Publish your first real paper
EDA is an open-access, peer-reviewed venue for high school, undergraduate and professional researchers. Free to publish, free to read, with a DOI on every accepted paper, and every decision made by a person.
Opening soon
Papers published
0
Subjects covered
$0
Cost to authors
0+
Open from age
Everything a journal does, without the gatekeeping.
Read by people with the subject knowledge.
Every submission is screened, then sent to reviewers matched on the subject they actually work in. AI assists — it never decides.
- Matched to declared subject expertise
- AI reports are advisory and labelled
- Every decision signed by a person
Five steps from draft to DOI.
The same process a professional venue uses, with the waiting and the paywalls taken out. You can see exactly where your paper is at every moment.
- 01
You submit
Upload a PDF or LaTeX source, up to 25 MB. Add your co-authors, subject area and an optional ORCID iD. Choose whether to publish under your name or a pseudonym.
Author
~15 min
- 02
Automated checks
Format validation, malware scan, similarity checking against published literature, AI-text signals and image forensics. Duplicate submissions are caught by file hash.
Automated
~10 min
- 03
AI review panel
Five specialist agents read the paper and write structured reports on methodology, novelty against prior art, research integrity, ethics and clarity. Every report is labelled advisory.
AI, advisory only
~1 hour
- 04
Human editors and reviewers
A subject editor triages, then assigns at least two human reviewers with matching expertise. Reviewers see the AI reports in a sidebar and are free to disagree with all of them.
Editor + 2 reviewers
~4 weeks
- 05
Decision and publication
The editor writes a signed decision letter. Accepted papers are copyedited, given a permanent identifier, published to a fast CDN and pushed to Google Scholar and OpenAlex.
Editor
~1 week
AI assists.
Humans decide.
Five AI agents read every submission and write structured reports on methodology, novelty, integrity, ethics and clarity. They are fast, they are thorough, and they catch things a tired reviewer misses at midnight.
They also have no authority whatsoever. An AI report is an input to a human editor, clearly labelled as advisory, and nothing else. A venue where a model decides what counts as knowledge would not be worth publishing in — so we did not build one.
Read the full AI use policyAI never decides
An AI report cannot accept or reject a paper. Only a named human editor can, and their decision letter is attached to the record permanently. If an editor disagrees with every AI report, nothing stops them.
Every AI report is disclosed
You see exactly which model read your paper, which prompt version it ran, what it said and what it cost. Reviewers see the same. Nothing about the AI layer is hidden from authors.
Your paper is never training data
Submissions are not used to train models, ours or anyone else's. Paper text is treated as untrusted input and is sandboxed away from tool access.
A permanent, citable record.
A citable DOI
Every accepted paper gets a registered DOI that resolves forever. Every version gets its own, plus one that always points at the latest.
Free, permanently
No article processing charges, no submission fees, no reading fees. We are pre-launch, funded by donations while we work toward nonprofit status.
You keep copyright
Published under CC BY 4.0 by default. You own your work; you grant us a licence to distribute it.
Real peer review
At least two human reviewers with subject expertise, plus an editor. You get a written decision letter, not a form rejection.
Versioning that makes sense
Submit a v2 and the old version stays citable. Corrections and retractions are published openly, never quietly deleted.
Indexed where it counts
Google Scholar, OpenAlex, an OAI-PMH endpoint and a public API. Your work is findable by the people who search for it.
Every accepted paper resolves forever.
Accepted papers get a registered DOI. Every version keeps its own identifier, plus one that always points at the newest. Drafts and in-review versions get a free ARK, so nothing in the pipeline is unaddressable.
Cite as
Mekonnen, A., & Park, J. (2026). Attention sparsity patterns in small transformers. EDA. https://doi.org/10.0000/eda.2026.0142
A note from the founder
I started EDA because good work kept disappearing.
I fell for physics the way most people fall for a song — all at once, and then permanently. What I did not expect was how much of the work around me never went anywhere. Friends spent months on real questions, wrote real papers, and then the file sat in a folder until the school account closed.
That is not a talent problem. It is an access problem. Journals were not built for a sixteen-year-old with a good result and no institution behind them, and the ones that will take your money are not the ones you want.
So EDA is the thing I wanted to exist: somewhere your paper is read by a person who knows the subject, and if it holds up it gets a DOI and stays findable for good. Free to publish, free to read. The reviewing is human — AI helps the reviewers, it never decides.
