iTagPDF: Towards Finally Automating PDF Accessibility
The DOI matches its Crossref record; 36 of the 36 references it cites were found in a registry or an index; 20 of the 50 sources its claims cite were checked for support.
Abstract
Most academic research is ultimately disseminated through documents in the PDF format. This format has advantages in flexibility and portability, but presents challenges for accessibility that have stubbornly resisted solutions despite decades of attempts. Tagging PDFs is hard to automate because tags are currently generated visually, not semantically, which makes the output cluttered and manual correction tedious and error-prone. Ironically, this semantic structure already exists during authoring but is discarded during PDF rendering. This raises an obvious question, can we use this lost semantic information to better automate tagging in PDFs? In this paper, we develop iTagPDF, a system that refines generated metadata using the semantics in the source documents of research papers. We demonstrate that the metadata generated by iTagPDF already surpasses what authors currently submit to ACM conferences on many criteria. Our approach represents a concrete step toward finally automating accessibility remediation in research paper PDFs.
Overview
CoChat Summary
TL;DR
iTagPDF automatically generates PDF accessibility metadata by combining visual PDF processing with semantic context from LaTeX source documents. It outperforms Adobe Acrobat auto-tagging and author-submitted PDFs on many accessibility criteria, though it still relies on author verification and has limitations for some content types.
Key findings
- On 40 ACM CHI/ASSETS paper PDFs, iTagPDF achieved 95.28% bounding box accuracy, 95.96% classification accuracy, and 97.26% reading order accuracy.
- In a comparison on 25 papers, iTagPDF scored 81.36% average accessibility criteria versus 49.68% for Adobe Acrobat auto-tags and 55.59% for author-submitted PDFs; normalized scores were 95.19%, 87.03%, and 80.97%, respectively.
- iTagPDF surpassed author-submitted metadata on reading order (96.42% vs. 88.62%), headings tagged (91.58% vs. 71.92%), lists tagged (83.09% vs. 46.46%), tables tagged (89.79% vs. 66.67%), and captions tagged (82.92% vs. 0.0%).
- Adobe Acrobat performed better on table structure in the smaller subset (81.29% vs. iTagPDF’s 72.40%, N=5), and authors’ submitted PDFs had higher figure alt-text coverage (81.60% vs. iTagPDF’s 75.20%).
Methods
- iTagPDF processes PDF page images with a YOLOv11 object detector fine-tuned on DocLayNet, extracts text with Tesseract OCR, parses LaTeX source into semantic chunks, and aligns visual regions to source chunks using word-overlap heuristics plus a GPT-4o LLM agent.
- It generates document-level tags such as `<H1:H3>`, `<P>`, `<Caption>`, `<Table>`, `<L>`, `<Figure>`, `<Artifact>`, `<Note>`, `<Formula>`, `<BibEntry>`, and `<Author>`, plus content-specific tags `<TH>`, `<TD>`, and `<LI>`, then embeds metadata using iText Core.
- Evaluation used 40 papers (20 CHI, 20 ASSETS; 554 pages; over 10,000 visual segments) from arXiv LaTeX sources and ACM Digital Library PDFs, with manually computed Box Error Rate, Classification Error Rate, and Reading Order Error Rate; accessibility criteria were compared against Acrobat auto-tagging and author-submitted PDFs on a 25-paper subset.
Limitations
- The paper states iTagPDF currently focuses on LaTeX-authored research paper PDFs and may not extend directly to Microsoft Word or Adobe InDesign sources.
- It relies on heuristics and LLM-in-the-loop mapping, which can fail when no reliable mapping is found, especially for figures or captions, leading to redundant `<Artifact>` tags; LLM outputs are not strictly deterministic, though the paper reports near-deterministic behavior for most tags.
- The paper states accuracy is limited by off-the-shelf modules, iTagPDF does not produce some tags such as `<Link>`, does not identify inline formulas, does not handle references as list elements, and still relies on author verification for figure alt text.
- Evident from the text: table structure evaluation used only N=5, formula tagging N=2, and alt-text quality was not evaluated.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
17 Valid · 9 Not valid · 16 Not checked · 83 Own results and interpretations
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Although a number of tools exist for PDF remediation [1, 7, 32], they mostly work in the same way -the tool visualizes each page of the PDF document including any existing metadata information and the user must manually provide (or fi x) the metadata.”
