12  Research Publication

(Modified on 2026-08-11)
1,584 words

The end goal of most scientific or engineering research is to publish a paper or present the work publicly at an appropriate (scholarly or professional) venue. Not all publication publishers/venues are equally valued or respected—they are rated (informally and formally) by their “impact.” We all hope to publish in “high-impact” venues, within our field.

In many fields, the highest impact venues are journals, which generally publish long (10–50+ page) articles, and sometimes shorter “reports,” book reviews, and commentaries. In other fields, especially in engineering, the highest impact publications are proceedings articles; “Proceedings” are books published by a conference: typically, conference participants give spoken presentations at the conference and also get to publish a paper in the proceedings. Proceedings papers are usually much shorter than journal articles, often in the 4–8 page range.

Members of GTCMT labs publish in journals/proceedings like:

Your research advisor will help you become familiar with appropriate venues for your own work.

For proprietary internal research within a company, you might not be encouraged, or even allowed, to publish research based on your work. However, written reports and presentations of results within the company are still expected.

Peer review

High-impact scholarly research, whether in the humanities, sciences, or engineering disciplines, goes through a process called peer review. Work which is published without peer review will generally be regarded as low impact—at best. However, because the peer-review process can be slow, some fields (engineering advances are progressing rapidly) like to share “preprints” of work online, without peer review.

Who

Peer review is typically, but not always “blind”—meaning that the reviewer doesn’t know the identity of the authors, and vice versa. If a journal/conference tells you to make your submissions “anonymous,” make sure you don’t include anything that could give away your identity—at ISMIR, we (are supposed to) throw out your submission if it’s not anonymized.

For journals, the journal’s editor(s) will be in charge of reviewing all submissions. They will typically assign an “action editor” to each submission they receive—the action editor being somebody the editor knows who is an expert in the submission’s topic. The action editor will read the submission then search for 2–4 reviewers—again, looking for people they know who are experts in the field. Some journals ask you to provide a list of possible reviewers to contact.

The review process for conferences and proceedings can vary a lot. Conferences will assemble a “program committee” to review all submissions; program committees may be a small group who review all the submissions together, or a larger group who each review 3–5 submissions. Often conferences then have “meta-reviewers” who act a bit like action editors. Some conferences only require that you submit a short abstract, and the review process is minimal; Other conferences thoroughly peer-review all submissions, though not quite as thoroughly as journals do.

The whole peer-review process is voluntary!

The whole peer-review process is voluntary! The editors, action editors, and reviewers are academics who volunteer their time. I serve as a reviewer for ISMIR each year, and typically review 5–6 journal articles every year. It often takes me 2–4 hours to do write a good review.

Process

In a complete, journal-style peer review, the full process is something like this:

  1. Each reviewer independently reads the paper and provides:
  • A quick summary of the paper
  • A decision
  • Comments to the authors
    • These can include big picture comments, as well as minor details (like typos).
  1. The action editor (or meta reviewer) will read the reviews and then decide what to do
  • If the reviewers disagree, the action editor may ask the reviewers to consider each others’ reviews and discuss the paper to (attempt to) resolve any disagreement.
  1. The action editor will pass on the decision and all the reviewers’ comments to the authors.
Decision

The decision options are usually:

  1. Accept outright
  2. Accept pending revisions — a.k.a., minor revisions
  3. Revise and resubmit — a.k.a., major revisions
  4. Reject outright

Outright acceptance is very rare; You are much more likely to get acceptance pending revisions or resubmit. In either case, the reviewers (should) provide a detailed list of things they want changed in the paper. These may include writing details or major changes to the project itself (which may not be possible at this point!). In the case of “Accept pending revisions,” if you make the revisions they want, the paper is in. In the case of the (more common) “Revise and resubmit,” the (same) reviewers will be asked to review the submission, and the process happens again—there is often multiple rounds of revision!

If the reviewers outright reject your paper, they are saying they don’t think you should bother trying again (at this journal). They should provide you helpful feedback anyway.

I will note that the quality of reviews varies a lot. If you are ever given the chance, you should really try to do a good job and give the authors you review helpful, constructive, (and detailed) feedback. Some reviewers can be lazy, providing little feedback, or can be unfriendly—this is really lame, but it happens.

Some conferences and journals, including ISMIR, are adopting “open review,” where reviews are published after the process. This should (hopefully) encourage reviewers to do a good job, and be kind!

