CLaCLab at SocialDisNER: Using Medical Gazetteers for Named-Entity Recognition of Disease Mentions in Spanish Tweets
This paper discusses a method for identifying mentions of diseases in Spanish tweets, which is important for tracking health trends, especially during pandemics.
This video presentation explains the key concepts from the paper in plain language.
Content & Liability Disclaimer
This article and its accompanying video are automated summaries derived from the original research paper by Unknown authors. The original research was conducted solely by the paper's authors; PDFdigest did not conduct any of the research and makes no claims of ownership over the underlying scientific work.
The video narration is generated by artificial intelligence and references the paper's authors for attribution. The video is not narrated by any of the paper's authors. This content may contain inaccuracies, omissions, or misinterpretations of the original research. First-person language (e.g., "we found", "our results") reflects the original authors' voice, not PDFdigest's. Always read the original paper for accurate, verified information before making any decisions based on this content.
This content is provided "as is" without any warranties, express or implied. Simulated systems OÜ, its officers, directors, employees, and agents shall not be liable for any direct, indirect, incidental, special, consequential, or punitive damages arising from your use of, reliance on, or access to this content, including but not limited to errors, omissions, or misinterpretations of the original research. This disclaimer applies to the fullest extent permitted by applicable law.
- 1 M RoBERTa is the baseline model without one-hot features, custom tokenization, or transformer encoder.
- 2 We separate word components in composites because disease mentions can occur there.
- 3 We classify tokens as B, I, or O based on their position in a disease name.
- 4 We concatenate H with G umls, G distemist, and G silver to produce Z.
Introduction
Finding disease mentions in multilingual tweets aids epidemiologists during pandemics. So-cialDisNER at SMM4H 2022 recognizes disease mentions in Spanish tweets.
Disease mentions use lay or professional language and appear in hashtags, usernames, or tweet text.
Sumemos 1 millu00f3n de pasos por la epilepsia, #1MillonDePasos, #Epilepsia, #ADosMetrosDeDistancia #InvestigaEpilepsia @RetoDravet #juntosesmejor @FundacionDravet.
Ad hoc gazetteer lists often have limited coverage.
Methodology
Sequence analysis for BIO tagging requires good tokenization. We frame NER as a sequence labeling task using the BIO tagging scheme.
Study Design
Results & Findings
We separate word components in composites because disease mentions can occur there. We classify tokens as B, I, or O based on their position in a disease name.
- We separate word components in composites because disease mentions can occur there.
- We classify tokens as B, I, or O based on their position in a disease name.
- We concatenate H with G umls, G distemist, and G silver to produce Z.
- A linear classifier feeds Y to produce matrix I for BIO classification.
- Section 2.2 described our submission model M sub.
M RoBERTa is the baseline model without one-hot features, custom tokenization, or transformer encoder.
We separate word components in composites because disease mentions can occur there.
How PDFdigest Helps You Understand Research
Instant Paper Analysis
Get structured summaries and key findings from dense PDFs in seconds.
Visual Explanations
Turn complex methods, figures, and results into clearer visual breakdowns.
AI-Powered Q&A
Ask focused questions and get answers grounded in the paper.
Motivation
Finding mentions of diseases in tweets in all languages has become an important tool for epidemiologists, especially during a pandemic. The SocialDisNER task focuses on recognizing disease mentions in Spanish tweets, which can appear in various formats.
System
The system consists of a pipeline with four components: a tokenizer, word embeddings from RoBERTa Large, gazetteer lists for disease recognition, and a linear classifier for BIO tagging. The approach emphasizes simplicity and knowledge injection from domain resources.
Model
The performance of the submission model on the test set shows a minor decrease, indicating robustness in the technique used.
Figures Explained
The paper’s visual material highlights the workflow and the main system components.
- Figure 1: GoldGaz: disease names compiled from the gold annotations of SocialDisNER training and validation dataDistemistGaz: (Miranda-Escalada et al., 2022) Spanish disease gazetteer compiled from Snomed-CT (Donnelly and others, 2006).
Frequently Asked Questions
Disease mentions use lay or professional language and appear in hashtags, usernames, or tweet text. The i-th row of H is the contextualized embedding vector for the i-th token.
Sequence analysis for BIO tagging requires good tokenization. We frame NER as a sequence labeling task using the BIO tagging scheme.
We separate word components in composites because disease mentions can occur there. M RoBERTa is the baseline model without one-hot features, custom tokenization, or transformer encoder.
Ad hoc gazetteer lists often have limited coverage.
This paper discusses a method for identifying mentions of diseases in Spanish tweets, which is important for tracking health trends, especially during pandemics.
Yes. PDFDigest can turn this paper into a structured explanation, key takeaways, visual summaries, and a narrated video when available.