README.md

---
title: "Healthcare AI Research Institute (HARI) - SNUH"
emoji: 🏥
colorFrom: blue
colorTo: indigo
sdk: static
pinned: true
---

## 📢 Announcement

Stay updated with the latest major developments from the **Healthcare AI Research Institute (HARI) at Seoul National University Hospital**!

---
**December 12, 2025**

* 🧠 **Model Release: A new medical reasoning model [`hari-q2.5-thinking`](https://huggingface.co/snuh/hari-q2.5-thinking) is now available!**
   - Our model achieved 89.2% accuracy on the KMLE and  and 88.36% on the USMLE QA benchmark.


---
**June 12, 2025**

* 🧠 **Model Release: Our first medical reasoning model [`hari-q3`](https://huggingface.co/snuh/hari-q3-14b) achieved 84.1% accuracy on the KMLE and is now available!**

---
**May 24, 2025**
* ✨ **New Dataset: [`KorMedLawQA`](https://huggingface.co/datasets/snuh/KorMedLawQA) is now available!**  
   - We are thrilled to release **KorMedLawQA**, a new dataset featuring high-quality multiple-choice questions based on South Korean medical law. Each question is accompanied by detailed legal reasoning, designed to support the development of specialized medical LLMs for the Korean healthcare environment, assist in KMLE preparation, and serve as a valuable benchmark.

---
**May 1, 2025**
* 🧠 **Model Release: [`hari-q2.5`](https://huggingface.co/snuh/hari-q2.5) achieving 84.6% accuracy on the KMLE is now available!**
   - This advanced medical LLM is designed to support clinical decision-making and medical education. Stay tuned for more updates!

---
**March 21, 2025**
* 📊 **Dataset Release: [`ClinicalQA`](https://huggingface.co/datasets/snuh/ClinicalQA) dataset is now available!**  
   - We have released a dataset of over 1,000 high-quality question-answer pairs covering diverse clinical scenarios to enhance medical AI development and benchmarking.

---

🤗 **Join us in shaping the future of AI-driven healthcare!**

---
## 🌍 Vision & Mission  
🔹 **Vision**: Shaping a sustainable and healthy future through pioneering AI research.  
🔹 **Mission**:  
   - Develop **cutting-edge AI technologies** for clinical and operational excellence.  
   - Foster **collaborations** between clinicians, AI researchers, and industry leaders.  
   - Lead **global healthcare AI commercialization** and regulatory frameworks.  
   - Train **next-generation AI-driven medical professionals**.  

---
## 🏗 Research Platform  
- **AI-Powered Medical Data Research Platforms**: SUPREME, [SNUHUB](https://snuhub.snuh.org), DeView, [VitalDB](https://vitaldb.net), [Korea Health Data Platform](https://khdp.net)
- **Supercomputing AI Infrastructure**: NVIDIA B200, H100, A100 GPUs, Quantum Computing Initiative.  

---
## 💡 Key Initiatives  
### 🔬 AI Development for Clinical Application  
- AI-assisted **Clinical note summarization**, **medical imaging diagnostics**, **real-time patient monitoring**, and more!
- AI-powered **electronic health records (EHR) automation**.

### 📡 Global AI-Driven Healthcare Data Platform  
- Secure **cross-institutional medical data-driven AI research collaboration**.  
- **Responsible AI frameworks** for safe AI adoption in healthcare.  

### 🎓 **AI Training & Education**
- **Basic AI Education Program**: Designed for **physicians, researchers, and medical students**, covering fundamental AI concepts, medical data analytics, and ethical AI usage.
- **Advanced AI Research Program**: Focused on AI model development, deep learning in healthcare, and clinical AI validation, aimed at senior researchers and specialists.

---
## 🤝 Collaborate with Us!  
HARI is open to partnerships with:  
- **AI research institutions** and **academic collaborators**.  
- **Healthcare technology companies** and **startups**.  
- **Government agencies** shaping AI policy in healthcare.  

For inquiries and collaboration, please contact us at:  
📧 **hhoon@snu.ac.kr**  
🌐 [Seoul National University Hospital](https://www.snuh.org)  

---
🚀 *Together, we are shaping the future of AI-driven healthcare!* 🤗