Ask HN: How to Publish on Your Own? | Hacker News Reader

Ask HN: How to Publish on Your Own?

Looking for guidelines to self-publish in various journals/conferences(especially related to computer science). Most of the wisdom related to processes in this space is passed on in a streamlined way to incoming grad students and researchers in an academic setting. But let's say I am a clerk sitting in a patent office and I have few experiments that I worked on with thorough proofs and results. How do I get this officially published(not about formats but about the processes)? In general, these are my questions(TIA):

1. How to choose journals/conferences to publish? 2. Do top publications prefer submissions from universities only? (What's the ground reality here?) 3. Any blogs/articles explaining in detail about this? 4. Would Arxiv let me publish without associating myself with a university?

View on HN

No university affiliation is required by arXiv or any other journal I know of. You should put some affiliation, but it can be your own thing like _Full Metal Jason Labs_.

However, it's hard to write well for an audience in a given field without having gone to grad school in that field. Readers will sense that your paper is from an outsider. So stick to topics that people will believe an outsider about. Proofs and working systems are good, criticisms of common methodologies are bad.

When you publish, especially on arXiv, you still have to get people to read it. Zero people wake up in the morning thinking "I'll just browse for new papers from people I've never heard of." You might email a few authors of recent papers you're citing saying, "Hey, I really liked your paper on ___. I've just published something related on arXiv that addresses a problem you mention in your future work section..."

Thank you for the pointers. How do PhD admissions look at self-publications? What are your thoughts on below two situations evaluated by the admissions committee at a good university(say MIT/Caltech)?

1. A strong LOR from a PI managing a known lab(but with few moderate pubs on my name) vs. 2. No great LORs but self-publishing one or two good papers(quantified with good number of citations and overall quality)

It's definitely better to have a strong LOR from a PI, even moreso from a PI at the school to which you're applying. "No great LORs but self-publishing" can be shortened to "no great LORs."

Why do you see these things as mutually exclusive? You could self-publish and use it as an "in" to working for a PI over the summer. If you do well and the PI likes you, they likely have a lot of sway with admissions and can help you get into whatever grad program they work with.

Disclaimer: Never applied/went to grad school but know many people who did... I've seen the strategy I described work at a high tier school.

> You could self-publish and use it as an "in" to working for a PI over the summer

This. Publishing is not an end onto itself. Think of it as a chance to get your foot in the door of the research community.

papers in top journals/conferences supervene everything but let's be realistic: if you're asking here you're probably not publishing in nature/annals/nips etc so unless there's something else exceptional about you (that you're publishing in mediocre journals as a high school student) LORs would carry more weight.

You are right :) If I had something that's good enough that it gets through nature then it's a different game altogether. But let's say even if you had something that was good enough to be published in Nature, what are the chances that it will get through if I directly send it in myself as a single author?(do they even accept such submissions?)

If you had something that good I'd find someone on twitter or locally who would read it and if it was good send it to someone they knew who could help get it published. And if they thought it was good they could forward it on to people. Basically building up a chain of trust via vouching.

In most schools, PhD admissions are mostly decided by the profs you'd be working for, and they are well-equipped to judge papers on their contents, not where they're published. They may also assign papers to their grad students (your future colleagues) to read and rank. In either case, they're being read by people who know the literature and hope to find people they'll be able to do good work with.

I myself don't put much faith in LoRs unless they're from someone I know pretty well. Other people differ.

If you're sure you can produce a good paper, I think that's better than a LoR. But research can be hit-or-miss.

Your self-published paper on arXiv is not going to get great (lots of) citations. Being a sole basement genius yet doing work that will be noticed by mainstream academia doesn't exist (any more). Even for grad students it's hard finding subjects and angles om them that are publishable.

One counterexample: https://arxiv.org/abs/1509.09308v1

A strong letter of recommendation with a few publications is definitely strongest.

> Zero people wake up in the morning thinking "I'll just browse for new papers from people I've never heard of."

Minor comment, but that's fairly dependent on the field. Every morning I am at work I take ~5–10 minutes to skim the titles of the day's new arXiv/astro-ph postings[0] to see if there's anything interesting. If there is, I'll mark it for later, save it, or spend some time reading/skimming it. That's fairly common among my peers in the field. There are even websites that piggyback on arXiv to facilitate voting on papers to create agendas for discussion groups or journal clubs[e.g., 1].

