🛡️ MIL SHIELD - NexTech Hackathon @ NMIET: TeamEdgeHackers

An Advanced Military-Grade Network Security System for Army & Defense - Powered by Machine Learning & Large Language Models

🎯 Project Overview

MIL SHIELD is an intelligent cybersecurity platform designed specifically for Army and Defense operations, combining traditional machine learning models with cutting-edge Large Language Model (LLM) reasoning to provide real-time network threat detection and automated response capabilities. Developed for the NexTech Hackathon at NMIET (Narsee Monjee Institute of Engineering and Technology), this system represents the next generation of military-grade network security tailored for defense infrastructure protection.

🌟 Key Features

• Military-Grade Security: Multi-layered AI defense combining Random Forest, Autoencoder, and LLM models
• Real-time Threat Detection: Live network monitoring with instant alerts for defense operations
• Intelligent Decision Making: AI-powered response recommendations for military scenarios
• Visual Attack Simulation: Interactive camera-based attack scenarios for training purposes
• Command & Control Dashboard: Military-style Streamlit interface for monitoring operations
• Comprehensive Analytics: Detailed visualization and threat intelligence reporting
• Field-Ready Deployment: Optimized for edge computing in defense environments
• Secure Communication: Built for classified network environments
• Threat Intelligence: Advanced pattern recognition for emerging cyber threats

📁 Project Structure

📦 NexTechHackathon-TeamEdgeHackers/
├── 🤖 modelTraining/              # Core ML model development
│   ├── models/                    # Model architectures
│   ├── preprocessing/             # Data preprocessing pipeline
│   ├── training_scripts/          # Training workflows
│   ├── data/                      # Datasets and processed data
│   ├── exports/                   # Trained models and scalers
│   ├── visualization/             # Model performance visualizations
│   └── gui-streamlit.py          # Interactive dashboard
│
├── 🎯 attack/                     # Attack simulation environment
│   ├── index4.py                 # Advanced PyQt5 attack simulator
│   ├── stream.py                 # Webcam streaming server
│   ├── attack_data_exporter.py   # Attack data generation
│   └── data/                     # Attack scenario datasets
│
├── 🧠 modelPlusLLMPrediction/     # LLM-enhanced decision system
│   ├── src/                      # Core system modules
│   │   ├── core/                 # Decision making engines
│   │   ├── network_io/           # Network event handling
│   │   └── utils/                # Utility functions
│   ├── configs/                  # Configuration files
│   ├── models/                   # Pre-trained model storage
│   ├── scripts/                  # Deployment scripts
│   └── final_demo.py            # Complete system demonstration
│
└── 📋 Configuration Files
    ├── .gitignore               # Git ignore patterns
    └── README.md               # This file

🚀 Quick Start Guide

Prerequisites

• Python 3.8+ with pip
• NVIDIA GPU with CUDA support (for LLM inference)
• Webcam (for attack simulation)
• Windows 10/11 or Linux

1️⃣ Environment Setup

# Clone the repository
git clone https://github.com/mohitksahu/NexTechHackathon-TeamEdgeHackers.git
cd NexTechHackathon-TeamEdgeHackers

# Create virtual environment
python -m venv .venv

# Activate environment (Windows)
.venv\Scripts\activate
# Or on Linux/Mac
source .venv/bin/activate

2️⃣ Install Dependencies

# Install model training dependencies
cd modelTraining
pip install -r requirements.txt

# Install LLM prediction system
cd ../modelPlusLLMPrediction
pip install -r requirements.txt

# Install attack simulation tools
cd ../attack
pip install -r req.txt

3️⃣ Launch the System

Option A: Interactive Dashboard

cd modelTraining
streamlit run gui-streamlit.py

Option B: LLM-Enhanced Analysis

cd modelPlusLLMPrediction
python final_demo.py

Option C: Attack Simulation

cd attack
python index4.py  # Launch PyQt5 simulator
# Or
python stream.py  # Start webcam streaming

🎮 Usage Examples

📊 CyberAttack Dashboard

The interactive dashboard provides real-time monitoring capabilities:

1. Download Network Logs: CSV files with network traffic data
2. Real-time Analysis: Live threat detection with ML models
3. Visualization: Interactive charts and attack pattern analysis
4. Alert System: Immediate notifications for detected threats

