An autonomous and Distributed Deep Learning Intrusion Detection System for Internet of Things (IoT)

Problem Setting

Massive attacks on vulnerable IoT devices cause data leakage or disruption of devices and systems.
Existing Intrusion Detection Systems (IDS) cannot detect zero-day attacks or introduce a high false alarm rate.
The highly dynamic attack landscape in IoT, the rapid growth and heterogeneity of IoT devices introduce new challenges for detecting attacks


DÏoT offers a unique network intrusion detection technology utilizing distributed deep learning-based algorithms for enabling IoT users to effortlessly and quickly protect their IoT devices and networks against attacks.

  • Autonomous and efficient system based on distributed deep learning algorithms
  • Zero-day attack detection, no false alarms
  • Superior performance compared to state-of-the-art technologies
  • Proven track record in IT and AI security research and successful industry collaborations of the heco-founders

Advanced Technology

  • Novel network traffic modelling techniques that can detect the potential malicious traffic at packet-level granularity.
  • Advanced AI algorithms boost accuracy and efficiency.
  • Optimized distributed deep learning scheme boosts model training process while preserving privacy of IoT user data.

Customer Benefits

  • Reliable, autonomous, and cost-effective solution to protect IoT devices and networks
  • No tedious and error-prone setup required
  • Privacy-preserving: Sensitive data of IoT users not shared with others

The DÏoT project recently received funding from the BMBF as part of the StartUpSecure program.