Cybersecurity & Defence

In the modern cybersecurity and defence landscape, strong machine learning models directly rely on access to data. Federated learning is a key technology to solve the data challenge where data cannot be shared between parties or moved off site.

The AI challenge

Common cases where data is out-of-reach in machine learning projects.
Don't want to share
Sensitive data kept private to protect privacy, ensure confidentiality, prevent misuse, and guard against cybersecurity threats.
Examples
  • Cyber incident reports
  • Operational security data
  • Intelligence gathering
Cannot share
Data that cannot be shared due to ownership rights, confidentiality agreements, or practical issues in transferring edge-generated data.
Examples
  • Proprietary defence technologies
  • Confidential intelligence briefings
  • Tactical edge data
Not allowed to share
Information that is restricted from being shared due to legal, regulatory, or compliance reasons to ensure data privacy and security.
Examples
  • Classified military communications
  • Export-controlled information
  • Personal data

Federated learning

Examples of how federated learning can enable or enhance AI in cybersecurity and defence.

AI at the edge

In intelligence gathering, surveillance, and reconnaissance, large amounts of data is gathered through sensors.

AI challenge

Edge processing poses significant challenges, in some cases due to the limited bandwidth and in some cases due to the amount of data generated.

Federated learning

Federated learning can significantly enhance data quality and model accuracy by:

  • Federated learning allows models to be trained directly where the data is generated.
  • By processing data locally, federated learning minimizes the need to transmit sensitive or large volumes of data.
  • The ability to train models on data from multiple sources directly on the edge helps overcome challenges associated with data scarcity.

Federated learning offers a promising solution by optimizing data processing and AI model training directly at the source of data collection.

Failure Prediction

Using AI for predictive maintenance and failure prediction to help forecast failures in vehicles and machinery.

AI Challenge

Accessing data for developing AI models faces several challenges in the military context:

  • Fragmented and inaccessible data across different systems
  • Security and privacy concerns complicate data sharing
Federated learning

Federated learning can improve data quality and model accuracy because:

  • Models can learn from the diverse and distributed data sets
  • Sensitive data can be processed locally, reducing the risk of compromising operational security
  • By training on data from multiple sources, federated learning can mitigate the issue of data scarcity

Intrusion detection

AI models can learn to recognize complex patterns and anomalies that may indicate a cyber intrusion.

AI Challenge

Involves analyzing heterogeneous and high-volume data streams. Challenging because:

  • Of the complexity and variability of network traffic
  • Difficult to keep confidentiality and integrity of sensitive information while using it to train AI models
Federated learning

Federated learning can enhance intrusion detection by:

  • Learning from diverse data sources across various network segments without centralizing sensitive data
  • Processing data locally, which is crucial for adhering to privacy regulations and safeguarding sensitive information

Project portfolio

A selection of our current public cybersecurity projects
IoT IDS
An advanced Intrusion Detection System (IDS) for IoT using federated learning, enhancing security and privacy by leveraging decentralised data analysis without compromising data privacy.
Learn more
LeakPro
A platform to evaluate the risk of information leakage in machine learning applications and to identify/validate realistic attack vectors.
Learn more
Secure Enclaves (TEE)
A solution for mitigating the challenge of protecting and ensuring trusted execution of machine learning on local clients using secure enclaves.
Learn more

Contact

For inquiries on enhancing AI capabilities with federated learning, please fill out the form below.
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.