Skip to main content

Pioneering Practical Applications of Fully Homomorphic Encryption

A Revolutionary Proof of Concept

We are pleased to announce our groundbreaking collaboration with OCBC Bank in 2023, where we successfully applied our state-of-the-art Fully Homomorphic Encryption (FHE) techniques to pre-trained AI models in the cloud. This project marks a significant milestone in advancing the field of encrypted AI inferences, demonstrating a new standard in practical applications and solidifying our role as innovators in this domain.

The Successful Applications in the Project

Our journey with OCBC encompassed three key applications, each demonstrating the robustness and versatility of our FHE solutions:

  1. Speaker Verification on Encrypted Extracted Feature Vectors
  • Purpose: Confirming the identity of a speaker against a referenced voice.
  • Applications: Enhancing security in areas like entry control, privileged information access, voice banking, fraud detection, and more.
  • Results: Achieved a remarkable speed of 3ms per verification with a high accuracy rate of 99.2% on the Vox1-O test set.
  1. Facial Verification on Encrypted Extracted Feature Vectors:
  • Purpose: Verifying identities using facial recognition against stored photos.
  • Applications: Widely applicable in user identity verification, access control, automating know-your-customer (KYC) process with facial recognition, and other sensitive scenarios.
  • Results: Delivered verification in just 8ms with a 98.1% accuracy rate on the Kaggle-Avengers Face dataset.
  1. Predicting Default Payments on Encrypted Tabular Data:
  • Purpose: Assessing credit card payment defaults using encrypted data.
  • Applications: Vital for financial institutions to predict and manage credit risks, and can also be used for similar applications such as portfolio management by adjusting strategies according to credit risk, and using predictive analytics for regulatory compliance for risk management and reporting.
  • Results: Exceptionally fast predictions at 1ms with an 82.65% accuracy rate.

The Broader Impact: Beyond Banking and Finance

The success of this project with OCBC serves as a compelling testament to the practicality and efficiency of our solutions. The applications are vast and varied, extending well beyond banking and finance.

  • Enhanced Security: Our technology ensures robust security in identity verification processes, crucial for sectors like healthcare, government, and technology.
  • Data Privacy Compliance: With increasing concerns about data privacy, our HE techniques offer a compliant way for companies to leverage AI without compromising on data security.
  • Efficiency and Accuracy: The results from the OCBC project highlight that our solutions do not sacrifice speed or accuracy, making them suitable for real-world applications where time and precision are essential.

A Future-Ready Approach

Our collaboration with OCBC is not just a milestone but a beacon for future possibilities. We are excited about extending our technology to various sectors, including finance, technology, healthcare, and smart cities. Our approach ensures that sensitive data remains encrypted throughout the AI processing cycle, providing a groundbreaking solution for privacy-enhancing AI and analytics.

Explore How Your Organisation Can Benefit From Our Technology

To industries and organizations navigating the complex landscape of data privacy and analytics, we invite you to explore the possibilities with us. Whether you are looking to conduct your own proof of concept or seeking deeper insights into our technology, our team is ready to assist.

Contact us today to discover how our cutting-edge solutions can empower your business with privacy-compliant AI and analytics. Let’s embark on this journey towards a more secure and efficient future together.

This site is registered on wpml.org as a development site. Switch to a production site key to remove this banner.