Revisions and Response

When asked to make revisions, you will need to revise you paper as best as you can to satisfy the reviewers. You will also write a detailed “response letter,” where you explain how you’ve revised the paper to satisfy each of the reviewers’ comments.

In many cases, the reviewers ask you to change things that are too late to change—if you already ran the experiment, its too late to change the design! In these cases, you’ll make whatever changes you can to the paper/analysis, including adding discussion of the problem the reviewer pointed out—you might say “We later realized that this was a poor design choice, but it was too late. In future research, we’d like to fix this problem, but for now we’ll have to consider our results tentative.” You’ll explain all of this in your response letter, and hope that the reviewer (and action editor) are okay with it. In other cases, you might actually disagree with the reviewer about a change. You can say so in your response letter (and explain why) and its up to the action editor to decide if you need to make the change or not.

Once you resubmit, if the action editor considers the revisions minor enough, they might simply accept the paper without sending it back to the reviewers. More likely, the revised paper will be sent back to the same reviewers. Occasionally, the reviewers might not be available, and you’ll get fresh reviewers, which is the worse case because you’re really starting from scratch with them.

Don’t forget to thank the editor and reviewers!

Timeline

The peer review process (for journals) can be painfully slow. Everybody involved is a volunteer, so reviews are often low priority. Once a submission comes in, the action editor will send out emails trying to find people to agree to review the paper. It might take a few weeks to get 3–4 reviewers to agree. They’ll then give the reviewers ~2–3 months to do the review; reviewers will then turn their reviews in late (most of the time). So, it can easily be 4–9 months before you even hear back!

If you get to do do revisions, they’ll usually give you ~3 months to do the revisions. Then the reviewers will get some time to review the revisions, etc. When a paper is finally accepted, it’ll often be a few weeks or months before the editor gets around to preparing a final “print ready” version for the publication. At this point, they’ll ask you to (quickly, within a week) review the final copy for errors.

The whole process (especially if there are multiple revisions) can easily take 9–24 months, or even longer. Some journals are more efficient than others.

For proceedings (and conference presentations), the upcoming conference provides a strict deadline, so things progress on a consistent schedule. This makes getting proceedings papers published much faster, but the quality of the peer review is not as high.

Conferences

Conference presentations do not have as much “impact” as publications (unless the conference also has a peer-reviewed proceedings). However, conferences are an essential place to stay abreast with the latest research and network with other people interested in the same things as you. They are a lot of fun and I highly recommend that you try to present your work at a conference during your MS studies—the school will help pay for it.

Members of GTCMT labs often present at conferences like:

  • IEEE International Conference on Human-Robot Interaction
  • International Conference on Acoustics, Speech, & Signal Processing
  • International Conference on Human-Agent Interaction
  • Sound and Music Computing Conference
  • The Society for Music Perception and Cognition (SMPC), national and international conferences.
  • The Society for Music Theory (SMT), national conference (very hard to get into).

Presentations vs Poster

Conferences will often have a mix of presentation and poster sessions. (There can also be other sessions, like panels and symposia.) Presentations can be anywhere from four to thirty minutes in length, depending on the conference, and are more prestigious. Posters, where you stand by your poster amongst many other posters, are less prestigious—though some people actually like them more.

Review

Getting into a conference (with no proceedings) is generally easier than publishing work. You’ll usually only submit an abstract or a proposal (300–500 words). A small (4–7) panel of researchers will serve as the “program committee” for the conference—they will review all the proposals. The program committee is usually publicly announced—the review process may or may not be blind.

Submissions and review often takes place 6–10 months before the conference is held. Unlike publications, it is possible (and common) to submit a project to a conference before the work is complete. This allows people to present really current work. Generally, your abstract must make it very clear what parts of the project are incomplete, so the program committee can take that into account.

Mantras

We can start to piece together and summarize the epistemology of science in our first set of course mantras.

Be critical, but not an asshole

As good scientists, we want other researchers to help us by criticizing our work, but we don’t want them to hurt our feelings. As good people, we should reciprocate. Criticism matters, but can easily cause ill will and damage your reputation in the community. Remember: All research is flawed, including your own. Always give polite, constructive criticism, even when you hate an idea or its execution. And always find something nice to say! This applies to criticizing yourself, too: don’t beat yourself up.