What the sources say
- Not checked yetcites [1] [1] Adobe Inc. 2023. Adobe Acrobat Pro DC Accessibility Features . Adobe Systems Incorporated. https://helpx.adobe.com/acrobat/using/createverify-pdf-accessibility.html. -
- Not checked yetcites [7] [7] Equidox Software Company. 2025. Equidox PDF Accessibility Software. https://equidox.co/pdf-solutions/pdf-accessibility-software/. Accessed: 2025-04-17. -
- Partly supportedcites [32] [32] Felix Maximilian Schmitt-Koopmann, Elaine May Huang, Hans-Peter Hutter, and Alireza Darvishy. 2025. Towards More Accessible Scientific PDFs for People with Visual Impairments: Step-by-Step PDF Remediation to Improve Tag Accuracy. In Proceedings of the 2025 CHI Conference on Human Factors in Computing Systems (CHI '25) . Association for Computing Machinery, New York, NY, USA, Article 48, 16 pages. doi:10.1145/3706598.3713084 - sourceExperimental — accuracy not yet measured
The Related Work section says professional PDF remediation tools have workflows that differ only slightly and let users edit the PDF structure tree directly. It does not say those tools all visualize each page with existing metadata or require users to provide or fix metadata manually. Section 3.8 describes page review and optional metadata editing in the authors’ own prototype, not in existing tools generally. (LLM, conf=0.84)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
Every source the paper cites for this claim was checked and supports it.
In the paper
“Wang et al. [39] expanded this analysis through automated evaluation of roughly 11k academic PDFs from diverse venues and noted that only 2.4% papers were compliant with Adobe-5 standards.”
What the sources say
- Supportedcites [39] [39] Lucy Lu Wang, Isabel Cachola, Jonathan Bragg, Evie Yu-Yen Cheng, Chelsea Haupt, Matt Latzke, Bailey Kuehl, Madeleine van Zuylen, Linda Wagner, and Daniel S. Weld. 2021. Improving the Accessibility of Scientific Documents: Current State, User Needs, and a System Solution to Enhance Scientific PDF Accessibility for Blind and Low Vision Users. arXiv:2105.00076 [cs.DL] https://arxiv.org/abs/2105.00076 - source
Table 2 (Section 3.1, p. 7) reports Adobe-5 Compliance of 2.4% for the full dataset of 11,397 PDFs. Section 3.3 (p. 10) also reports that 275 papers, or 2.4%, met all five criteria. (LLM, conf=0.95)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“However, while Ally was notably inaccessible due to the dexterity required for drawing reading order arrows with a mouse, PAVE 2.0 cannot automatically detect elements such as captions and footnotes because of limitations in the classes included in its training dataset.”
What the sources say
- Supportedcites [32] [32] Felix Maximilian Schmitt-Koopmann, Elaine May Huang, Hans-Peter Hutter, and Alireza Darvishy. 2025. Towards More Accessible Scientific PDFs for People with Visual Impairments: Step-by-Step PDF Remediation to Improve Tag Accuracy. In Proceedings of the 2025 CHI Conference on Human Factors in Computing Systems (CHI '25) . Association for Computing Machinery, New York, NY, USA, Article 48, 16 pages. doi:10.1145/3706598.3713084 - source
Section 5.1.1 (Tag Accuracy Results; page not provided) explicitly says that limitations in the training dataset’s classes prevent PAVE 2.0 from detecting captions and footnotes, which it typically tags as paragraphs. This supports the PAVE 2.0 claim attributed to the source. (LLM, conf=0.95)
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“However, even 'silent' tools that automate accessibility with minimal human intervention still require user oversight and understanding to perform effectively [23].”
What the sources say
- Supportedcites [23] [23] Peya Mowar, Yi-Hao Peng, Jason Wu, Aaron Steinfeld, and Jeffrey P Bigham. 2025. CodeA11y: Making AI Coding Assistants Useful for Accessible Web Development. In Proceedings of the 2025 CHI Conference on Human Factors in Computing Systems (CHI '25) . Association for Computing Machinery, New York, NY, USA, Article 45, 15 pages. doi:10.1145/3706598.3713335 - source
The Discussion (p. 12) says that although the tool can 'silently' improve accessibility, users need some understanding to use it effectively and developers must manually implement more challenging aspects. The Discussion (p. 11) also says developers need to intervene in AI-generated code to implement accessibility features effectively. (LLM, conf=0.94)
plainstates a possible cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“However, challenges remain in disambiguating symbols and mapping their intended meaning [20].”
