The project aims to develop an AI-powered chat system using RAG for a question and answer-based healthcare application. The primary objective is to assist users in identifying potential diseases based on the symptoms they provide. The system engages users in a conversational manner, asking relevant questions to gather detailed information about their symptoms and, subsequently, suggesting possible diagnoses.
Data Used:
We used three data sources: Manual of Surgery Volume First General Surgery Sixth Edition Principles of Public Health Surgical Anatomy Which belong to the Gutenberg project and are in the public domain.
Changes in the Prompt:
The prompt outlines the system's functionality, emphasizing the importance of asking clear questions, maintaining a friendly tone, and avoiding the invention of medical conditions. One significant modification is the incorporation of the RAG framework, enabling the system to generate responses in adherence to specific rules, ensuring transparency and accuracy. Additionally, the prompt includes explicit instructions to provide hyperlinks to the sources for each piece of information, enhancing traceability and credibility.
Beneficiaries:
The application targets users seeking preliminary insights into potential health issues based on their symptoms. It benefits individuals by providing a convenient and accessible platform to gather information about their health concerns. The friendly and understanding tone in interactions aims to alleviate user concerns while making it clear that the suggestions are not a substitute for professional medical advice. Healthcare professionals may also find the tool useful for providing initial guidance to users.
Project name
Doctor Robot
Description
Project Description:
The project aims to develop an AI-powered chat system using RAG for a question and answer-based healthcare application. The primary objective is to assist users in identifying potential diseases based on the symptoms they provide. The system engages users in a conversational manner, asking relevant questions to gather detailed information about their symptoms and, subsequently, suggesting possible diagnoses.
Data Used:
We used three data sources: Manual of Surgery Volume First General Surgery Sixth Edition Principles of Public Health Surgical Anatomy Which belong to the Gutenberg project and are in the public domain.
Changes in the Prompt:
The prompt outlines the system's functionality, emphasizing the importance of asking clear questions, maintaining a friendly tone, and avoiding the invention of medical conditions. One significant modification is the incorporation of the RAG framework, enabling the system to generate responses in adherence to specific rules, ensuring transparency and accuracy. Additionally, the prompt includes explicit instructions to provide hyperlinks to the sources for each piece of information, enhancing traceability and credibility.
Beneficiaries:
The application targets users seeking preliminary insights into potential health issues based on their symptoms. It benefits individuals by providing a convenient and accessible platform to gather information about their health concerns. The friendly and understanding tone in interactions aims to alleviate user concerns while making it clear that the suggestions are not a substitute for professional medical advice. Healthcare professionals may also find the tool useful for providing initial guidance to users.
Language
English
Project Repository URL
https://github.com/maxquinteros/desert-coders-azure-search-openai
Deployed Endpoint URL
https://app-backend-rweaj6ni7t444.azurewebsites.net/#/
Project video
https://www.youtube.com/watch?v=I_Z5YxrX6hM
Team members
maxquinteros, JosephFloresA, PabloNAGG
Showcase Consent
Yes