[0] https://arxiv.org/list/astro-ph/new

[1] https://voxcharta.org/about/about-this-website/

> No university affiliation is required by arXiv

No university affiliation is required, but you are required to 'belong to the scientific community', and to prove that you will need someone to vouch for you and your work.

It's actually a higher bar of entry for arXiv than it is conferences and journals!

>Zero people wake up in the morning thinking "I'll just browse for new papers from people I've never heard of."

Make that 1 person.

I can tell you from experience that academia is as cliquey as it gets. Depending on the field even having a good university affiliation may not see you getting through the submission and review process if you don't have one of the clique-members as a co-author. Reviewers will often find a reason to reject your submission (it's all too easy citing 'insufficient validation' or characterising the topic as 'outlandish').

I feel dirty for even saying this but the best way to get your work published in a good journal is to let a member of said clique, typically a professor with a good amount of publications on the topic, piggy-back on your article. Their presence on the author-list will lend credence to your work.

If you choose not to go the above route, and power to you for not playing their game, things get tricky. You can try applying to some well-known conferences and best-case you can get your work presented and published as proceedings (very common in the IEEE world). Depending on the topic you can try and submit to a decent journal with a strong impact-factor but depending on the editor you may not even get to the review process.

Lastly, it's mentioned elsewhere here, do not publish to rando journals that expect a fee. Even if your work is fantastic publishing in one of those will cast doubt on every single one of your words cause they're seen as completely predatory and lacking in credibility.

Makes sense. Thanks :)

World has changed quite a lot since 1905, my friend. Today, research is not done in silo and publishing new works typically requires significant understanding and knowledge about previous works and collaborators (unless you are doing pure math but even then...). If you are writing research paper in CS, you should have extensive citations in your paper. Top tier CS/ML conferences have typical acceptance rate of 25% or less. The 75% of the papers they reject often have nothing technically wrong with it except minor negligence like missing citations, not enough experiments, missing some baseline, not having confidence intervals in metrics and so on. It is being said that experienced reviewers often can tell you with high accuracy whether paper would be accepted or rejected in about 60 seconds because rejected papers usually has some signs of inexperience of author reflected in structure and presentation. Choosing right conferences is also extremely important because today’s conferences are very focused on specific topic and style. If you go bit off and you face rejection regardless of content of your paper merry because area chair and reviewers would say they are not confident. Typically you should know who has done related work in the field and which are the related papers in past 5 years at least. You should have at least one of those people as your co-author. This is needed because today’s research publications need very specific format, structure, target venue and you need someone to guide and review your work. It’s not just about correctness or achieving some great thing, you also need to compare your work with previous work using experiments and precisely layout pro and cons. Typically this is done by you being PhD student and your advisor helping you to navigate the space. Even with 2 or 3 experienced co-authors, typically writing good paper might take couple of weeks of brainstorming about presentation, content, back and forth, rewriting, rephrasing and re-reviewing. I don’t want to discourage you but if you have to ask these questions, most likely you are no where close to publishing your work. Publishing on arxive is pointless because it is considered non-peer reviewed and it would rarely get you any audience. I would suggest to look up ACM conferences in your field (IEEE is another option but I don’t like them).

No problem at all :) Thanks a lot for the info. I am actually looking for such realistic views about things to set a clear perspective on how to proceed forward. Will definitely look into that!

A clerk sitting in a patent office with a PhD or without a PhD?

To be more specific: Do you have a degree in the university? In which field? Is it related to what you want to publish?

In which field yo want to publish? In physics they like papers in journals. In computer science, they like conference presentations.

Beware that there are many predatory journals that publish any article if you pay the fee. Nobody cares about them, it's borderline a scam.

From the other comment:

> Proofs and working systems are good, criticisms of common methodologies are bad.

I agree. Be very clear about what is your new result. It is very difficult to convince people about something like "Quantum Mechanics is wrong" unless you have a big trajectory or a very solid proof. It is easier if you have a new algorithm that outperforms the 2 or 3 more popular algorithms in the 2 or 3 more popular benchmarks.

