INL Innovation Spotlight

Innovative Data Concealment for Secure AI Research: The DIOD Methodology

The DIOD methodology offers a groundbreaking approach to share critical data for AI research, ensuring confidentiality while maintaining data utility.

Overview:      

In the era of big data, it is crucial to share information across platforms and organizations for innovation, especially in fields like AI research. However, the risk of sensitive data being reverse-engineered or compromised poses a significant challenge. Traditional data anonymization techniques often fall short, either by limiting data utility or failing to fully protect against data breaches.

The DIOD (Deceptive Infusion of Data) methodology emerges as a solution, particularly relevant for industries where data sharing is essential yet risky, such as defense, healthcare, and energy. Its market potential is vast, considering the increasing reliance on AI for materials discovery, energy optimization, and security.

Description:   

The DIOD methodology is an innovative approach to data sharing that successfully hides the identity of the system from which data originates, while still maintaining the functional dependencies required for AI research. It employs a non-invertible process to introduce deception into the data, ensuring the confidentiality of the original system's governing laws. Unlike traditional methods that can often degrade data quality or provide incomplete protection, DIOD preserves the crucial correlations needed for AI analysis. This enables researchers to utilize the data without jeopardizing the exposure of proprietary information.

Benefits:          

• Enhanced Security: Enables data sharing while protecting sensitive system information.
• Preserved Data Utility: Maintains crucial correlations and functional dependencies for AI research.
• Scalability: Provides an efficient solution for different data sizes and types.
• Compatibility: Applicable across various scientific and industrial sectors without compromising data integrity.
• Innovation in Anonymization: Represents a significant advancement beyond traditional data protection methods such as k-anonymity and encryption.

Applications:    

• Defense and Military: Facilitating secure sharing of data related to new technologies, while maintaining the confidentiality of critical information.
• Healthcare: Enabling the sharing of patient data for research purposes, while ensuring the full protection of personal information.
• Energy Sector: Facilitating the exchange of data on energy generation and storage innovations, while safeguarding proprietary processes.
• AI and Machine Learning Research: Providing benchmark datasets for the development and testing of AI algorithms, without any concerns regarding the origin of the data.

Development Status: 

Technology Readiness Level (TRL) 1: Basic principles observed and reported.

IP Status:        

Provisional Patent Filing No. 63/515,835, “Systems and Methods for Objective Management,” BEA Docket No. BA-1494.

INL Tech Partnerships: Your Gateway to Innovation

INL offers strategic access to proprietary technology, enhancing small business growth and contributing to economic and public advancement. We cater licensing terms to each business we work with, ensuring mutually beneficial agreements. Engage with our diverse technology offerings to propel your company forward.

Learn more about our licensing opportunities and the support we provide at https://inl.gov/technology-deployment/. For specific discussions on how your business can benefit, please contact Andrew Rankin at [email protected].

INL’s Technology Deployment department focuses exclusively on licensing intellectual property and partnering with industry collaborators capable of commercializing our innovations. Our goal is to commercialize the technologies developed by INL researchers. We do not engage in purchasing, manufacturing, procurement decisions, or providing funding. Additionally, this is not a call for external services to assist in the development of this technology.