention of this practice on Twitter. Reviewer: “Your paper is not novel because it is very similar to this other paper on arXiv.” The recommended paper and the paper under review are the same paper.
For any type of recommended publication, arXiv, conference, or journal, you need to determine its public release date. If the paper is a conference paper, the public release date is when the proceedings are avaliable online, not when the paper was presented at the conference. Secondly, you need to determine the paper under review’s submitted date. These two items are needed so you aren’t the inconsiderate reviewer who recommends a paper that was released after the paper under review was submitted. To explain my whole thought process on this issue, I created a flowchart, where the recommended paper is shortened to ‘Rec’d paper’. Also key to understanding the flowchart is my review structure, where sections include ‘Positives and Negatives of the paper’ and ‘Suggestions,’ and I specify that the authors do not have to respond to anything in the ‘Suggestions’ section.
Figure 1. A flowchart for determining how to recommend a related paper in a review. A non-visual description of the flowchart follows. First, determine the paper under review’s submission date, this is \(x\). Then, determine the recommended paper’s public release date, \(y\). For all types of papers, arXiv, conference, and journal, determine whether \(y-x > 3\) months. If \(y-x > 3\) months is “no,” then mention the recommended paper in the ‘Suggestions’ section of the review. If \(y-x > 3\) months is “yes,” does the recommended paper solve a very, very similar problem as the paper under review? If “yes” to both of these questions, then the recommended paper can be mentioned in the ‘Positives and Negatives’ of the paper section. Otherwise, it should be mentioned in the ‘Suggestions’ section.
In Figure 1, the flowchart shows that one can always recommend a related paper. Timing and relevance to the work matters, and has an impact on how the authors and editors should weight your recommendation: either “you have missed something important,” or, “here’s something for the future that might be neat, for your information only.” In the flowchart, I chose three months as an appropriate buffer time frame, to account for Google Scholar updates and searches to be run. However, this buffer could be set longer than three months. I do not recommend shortening it.
Q5. Can I list arXiv papers on my C.V.?
A5. Yes. See Q1. Like arXiv, there are many opinions about C.V.s. I recommend that people with a few publications list their arXiv papers together with other publications, under a general ‘Publications’ heading. Once you have more publications, you may choose to separate them by type, ‘Journal,’ ‘Refereed conference publications,’ ‘Technical reports,’ etc. I have some C.V. templates (in latex) for a few different career stages here.
And since this is an opinion piece, I’ll mention that I would much rather see arXiv papers on a C.V. than ‘Submitted papers’ with the conference or journal name listed. Papers get rejected, a reality of which modern academics are very aware. It would be great if your paper gets accepted to a Transactions journal, but while the months (years?!) slide by, I think it is stronger to list the paper as a standalone arXiv paper under a technical reports section than as a submitted paper.
Q6. Why do people post their papers on arXiv?
A6. This post, “Hands off arXiv!” goes into some detail about the benefits to authors of posting your paper on arXiv.
For me, arXiv’s main benefit is the increase in accessibility to research work. Universities in richer nations and some companies support subscriptions to the most common databases. These subscriptions are very expensive (see also, the Elsevier wars of 2019-) and do not cover all those who may be interested in the research. ArXiv provides a way to post one’s paper so that it can be accessed internationally, with a unique ID.
Q7. I don’t want to deal with posting my paper on arXiv. Do I have to?
A7. No. This is a “you do you” situation.
If your paper is not open-access, I think it is important to post a non-paywalled version, somewhere, to allow those without subscriptions to view the paper. Popular options include personal websites and ResearchGate. Some scientific organizations have well-developed open-access repositories, such as INRIA’s HAL.
Q8. I want to submit my first arXiv paper. What are important things to know?
A8. Once you submit your paper, you cannot take it down (see the license information). Consequently, you should discuss with all your co-authors your thoughts about arXiv submission and ensure that you have all of the relevant institutional permissions to submit.
The actual submission process using latex can take some getting used to, budget some time. Also, your .tex source will be stored on arXiv, so any comments you have in that file will remain. Getting submissions ready to go has sped up recently with a new tool, arxiv-latex-cleaner, which will resize images to get them under the size limit and remove comments in the .tex file.
Once you get your paper uploaded, you may also need to find an endorser. Follow the instructions from arXiv, and don’t forget to attach a copy of your paper when requesting an endorsement.
Many thanks to early readers of this post for feedback!