What the sources say
- Supportedcites [20] [20] Frank Mittelbach, Ulrike Fischer, David Carlisle, and Joseph Wright. 2025. MathML and other XML Technologies for Accessible PDF from LATEX. In Proceedings of the 2025 ACM Symposium on Document Engineering (Nottingham, United Kingdom) (DocEng '25) . Association for Computing Machinery, New York, NY, USA, Article 19, 4 pages. doi:10.1145/3704268.3748669 - source
Section 2.2 (Intent) explains that the notation |x| can have several meanings depending on context and that an intent annotation can disambiguate it. It also says LaTeX can add these annotations only in some cases, supporting the claim that challenges remain. (LLM, conf=0.89)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
49 more claims · 2 with their cited sources checked
Create a free account to see all 125 claims and to verify the 6 references still open.
- It may state two separate claims as one. 36
- Probably not a substantial claim. 52
Citations
Cites
You Only Look Once: Unified, Real-Time Object Detection
Redmon, J., Divvala, S., Girshick, R., & Farhadi, A. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.91 · Cited by 53,803 · cited 1×
“Object Detection models, such as YOLO [31], trained or fine-tuned on document layout datasets ( e.g., PubLayNet [41], DocLayNet [28] or DocBank [13]) can often detect visual elements on pages like text and figures with high accuracy.”
DETRs Beat YOLOs on Real-time Object Detection
Zhao, Y., Lv, W., Xu, S., Wei, J., Wang, G., Dang, Q., Liu, Y., & Chen, J. (2024) · 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr52733.2024.01605 · Cited by 5,020
An Overview of the Tesseract OCR Engine
Smith, R. (2007) · Ninth International Conference on Document Analysis and Recognition (ICDAR 2007) Vol 2 · doi:10.1109/icdar.2007.4376991 · Cited by 2,376
Generative Agents: Interactive Simulacra of Human Behavior
Park, J. S., O'Brien, J., Cai, C. J., Morris, M. R., Liang, P., & Bernstein, M. S. (2023) · Proceedings of the 36th Annual ACM Symposium on User Interface Software and Technology · doi:10.1145/3586183.3606763 · Cited by 2,187 · cited 1×
“However, while [27] by (a) accessibility remediation tools (i, v in blue), (b) machine learning pipelines for document layout analysis (ii–iv in lilac), (c) vision-language models for object detection (vi–viii in green), and (d) iTagPDF (ix).”
Grounding DINO: Marrying DINO with Grounded Pre-training for Open-Set Object Detection
Liu, S., Zeng, Z., Ren, T., Li, F., Zhang, H., Yang, J., Jiang, Q., Li, C., Yang, J., Su, H., et al. (2025) · Lecture Notes in Computer Science · doi:10.1007/978-3-031-72970-6_3 · Cited by 2,074
YOLO-World: Real-Time Open-Vocabulary Object Detection
Cheng, T., Song, L., Ge, Y., Liu, W., Wang, X., & Shan, Y. (2024) · 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr52733.2024.01599 · Cited by 815
DocVQA: A Dataset for VQA on Document Images
Mathew, M., Karatzas, D., & Jawahar, C. V. (2021) · 2021 IEEE Winter Conference on Applications of Computer Vision (WACV) · doi:10.1109/wacv48630.2021.00225 · Cited by 513
PubLayNet: Largest Dataset Ever for Document Layout Analysis
Zhong, X., Tang, J., & Jimeno Yepes, A. (2019) · 2019 International Conference on Document Analysis and Recognition (ICDAR) · doi:10.1109/icdar.2019.00166 · Cited by 509
DocLayNet: A Large Human-Annotated Dataset for Document-Layout Segmentation
Pfitzmann, B., Auer, C., Dolfi, M., Nassar, A., & Staar, P. (2022) · ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD) · doi:10.1145/3534678.3539043 · Cited by 378
Lopez, P. (2009) · Lecture Notes in Computer Science · doi:10.1007/978-3-642-04346-8_62 · Cited by 242
DocBank: A Benchmark Dataset for Document Layout Analysis
Li, M., Xu, Y., Cui, L., Huang, S., Wei, F., Li, Z., & Zhou, M. (2020) · Proceedings of the 28th International Conference on Computational Linguistics · doi:10.18653/v1/2020.coling-main.82 · Cited by 217
Le, L. (2024) · International ACM SIGACCESS Conference on Computers and Accessibility (ASSETS) · doi:10.1145/3663548.3675651 · Cited by 50
FigurA11y: AI Assistance for Writing Scientific Alt Text
Singh, N., Wang, L. L., & Bragg, J. (2024) · Proceedings of the 29th International Conference on Intelligent User Interfaces · doi:10.1145/3640543.3645212 · Cited by 42 · cited 1×
“Researchers are actively studying interactive LLM-guided tooling for producing alt text [33] and future work could also look at evaluation of such automatic and semi-automatic alt text generation approaches.”
