The EPO published European patent application EP4062620A1 titled 'Systems and Methods for Efficient Computations on Split Data and Split Algorithms,' filed by TripleBlind Holdings, Inc. The patent covers computational techniques that enable multiple parties to perform machine learning operations on partitioned datasets without sharing underlying data. Key technology areas include split learning algorithms, privacy-preserving neural network computations, and distributed data processing methods with applications in AI training, financial analytics, and secure multi-party computations.

Organizations developing or deploying AI/ML systems that process sensitive or partitioned data across multiple entities should review this patent for potential licensing implications. Technology companies in fintech, healthcare analytics, and privacy-preserving computing should conduct freedom-to-operate analyses to assess whether their systems may require licensing. This is an informational patent publication; compliance teams should monitor for related patents in this technology space as the privacy-preserving AI sector evolves.