Event Detection from Social Media for Epidemic Prediction (2404.01679v2)

Abstract: Social media is an easy-to-access platform providing timely updates about societal trends and events. Discussions regarding epidemic-related events such as infections, symptoms, and social interactions can be crucial for informing policymaking during epidemic outbreaks. In our work, we pioneer exploiting Event Detection (ED) for better preparedness and early warnings of any upcoming epidemic by developing a framework to extract and analyze epidemic-related events from social media posts. To this end, we curate an epidemic event ontology comprising seven disease-agnostic event types and construct a Twitter dataset SPEED with human-annotated events focused on the COVID-19 pandemic. Experimentation reveals how ED models trained on COVID-based SPEED can effectively detect epidemic events for three unseen epidemics of Monkeypox, Zika, and Dengue; while models trained on existing ED datasets fail miserably. Furthermore, we show that reporting sharp increases in the extracted events by our framework can provide warnings 4-9 weeks earlier than the WHO epidemic declaration for Monkeypox. This utility of our framework lays the foundations for better preparedness against emerging epidemics.

• The paper introduces a novel method to detect events from social media for real-time epidemic prediction.
• It employs data mining and natural language processing techniques to extract actionable health signals.
• Experimental results reveal improved prediction accuracy over traditional models, aiding proactive public health responses.

Guidelines for EMNLP 2023 Submissions: A Detailed Overview

Introduction to Submission Guidelines

The document under review serves as a supplementary guide complementing the general instructions for authors preparing submissions for the EMNLP 2023 conference. It particularly addresses how to effectively use the LaTeX style file designated for EMNLP 2023, in alignment with the formatting requirements specified by the Association for Computational Linguistics (ACL). This guide is not only pertinent for papers under review but also extends its applicability to the final versions of the accepted papers, showcasing an example manuscript that adheres to the stipulated formatting and style guidelines.

Document Preparation and Formatting

LaTeX Templates and Commands

The guidelines provide comprehensive instructions regarding the preparation of a manuscript using LaTeX, emphasizing the utilization of pdfLaTeX for generating the final PDF file. A notable mention is the suitability of XeLaTeX for manuscripts that incorporate non-Latin scripts. Authors are guided through the process of employing appropriate commands to handle special characters—a common necessity in bibliographies—ensuring correct rendering and sorting of references.

Title and Author Formatting

Instructions delineate the procedure for setting the title and author information, highlighting the utilization of commands such as \title, \author, and various conjunctions for formatting multiple authors. It specifies the default size for the title and author box and gives directions for adjusting this, should there be a need for more space.

Manuscript Structure

Managing Footnotes and Citations

The document delineates the insertion of footnotes and the correct format for tables and figures, ensuring that captions adhere to default sizes. It addresses potential compilation errors associated with hyperlinks, particularly when using older LaTeX versions, and recommends an update to mitigate these issues.

For citations, a table is provided to clarify the syntax supported by the style files, which are based on the natbib package. This aids authors in employing various citation styles effectively within their manuscripts.

References Section

Authors are guided on the preparation of the references section, encouraging the inclusion of DOIs or URLs in BibTeX entries whenever possible. This facilitates linking the paper titles in the references section to online resources, enhancing accessibility and ensuring compliance with the ACL Anthology's referencing standards.

Additional Components

Appendices

Guidelines are offered on how to incorporate appendices into the manuscript, transitioning the section numbering to letters and providing an illustrative example of an appendix section.

Ethical and Limitation Disclosures

It is mandated that submissions include a "Limitations" section to discuss the constraints of the work, complementing the strengths discussed in the main text. Furthermore, an ethics statement is encouraged, elaborating on the broader impact and ethical considerations of the research, aligning with the ACL Ethics Policy.

Conclusion

The guide for EMNLP 2023 submissions serves as an essential tool for authors, ensuring that manuscripts not only conform to the technical standards of LaTeX formatting but also adhere to ethical guidelines and recognition of limitations. This structured approach to manuscript preparation is envisioned to streamline the review process and elevate the quality of submissions, promoting clarity, precision, and accountability in research dissemination.

In anticipation of future developments in AI research, adherence to these detailed guidelines paves the way for more robust and coherent scientific communication. As AI research continues to evolve, the integration of comprehensive ethical considerations and acknowledgment of research limitations will remain paramount in fostering responsible and impactful scientific advancements.