CONCEALING ALGORITHM LOGIC IN SECURE HOMOMORPHIC COMPUTATION

VAlue proposition

This technology, PrivaCT, offers a significant value proposition in the realm of data privacy and security. By leveraging fully homomorphic encryption (FHE), PrivaCT enables computations to be performed on encrypted data without the need for decryption, thereby preserving data confidentiality. What sets PrivaCT apart is its innovative approach to protecting not just the data but also the functions applied to it. Traditional FHE methods often evaluate functions in cleartext, which can expose proprietary algorithms and make them vulnerable to side-channel attacks. PrivaCT addresses this by transforming private internal circuits into uniform external circuits, thereby eliminating any information about the underlying function. By offering a scalable and efficient solution for secure computation, PrivaCT establishes a new benchmark for privacy-preserving computation, making it an invaluable tool for industries that require robust data and algorithm confidentiality.

 

Description of Technology

The core innovation of PrivaCT lies in its ability to transform private internal circuits into uniform external circuits, effectively concealing the specifics of the underlying functions through a two-stage approximation process. This process involves using a universal function approximator to capture the function's behavior and then decomposing it into scalar components, each approximated by univariate methods. This strategy not only maintains high accuracy but also ensures a constant runtime, making PrivaCT highly efficient. This makes PrivaCT particularly valuable for industries such as healthcare, where the confidentiality of proprietary algorithms and personal data is paramount.

 

Benefits

  • Enhanced Data Privacy
  • Secure Algorithm Confidentiality
  • High Accuracy and Efficiency
  • Scalability
  • Broad Applicability
  • Resistance to Side-Channel Attacks
  • Constant Runtime

 

Applications

  • Healthcare Data Analysis
  • Financial Services
  • Cloud Computing
  • Supply Chain Management
  • Scientific Research
  • Government and Defense

 

IP Status

Patent Pending

 

LICENSING RIGHTS AVAILABLE

Full licensing rights available

 

INVENTORs: Vishnu Boddeti and Amina Bassit

 

Tech ID: TEC2025-0139

 

For more information about this technology,
 

Patent Information: