-
-
Signup Page: Prompts users to complete a questionnaire about body data, goals, exercise, habits, and diet.
-
Dashboard Page: Offers analysis of biological data, including sleep, heart rate, dietary records and medication logs.
-
Query Page: Provides a TCM wellness knowledge base for searching by keywords or tags.
-
Chatbot Page: Lets users log dietary habits and exercise routines while querying records and seeking wellness advice
Inspiration
Background
In 2024, wearable technology has advanced, focusing on health monitoring, AI integration, and remote patient care. Smartwatches now track blood pressure and offer personalized health insights, while IoT enhances functionality, promoting proactive health management.
Traditional Chinese Medicine (TCM) is gaining global recognition, with practices like acupuncture accepted in over 180 countries. Its integration with modern medicine provides a holistic healthcare approach, supported by research and WHO endorsement, improving wellness and patient outcomes.
AI is also transforming personalized healthcare by enhancing diagnostics and enabling predictive analytics for early disease detection. By analyzing medical data from wearables, AI is creating personalized wellness programs, while IoT allows real-time health tracking.
Current Challenges
The health app market is fragmented. Some apps, like Apple Watch, focus on basic monitoring, while advanced tools only intervene after issues arise. This leaves a gap in preventative healthcare where real-time data could be used for proactive advice.
Our Solution
Our solution combines biological data with wellness practices from TCM and Western medicine. This integrated approach offers personalized advice on diet, exercise, and stress management, aiming to prevent health issues before they occur and providing a cutting-edge solution in smart healthcare.
How we built it
Releated Work
We researched bio-data analysis and LLM medical models, focusing on advancements in wearable devices and IoT systems for personalized health assessments. An IoT framework uses machine learning models like Random Forest and Logistic Regression to analyze metrics such as heart rate and BMI for tailored health recommendations. Simultaneously, LLMs like Health-LLM predict health conditions based on wearable data, transforming diagnostics and health predictions through fine-tuning with specialized datasets.
Technical Implementation
Based on our research, we are developing our platform in three key areas: full-stack development, bio-data analysis, and LLM model fine-tuning. We’ve created three main pages: a Health Dashboard, a TCM Knowledge Base, and an LLM Medical Chatbot, utilizing a TypeScript frontend, Python backend, and PostgreSQL database.
Our work involved data cleaning, analysis, visualization, and model training using SAS Viya. We applied advanced machine learning techniques like RNN and LSTM to ensure accurate data prediction and extrapolation.
Additionally, we used Nvidia AI Workbench to fine-tune LLMs for predicting health conditions based on wearable data, employing RAG to deliver personalized suggestions through real-time contextual information.
Testing and Optimization
In the final phase, we conducted thorough testing, hosted the back-end, launched the website, deployed code on GitHub, and produced detailed reports and demo videos. This process ensured seamless functionality and clearly demonstrated our product concept.
What We Learned
Product Development
We successfully navigated the entire product lifecycle within one month, deepening our understanding of development, data analysis, and AI model fine-tuning. This experience allowed us to explore new features for Nvidia AI Workbench.
Team Collaboration
As team leader, I focused on balancing challenges with effective implementation. Despite the high volume of tasks and tight deadlines, we prioritized effectively. The team's active learning and problem-solving were key to our success.
Skills Enhancement
Although this wasn’t our first hackathon, we pushed ourselves beyond basic projects to create a meaningful AI application using Nvidia AI Workbench.
Conclusion
In conclusion, we are grateful for the opportunity to learn more about Nvidia and Dell while completing our hackathon project. This experience provided us with valuable knowledge and skills, and we hope our efforts yield a positive outcome in the competition.
Built With
- nextjs
- postgresql
- python
- react
- sas
- typescript
Log in or sign up for Devpost to join the conversation.