AI-Driven Smart Communication

Tekdemir A., Kuğu E.

10th International Conference on Computer Science and Engineering, UBMK 2025, İstanbul, Turkey, 17 - 21 September 2025, pp.1461-1466, (Full Text) identifier

  • Publication Type: Conference Paper / Full Text
  • Doi Number: 10.1109/ubmk67458.2025.11207022
  • City: İstanbul
  • Country: Turkey
  • Page Numbers: pp.1461-1466
  • Keywords: Anomaly Detection, Artificial Intelligence, Autoencoder, Deep Learning, LSTM, Network Management, Q-Learning, Traffic Prediction
  • TED University Affiliated: Yes

Abstract

The growing volume of data traffic and increasing complexity of network infrastructures have made network management more challenging. This study explores the effectiveness of artificial intelligence techniques in managing networks using the UNSW-NB15 dataset. Deep learning models (LSTM and Autoencoder), anomaly detection algorithms, and reinforcement learning methods (Q-Learning) were applied to traffic prediction, anomaly detection, routing, and network security. The performance of these methods was thoroughly analyzed. Results show that AI-based approaches offer clear advantages over traditional techniques by delivering high accuracy and efficiency. These findings emphasize the vital role AI-driven systems will play in future network infrastructures.