Page 473 - AI for Good Innovate for Impact

AI for Good Innovate for Impact

2�3 Future Work

1. Data Collection, to make it more general and work for most digital devices we may need 4.4-Productivity
gathering more data not only in amount but also from different devices and threat types.
2. Model Development, this would enable us to build and fine tune a more general and
novel detection system.
3. Model development and training.
With our model we are trying to detect and identify the threat type as well

3 Use Case Requirements

REQ-01: Real-time Edge Monitoring

It is critical that the system performs real-time traffic inspection at the edge level (like routers,
IoT gateways) to immediately detect suspicious activity.

REQ-02: Cloud-Based Analytics and Response

It is critical that flagged or suspicious traffic is automatically forwarded to the cloud for deep
analysis, logging, and correlation with known threat intelligence databases.

REQ-03: Accuracy and Continuous Learning

It is expected that the system maintains high accuracy and low false positives by continuously
retraining its AI models using updated, diverse datasets.

REQ-04: Privacy-Aware Traffic Monitoring

It is critical that the system ensures data privacy and user consent, using a subscription-based
or contract-based model to authorize traffic inspection.
REQ-05: Scalability and Deployment Flexibility

It is expected that the solution can be deployed in various environments (e.g., smart homes,
offices, healthcare, manufacturing) with minimal configuration overhead.

REQ-06: Performance and Inference Speed

It is critical that the IDS provides fast inference times with minimal latency, allowing seamless
operation without impacting network performance.

437