Thanks! I have Masters in CS from a state uni(Have a thesis in Machine learning). Been in industry(FAANG) for 2 years as an ML engineer(I know that's not a great track to be in for research). But I am very much interested in getting into core research teams like FAIR / Google Research as a Research Scientist(trust me when I say that the teams set the bar very clearly with an entry minimum of phd though you are exceptional as an ML engineer and they are highly attached to certain unis too. That's reality). I am not in the 99th percentile with my grades though(3.5/4.0). I understand that's also an important factor with admissions. So I am working on boosting my profile up with some good pubs. Pointers?

Grades don't actually matter that much for admissions. The letters of recommendation are much more important, e.g. from the advisor that you did your thesis with.

A friend in scientific publishing tells me that China is in on the game.

Not wishing to say anything trade confidential, however, a major Western publishing house is having difficulty competing, the reason being turnaround speed. Try and get something published here and it takes a long time. Having a baby can be quicker.

Meanwhile in China you submit to your journal and you get it peer reviewed the following week and published shortly thereafter.

Try it, at least under the pretext of learning about how the world of scientific publishing works in the West as well as in China. Even if you do not get published there will be benefits in knowing what the score is.

Sorry I have no more details to share with you on the internal difficulties of the publishing business, there has been a lot of merger and acquisition activity going on as the business model is currently going through some pains.

It may seem disloyal going to get published by the underdog, however it isn't a biggie. Where do you think growth for journal subscriptions is coming from these days? China. The publishing community is gladly selling them 'our IP'.

I'd recommend starting by looking at the literature of the area you want to publish in. Familiarise yourself with the main journals, and try to get a feel for the customs of the field.

Almost all publications provide guidelines for authors, and most editors will be happy to answer questions on the submission criteria.

Speaking as a reviewer: the most important things, above everything else, are always (1) the quality of the science, and (2) the clarity of the presentation. If you get those right, you will have little problem publishing.

If you'd like any feedback, feel free to send me an email. My address is in my profile.

Definitely :) Thanks!

This is not snark: You can self publish to your heart's content via blog.

I've actually been interviewed by a paper (and misquoted and misgendered in the article they published) due to my blogging, and I was homeless at the time. I also was interviewed by a college student who wrote a piece about me and was asked to write a paid piece with a byline because of my blogs.

I've actually been interviewed at least three other times without it being published because people find my story hard to believe. Years ago, I was quoted in a book.

I'm not even making any kind of effort to get that kind of attention. In fact, I actually spent a lot of years trying to figure out how to deflect attention while finding my voice.

I imagine if you were making a concerted effort to figure out how to get taken seriously, you could get a lot more legitimate media attention than I've ever gotten, assuming your work is solid and the writing is good.

1. I find conferences on https://edas.info/rssConferences.php and then use my country's national ranking lists to figure out whether it's a good one https://www.tsv.fi/julkaisufoorumi/haku.php?lang=en, here, a score of 1 or better is worth submitting to.

2. I don't think "top publications" like found in http://csrankings.org/ turn down submissions because it's from an independent author. Instead, it would be harder because you are working alone. Now, you need more time, nobody is steering or affecting your opinions what is currently relevant, and nobody is giving you a fresh pair of eyes to identify your own biases. That's hard.

That helps! Thanks :)

I would like to share my experience on publishing my work in a top-tier conference, without any connections with academia. I have been working in industry and is considering shifting to academic research career. But the lack of publications made my chances very low. So I took up a research project, did research on it during "after the job hours". Once I was satisfied with the results, I created an account on IEEE paperplaza, wrote the paper as per the standards set by the journal and submitted. After five months of review, the paper got accepted. You can read more of my story here http://achuwilson.github.io/blog/2016/09/how_i_wrote_researc...

Thanks for the share! So you got a phd admit?

https://arxiv.org/help/submit

It looks like registering for Arxiv involves naming your institution but it is not described as a strict requirement. Why not give it a shot? Take a direct approach. Try to mimic the tone they are used to and be respectful, acting like you belong can go a long way. I imagine that existing work would be valued over soft qualifications. If you have work that belongs there I can’t see why they would turn you away. Try the same thing with other publishers that interest you.

I guess that's one way to move forward :) Thanks! Will give that a shot then.

Reply on news.ycombinator.com