News: Check out our newer demo at Paper To HTML.
Welcome to SciA11y!
This is an experimental prototype created by Semantic Scholar. It provides access to 1.5M open access scientific documents in accessible HTML format. Our system uses machine learning techniques to extract the semantic content of scientific papers and formats it in HTML for easier reading. Because of our reliance on statistical machine learning techniques, some errors are inevitable. We will continue to improve upon our models and would love to hear your feedback in the meantime. The papers included in this demo come from a static dataset; all papers have CC (non-ND) licenses and were published in or before April 2020. More about this prototype...
You can also upload your own PDF, which we process and render in HTML for reading. You can try this functionality here.
Example papers
2016 Elvira S. Poloczanska, Michael T. Burrows, Christopher J. Brown et al.
2016 Tsung-Hsien Wen, Milica Gasic, Nikola Mrksic et al.
2015 Mozhdeh Shahbazi, Gunho Sohn, Jérôme Théau et al.
2016 Leib Litman, Jonathan Robinson, Tzvi Abberbock
2015 Alberta Pasquale, Scott Preiss, Fernanda Silva et al.
2015 Miguel Ballesteros, Chris Dyer, Noah A. Smith
2015 Yu Zhao, Shun Duan, Bingran Yu et al.
2013 Julien IE Hoffman
2010 Michael Jerrett, Sara Gale, Caitlin Kontgis
2014 Michael V. Holmes, Folkert W. Asselbergs, Tom M. Palmer et al.
Preprint
To find out more about how we created this prototype, please read our preprint. Accessible PDF available here.
Team
- Lucy Lu Wang, Allen Institute for AI
- Isabel Cachola, Johns Hopkins University
- Jonathan Bragg, Allen Institute for AI
- Evie Yu-Yen Cheng, Allen Institute for AI
- Chelsea Haupt, Allen Institute for AI
- Matt Latzke, Allen Institute for AI
- Bailey Kuehl, Allen Institute for AI
- Madeleine van Zuylen, Allen Institute for AI
- Linda Wagner, Allen Institute for AI
- Dan S. Weld, Allen Institute for AI and University of Washington
Feedback
Please address questions or feedback to Lucy Lu Wang or Jonathan Bragg.