File size: 5,921 Bytes
9e65751
 
 
 
 
 
c16e89f
 
 
c6fd2dc
 
 
 
 
9e65751
 
c6fd2dc
9e65751
0fb6a2e
c6fd2dc
 
 
 
 
9e65751
c16e89f
 
 
c6fd2dc
c16e89f
 
 
c6fd2dc
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
9e65751
ce940cd
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
9e65751
c6fd2dc
9e65751
ce940cd
c6fd2dc
ce940cd
 
 
9e65751
ce940cd
 
 
 
206a5f7
9e65751
 
 
ce940cd
 
 
9e65751
 
 
ce940cd
c6fd2dc
9e65751
 
 
 
 
c6fd2dc
9e65751
 
c6fd2dc
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
---
tags:
- chatbot
- evaluation
- first-aid
license: mit
language:
- en
metrics:
- faithfulness
- answer_relevancy
- context_recall
- answer_correctness
- semantic_similarity
---

# πŸ₯ First Aid Tutor



## πŸ“˜ About the Project
First Aid Tutor is an **AI-powered chatbot** designed to evaluate the efficacy of **Retrieval-Augmented Generation (RAG)** in medical education for paramedical and medical students. It provides **first-aid guidance** based on medical literature to assess learning outcomes and response accuracy.

This repository contains the **codebase** for the **First Aid Assistant** chatbot, tested on **10 first-aid-related questions**.

## πŸ“’ **Ownership and Attribution**
> **⚠️ Disclaimer:**  
> - The **First Aid Assistant chatbot** is the property of **Mysore Medical College and Research Institute**.  
> - The uploaded **FIRST AID PDF** that forms the basis of the **Retrieval-Augmented Generation (RAG)** process is an **official course material** and the property of **Mysore Medical College and Research Institute**.

---

# πŸš‘ First Aid Tutor - RAG Chatbot

The **First Aid Tutor** is a **Retrieval-Augmented Generation (RAG)** chatbot designed to assist **medical and paramedical students** in learning **first-aid protocols**. It is strictly trained on a **verified first-aid guide** and only provides information present in the **First Aid PDF**.

## 🎯 Features
βœ… **Evidence-Based Responses** - The chatbot only retrieves and generates answers based on the **First Aid PDF**.  
βœ… **User-Friendly Interface** - Developed using **Gradio**, featuring a modern and intuitive UI.  
βœ… **Medical Education Research** - Used for evaluating **RAG efficacy in medical education**.  

## πŸ”¬ How It Works
- Uses **TF-IDF + Cosine Similarity** for retrieving contextually relevant information.  
- Queries are answered based strictly on **verified medical content**.  
- Integrated with **Gradio UI** for an easy-to-use chatbot experience.  

## πŸ“Œ Live Demo
πŸš€ **[Try the First Aid Tutor Chatbot](https://huggingface.co/spaces/DrSyedFaizan/First_Aid_Assistant)**

## πŸ“₯ How to Run Locally
You can run the chatbot locally using the following commands:

```
git clone https://github.com/SYEDFAIZAN1987/First-Aid-Tutor
cd First-Aid-Tutor
pip install -r requirements.txt
python raggpt.py
```


# 🩺 First Aid Assistant - Model Evaluation Report

This repository presents the evaluation results of the **First Aid Assistant** chatbot, which provides first aid guidance based on common emergency conditions. The model has been evaluated using the **RAGAS** framework with metrics that assess the quality of the generated answers.

## πŸ“Š **Evaluation Metrics**

The chatbot was evaluated based on the following RAGAS metrics:

- **Answer Relevancy:** Measures how relevant the response is to the user's question.
- **Answer Correctness:** Compares the generated response to the ground truth to assess factual correctness.
- **Semantic Similarity:** Evaluates how semantically similar the generated answer is to the reference answer.

---

## πŸš€ **Performance Summary**

| **Metric**               | **Average Score** |
|:--------------------------|:-----------------:|
| **Answer Relevancy**      | **0.94**          |
| **Answer Correctness**    | **0.91**          |
| **Semantic Similarity**   | **0.97**          |

---

## πŸ“ˆ **Detailed Results**

Here’s a snapshot of the evaluation for some sample questions:

| **Question**                                            | **Answer Relevancy** | **Answer Correctness** | **Semantic Similarity** |
|---------------------------------------------------------|----------------------|------------------------|-------------------------|
| What are the first aid measures for high fever in infants? | 0.93                 | 0.85                   | 0.98                    |
| What are the signs and symptoms of low blood sugar?       | 0.85                 | 0.98                   | 0.94                    |
| What does RICE stand for in first aid treatment?          | 0.99                 | 1.00                   | 0.98                    |
| What is the treatment of snake bite?                      | 0.96                 | 1.00                   | 0.98                    |
| How do you provide first aid for choking?                 | 0.96                 | 0.97                   | 0.98                    |

---

## πŸ“‹ **Key Insights**

- The chatbot performed exceptionally well in **semantic similarity** (average score of **0.97**), indicating that responses are closely aligned with the ground truth.
- **Answer correctness** is strong overall but showed slight variability, suggesting room for improvement in handling complex queries.
- The **relevancy** of responses remained consistently high, reflecting the model's ability to address user questions effectively.

---

## πŸ“ **Evaluation Artifacts**

- **RAGAS Evaluation Report:** [View Full Report](https://huggingface.co/spaces/DrSyedFaizan/FirstAidTutor_RAGAS_Evaluation)

---

## 🌟 **Conclusion**

The **First Aid Assistant** demonstrates reliable performance in answering first aid-related queries with high semantic accuracy and relevancy. Continuous improvement in factual correctness will further enhance its capability to provide life-saving information in emergency situations.

---


## πŸ“₯ Download & Re-Evaluate
You can **re-evaluate the chatbot** by running the following:

```
git clone https://huggingface.co/DrSyedFaizan/First_Aid_Assistant_Evaluation
cd First_Aid_Assistant_Evaluation
python eval.py
```

---

## πŸ“Œ Author
πŸ‘¨β€βš• **Dr. Syed Faizan** | πŸ₯ **Healthcare AI & Data Science**  
πŸ”— **Hugging Face Repo**: [First_Aid_Assistant_Evaluation](https://huggingface.co/DrSyedFaizan/First_Aid_Assistant_Evaluation)

πŸ“© **For collaborations, inquiries, or improvements, feel free to contribute or reach out!**