priorwork.fyi helps researchers discover relevant prior work by analyzing your document - be it a paper, draft or even just a problem statement (as a PDF) - and finding semantically similar papers from top ML conferences. The more detailed the input, the better the results.
The Process: Whether you're preparing a submission, conducting a literature review, or just starting a new topic — upload your PDF and get instant suggestions.
Privacy Note (for Reviewers): If you are using this service for a paper you are reviewing,
please know that the tool runs only on my machine and does not forward the contents to any
3rd party or store the PDF contents after the query is completed. The filename of the PDF is kept for displaying the history for logged in users.
Always consult the conference rules before use.
The tool uses semantic embeddings to understand the meaning of your document, not just keywords. Your paper is encoded into a high-dimensional vector and compared against our entire index to find the most relevant matches.
This means we can find related work even when different terminology is used — catching papers you might miss with traditional keyword search.
Our index contains 94,169 papers from major ML, NLP, and Computer Vision conferences:
The index is regularly updated as new conferences publish proceedings.
I'm Stefan, a researcher who got tired of manually hunting for related work. I built priorwork.fyi to solve my own problem and decided to share it.
Contact me here:
smagureanu+priorwork@gmail.com
Or look me up on LinkedIn: Stefan Magureanu