If you have something you have been sitting on, send it. That is the whole point.
Naol Demisse
Founder, EDA · Maryland
Twelve fields, one standard.
Computer Science
Algorithms, ML, systems, HCI
Mathematics
Pure, applied, statistics
Physics
Condensed matter, optics, astro
Biology
Molecular, ecology, neuroscience
Chemistry
Organic, materials, analytical
Quantum
Algorithms, error correction, hardware
Engineering
Robotics, civil, electrical
Earth & Climate
Climate, geology, oceans
Economics
Micro, behavioural, development
Social Science
Psychology, sociology, policy
Health Sciences
Public health, epidemiology
Interdisciplinary
Work that spans fields
High schoolers and professors, same bar.
We do not run an easier track for younger authors. That would make publication here meaningless for the students who earned it. What we do instead is give everyone a real review, written in language you can act on.
You did real work. It deserves a real record.
Science fair projects and independent research die in Google Drive. If it stands up to review, it should be citable. We take submissions from age 15 and up, with guardian consent where the law requires it.
Publish before grad school, not because of it.
Course projects, summer research and replication studies are all in scope. A published, reviewed paper with a DOI is a stronger application artefact than a line on a CV.
Full-rigour venue, no paywall.
We accept professional-grade work and hold it to professional standards. Same review process, same reviewers, no article processing charge. Many of our reviewers publish here too.
A venue for minors has to be built carefully.
A DOI is permanent. That is the feature, and it is also the risk: publishing a 15-year-old’s name attached to their work is a decision that cannot be undone later. So we built the consent and attribution machinery first, before we built anything else.
Read all policiesMinimum age 15
Verified guardian consent where the law requires it, recorded with a timestamp.
Publish under a pseudonym
Offered at submission, not buried. Your choice is permanent either way, so we make it prominent.
Similarity and integrity checks
Plagiarism, AI-text signals, image forensics and duplicate detection on every version.
Corrections in the open
Retractions get a public notice explaining why. We never quietly delete a record.
Reviewer anonymity
Reviewers stay anonymous to authors unless they choose to sign. Conflicts must be declared.
A real appeals process
Disagree with a decision and a different editor reviews it. Written response, every time.
Nobody pays to publish. Somebody pays for the servers.
Charging authors would defeat the point, and charging readers would too. So EDA runs on donations and grants, with every expense published. Reviewers and editors volunteer. A donation covers real things: DOI registration fees, cloud storage, and the review pipeline that reads every submission.
What it costs us
- DOI registration$0.25per accepted paper
- Storage and delivery~$0.02per paper per year
- AI review panel~$0.90per submission, capped
- Charged to authors$0.00and it stays that way
Is this a real journal or a preprint server?
Both, in sequence. Nothing is published until it clears human peer review, so it is not an open preprint archive. But we are not a traditional journal either: there is no subscription, no APC, and reviews are published alongside accepted papers when authors opt in.
Does AI write the reviews?
AI writes advisory reports. Humans write the reviews that count and make every decision. The AI reports exist to catch things fast, give authors more feedback than two reviewers could, and flag integrity problems early. They are clearly labelled and never binding.
How is a student paper held to a real standard?
The same way any paper is: reviewers with subject expertise, a rubric, and an editor who can reject. We reject work that does not meet the bar. A permissive venue would be worthless to the students who actually did good work.
Who pays for this?
Donations. EDA is pre-launch and runs as a sole proprietorship while we work toward nonprofit status, so donations are not tax-deductible. Our cloud costs, DOI registration fees and running expenses are published on the transparency page. Reviewers and editors are volunteers.
What if I am under 18?
You can submit from age 15. If you are under 16 and in a jurisdiction that requires it, we collect verified guardian consent before your paper enters review. You can also publish under a pseudonym, because a DOI is permanent and you may not want your legal name attached to work you did at 15.
Can I withdraw a published paper?
You can request a retraction, and we will publish a retraction notice explaining why. We cannot make the DOI stop resolving, because permanence is the whole point of a DOI. The record stays, marked clearly as retracted.
Your research is ready.
Is it citable yet?
Submitting takes about fifteen minutes. Screening comes back the same day. If it is not ready, you will get told exactly why, and you can revise and resubmit.