CasiMedicos-Arg: A Medical Question Answering Dataset Annotated with Explanatory Argumentative Structures
Project Overview
The document highlights the innovative use of generative AI in education, particularly through the development of the CasiMedicos-Arg dataset, a multilingual Medical Question Answering (QA) resource designed to enhance the explanatory abilities of medical residents. By providing annotated argumentative structures for 558 clinical cases in four languages, this dataset serves a crucial role in training medical professionals to justify their diagnoses and treatment decisions. It addresses a significant gap in AI tools for the medical field, promoting the application of argument mining and generative techniques to foster better explanations in clinical contexts. The findings suggest that by integrating such AI-driven resources, medical education can improve, leading to more competent healthcare practitioners who can articulate their reasoning effectively. Overall, the document underscores the potential of generative AI to transform educational practices in specialized fields like medicine, enabling learners to engage with complex material through structured, annotated support.
Key Applications
CasiMedicos-Arg: A multilingual dataset for Medical Question Answering with annotated argumentative structures
Context: Medical education for resident doctors preparing for licensing exams
Implementation: The dataset was created by enriching clinical cases with argumentation annotations, including claims and premises, derived from explanations provided by medical doctors.
Outcomes: The dataset enhances the ability of AI systems to generate explanations for medical diagnoses, supporting educational objectives for medical residents.
Challenges: The lack of existing datasets that include argumentative explanations and the need for multilingual support in medical QA tasks.
Implementation Barriers
Data Availability and Model Performance
Existing medical QA datasets are primarily in English and do not include argumentative explanations. Current medical QA benchmarks focus on correct answer identification but not on generating argumentative explanations.
Proposed Solutions: The CasiMedicos-Arg dataset provides a multilingual approach by translating and annotating clinical cases in four languages, aiming to improve the quality of AI-generated explanations by offering a structured argumentative framework.
Project Team
Ekaterina Sviridova
Researcher
Anar Yeginbergen
Researcher
Ainara Estarrona
Researcher
Elena Cabrio
Researcher
Serena Villata
Researcher
Rodrigo Agerri
Researcher
Contact Information
For information about the paper, please contact the authors.
Authors: Ekaterina Sviridova, Anar Yeginbergen, Ainara Estarrona, Elena Cabrio, Serena Villata, Rodrigo Agerri
Source Publication: View Original PaperLink opens in a new window
Project Contact: Dr. Jianhua Yang
LLM Model Version: gpt-4o-mini-2024-07-18
Analysis Provider: Openai