Creating accessible PDFs for conference proceedings
Brady, E., Zhong, Y., & Bigham, J. P. (2015) · Proceedings of the 12th International Web for All Conference · doi:10.1145/2745555.2746665 · Cited by 35 · cited 4×
“Despite this, persistent challenges remain in ensuring that documents prepared in this format consistently include the metadata necessary to make them accessible to people with disabilities [2, 3].”
InfluentialSelf-citationAn Uninteresting Tour Through Why Our Research Papers Aren't Accessible
Bigham, J. P., Brady, E. L., Gleason, C., Guo, A., & Shamma, D. A. (2016) · Proceedings of the 2016 CHI Conference Extended Abstracts on Human Factors in Computing Systems · doi:10.1145/2851581.2892588 · Cited by 34
Self-citationCodeA11y: Making AI Coding Assistants Useful for Accessible Web Development
Mowar, P., Peng, Y.-H., Wu, J., Steinfeld, A., & Bigham, J. P. (2025) · Proceedings of the 2025 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3706598.3713335 · Cited by 31
Self-citationPDF Accessibility of Research Papers: What Tools are Needed for Assessment and Remediation?
Jembu Rajkumar, A., Lazar, J., Jordan, J. B., Darvishy, A., & Hutter, H.-P. (2020) · Proceedings of the Annual Hawaii International Conference on System Sciences · doi:10.24251/hicss.2020.512 · Cited by 28
Mathspeak: An Audio Method for Presenting Mathematical Formulae to Blind Students
Nazemi, A., Murray, I., & Mohammadi, N. (2012) · 2012 5th International Conference on Human System Interactions · doi:10.1109/hsi.2012.17 · Cited by 26
Wang, L. L., Cachola, I., Bragg, J., Cheng, E. Y.-Y., Haupt, C., Latzke, M., Kuehl, B., van Zuylen, M., Wagner, L., & Weld, D. S. (2021) · arXiv · doi:10.48550/arxiv.2105.00076 · Cited by 22 · cited 3×
“Wang et al. [39] expanded this analysis through automated evaluation of roughly 11k academic PDFs from diverse venues and noted that only 2.4% papers were compliant with Adobe-5 standards.”
InfluentialMaking the field of computing more inclusive
Lazar, J., Churchill, E. F., Grossman, T., van der Veer, G., Palanque, P., Morris, J. "., & Mankoff, J. (2017) · Communications of the ACM · doi:10.1145/2993420 · Cited by 21 · cited 2×
“We compare iTagPDF with two baselines: (i) Adobe Acrobat Pro Auto-Tagging, the most widely used PDF remediation software, and (ii) the higher standard of author-submitted PDFs to ACM CHI and ASSETS conferences [11].”
Kumar, A., & Wang, L. L. (2024) · The 26th International ACM SIGACCESS Conference on Computers and Accessibility · doi:10.1145/3663548.3675634 · Cited by 18 · cited 4×
“First, both Wang et al. [39] and Kumar et al. [10] found that accessibility compliance varies substantially by the typesetting platform, with Microsoft Word producing the most accessible PDFs while LaTeX invariably yielding the least.”
InfluentialPradhan, D., Rajput, T., Rajkumar, A. J., Lazar, J., Jain, R., Morariu, V. I., & Manjunatha, V. (2022) · ACM Transactions on Accessible Computing · doi:10.1145/3507661 · Cited by 15
(2019) · doi:10.1145/3308561.3353808 · Cited by 14
Self-citationLaTeX Tagged PDF—A blueprint for a large project
Mittelbach, F. G., & Rowley, C. (2020) · TUGboat · doi:10.47397/tb/41-3/tb129mitt-tagpdf · Cited by 13
Breaking the Accessibility Barrier in Non-Visual Interaction with PDF Forms
Uckun, U., Aydin, A. S., Ashok, V., & Ramakrishnan, IV (2020) · Proceedings of the ACM on Human-Computer Interaction · doi:10.1145/3397868 · Cited by 12
Author Reflections on Creating Accessible Academic Papers
Menzies, R., Tigwell, G. W., & Crabb, M. (2022) · ACM Transactions on Accessible Computing · doi:10.1145/3546195 · Cited by 12
A framework for improving the accessibility of research papers on arXiv.org
Brinn, S., Cameron, C., Fielding, D., Frankston, C., Fromme, A., Huang, P., Nazzaro, M., Orphan, S., Sigurdsson, S., Tay, R., et al. (2022) · arXiv · doi:10.48550/arxiv.2212.07286 · Cited by 11
Schmitt-Koopmann, F. M., Huang, E. M., Hutter, H.-P., & Darvishy, A. (2025) · Proceedings of the 2025 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3706598.3713084 · Cited by 8 · cited 5×
“Although a number of tools exist for PDF remediation [1, 7, 32], they mostly work in the same way – the tool visualizes each page of the PDF document including any existing metadata information and the user must manually provide (or fix) the metadata.”
