Page 471 - AI for Good Innovate for Impact

AI for Good Innovate for Impact

Use Case 16: AI Cyber Security Edge and Cloud based Intrusion

Detection System for IoT 4.4-Productivity

Organization: Addis Ababa Science and Technology University

Country: Ethiopia

Contact: Mr. Kibremoges Fenta , email: kibremogesf@ gmail �com

1 Use Case Summary Table

Item Cyber Security
Category Productivity

Problem Addressed The increasing vulnerability of interconnected IoT devices due to weak
access point security and the lack of real-time, intelligent monitoring
mechanisms.
Key Aspects of Solu- • AI-powered real-time intrusion detection deployed at the edge, like
tion routers
• Cloud-base analytics for flagged threats
• Continuous training and fine tuning for accuracy and speed
• Subscription/contact based access control for traffic monitoring
• Focus on privacy, safety, and digital trust

Technology Keywords Edge AI, Intrusion Detection, IoT Security,
Cyber Threat Intelligence, Real-time Monitoring, Deep Learning,
CNN-LSTM

Data Availability Modified and publicly available on kaggle Link https:// doi �org/ 10
�34740/ kaggle/ dsv/ 11932656
Metadata (Type of • Tabular data (CSV format)
data) • Multi-class network traffic( benign + multiple attack types)

• Models trained on class-balanced and cleaned version of CICIoT2023
• CNN-LSTM and Transformer model used for classification
• Evaluating using standard train-test split
• Accuracy, F1-score, recall, and inference speed optimized

435