🤖 LLM Decision Engine

The AI-powered decision system offers intelligent threat response:

# Example: Process a network event
from src.core.decision_maker import DecisionMaker

decision_maker = DecisionMaker()
result = decision_maker.process_event({
    "src_ip": "192.168.1.100",
    "dst_ip": "10.0.0.50",
    "protocol": "TCP",
    "threat_score": 0.85
})

print(f"Decision: {result['action']}")
print(f"Confidence: {result['confidence']}")
print(f"Reasoning: {result['explanation']}")

🎯 Attack Simulation

Interactive attack scenarios for testing:

1. Camera-based Simulation: Visual attack pattern recognition
2. Network Traffic Generation: Synthetic attack data creation
3. Real-time Response Testing: Validate system responses

🧪 Model Performance

Model	Accuracy	Precision	Recall	F1-Score
Random Forest	94.2%	93.8%	94.1%	93.9%
Autoencoder	91.7%	90.5%	92.3%	91.4%
Hybrid System	96.8%	96.2%	96.5%	96.3%

🔧 Configuration

Model Training Settings

# modelTraining/config.yaml
random_forest:
  n_estimators: 100
  max_depth: 20
  random_state: 42

autoencoder:
  hidden_layers: [64, 32, 16, 32, 64]
  epochs: 100
  batch_size: 32

LLM Configuration

# modelPlusLLMPrediction/configs/llm_config.yaml
mode: local
model_name: microsoft/DialoGPT-medium
quantization: q4_0
context_window: 2048

🛠️ Development

Adding New Models

1. Create model class in modelTraining/models/
2. Implement training script in training_scripts/
3. Update preprocessing pipeline if needed
4. Add visualization in visualization/

Extending LLM Capabilities

1. Modify modelPlusLLMPrediction/src/core/llm_agent.py
2. Update configuration in configs/llm_config.yaml
3. Add new decision logic in decision_maker.py

Custom Attack Simulations

1. Extend attack/index4.py with new scenarios
2. Add data generation in attack_data_exporter.py
3. Update visualization components

📈 Performance Optimization

• CUDA Acceleration: GPU-optimized inference
• Model Quantization: 4-bit quantization for efficiency
• Batch Processing: Optimized data pipeline
• Memory Management: Efficient resource utilization

🤝 Contributing

We welcome contributions! Please see our contributing guidelines:

1. Fork the repository
2. Create a feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

📋 Requirements

Core Dependencies

• tensorflow>=2.17.0
• scikit-learn>=1.5.1
• pandas>=2.2.2
• numpy>=1.26.4
• streamlit
• PyQt5
• opencv-python

LLM Dependencies

• torch
• transformers
• accelerate
• bitsandbytes

🏆 Team EdgeHackers

• Team Lead: Mohit K Sahu
• Team Memvers: Lokesh Jena Subham Mohapatra RudraBhuyan
• ML Engineers: Specialized in deep learning and cybersecurity
• Full-Stack Developers: Dashboard and system integration
• Security Analysts: Threat modeling and validation

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

🙏 Acknowledgments

• NexTech Hackathon @ NMIET organizers for this incredible opportunity to innovate for defense
• NMIET Faculty & Staff for supporting defense technology development
• Indian Army & Defense Forces for inspiring this security solution
• CIC-IDS2017 dataset for comprehensive cybersecurity training data
• Hugging Face for transformer models enabling advanced AI capabilities
• Streamlit for providing an intuitive dashboard framework
• Defense Research Community for continuous cybersecurity innovation
• References: (https://www.unb.ca/cic/datasets/index.html) (https://share.google/sU2mFIb9Vg6AZn8cT) (https://huggingface.co/collections/TinyLlama/tinyllama-11b-v11-660bb405bf46efd55c2094fc)

📞 Support

For questions and support:

• 📧 Email: [email protected]
• 💬 Issues: GitHub Issues

---

🛡️ Defending the Nation with AI-Powered Intelligence 🇮🇳

Built with ❤️ by Team EdgeHackers for NexTech Hackathon @ NMIET

"Securing Defense Networks, Protecting Our Forces"