InfluentialPaper to HTML: A Publicly Available Web Tool for Converting Scientific Pdfs into Accessible HTML
Wang, L. L., Bragg, J., & Weld, D. S. (2022) · ACM SIGACCESS Accessibility and Computing · doi:10.1145/3582298.3582299 · Cited by 6
Enhancing LaTeX to automatically produce tagged and accessible PDF
Mittelbach, F., & Fischer, U. (2024) · TUGboat · doi:10.47397/tb/45-1/tb139mitt-deims24 · Cited by 6
Will AI allow us to dispense with all or most accessibility regulations?
Vanderheiden, G. C., & Marte, C. (2024) · CHI Conference Extended Abstracts on Human Factors in Computing Systems (CHI EA) · doi:10.1145/3613905.3644059 · Cited by 6
Automatically producing accessible and reusable PDFs with LATEX
Mittelbach, F., Fischer, U., Carlisle, D., & Wright, J. (2024) · Proceedings of the ACM Symposium on Document Engineering 2024 · doi:10.1145/3685650.3685670 · Cited by 4
Mowar, P., Gupta, M., & Jain, M. (2024) · International ACM SIGACCESS Conference on Computers and Accessibility (ASSETS) · doi:10.1145/3663548.3675608 · Cited by 4
Self-citationFormA11y—Research and Development of a Tool for Remediating PDF Forms for Accessibility
Paliwal, S., Hoeflich, J., Jordan, J., Jain, R., Morariu, V. I., Siu, A., & Lazar, J. (2024) · ACM Trans. Comput. Hum. Interact. · doi:10.1145/3702317 · Cited by 3
MathML and other XML Technologies for Accessible PDF from LATEX
Mittelbach, F., Fischer, U., Carlisle, D., & Wright, J. (2025) · Proceedings of the 2025 ACM Symposium on Document Engineering · doi:10.1145/3704268.3748669 · Cited by 1
HTML papers on arXiv -- why it is important, and how we made it happen
Frankston, C., Godfrey, J., Brinn, S., Hofer, A., & Nazzaro, M. (2024) · arXiv · doi:10.48550/arxiv.2402.08954 · Cited by 1 · cited 1×
“Notably, Frankston et al. [8] observed that working directly from the source documents provided a superior starting point compared to PDFs, which is why platforms like ar5iv and the ACM Digital Library directly convert source files into accessible HTML.”
Research works printed but not linked: 1
- Adobe Inc.2023. Adobe Acrobat Pro DC Accessibility Features. Adobe Systems Incorporated. https://helpx.adobe.com/acrobat/using/create-verify-pdf-accessibility.html.
Cited by
(2000) · Time to knit · doi:10.1145/3772318.3809047
Metadata
- Authors
- Mowar, P., Steinfeld, A., & Bigham, J. P.
- Affiliations
- Carnegie Mellon University (US)
- Venue
- Proceedings of the 2026 CHI Conference on Human Factors in Computing Systems
- Publisher
- ACM
- Published
- 2026-04-13
- Pages
- 1-17
- OpenAlex
- W7154067339
- Semantic Scholar
- 87a9235ab1c2c8032212e34aee2a1607670cbfa7
- Cited by
- 2
- References
- 34
- Citations by year
- 2026: 2
- FWCI
- 17.47
- Type
- conference paper
- Language
- en
- Open access
- closed
- Locations
- CHI Conference on Human Factors in Computing Systems (publishedVersion)
- Topics
- Social Sciences › Human Factors and Ergonomics › Digital Accessibility for Disabilities
- Social Sciences › Business, Management and Accounting › Management Information Systems › Information Systems and Technology Applications
- Physical Sciences › Computer Science › Computational Theory and Mathematics › Mathematics, Computing, and Information Processing
- Keywords
- PDF accessibility, document tagging, assistive technology
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
The claim verdicts this report counts are experimental: their accuracy is not yet measured.
A language model drew the claims from the paper on 2026-10-25, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-25 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.