Data is an organization’s most precious resource. It’s used in countless numbers of ways, from making product decisions to understanding customer needs. Yet, data remains largely underutilized, often restricted by privacy concerns. A major conflict remains between gaining utility from a data set and protecting the privacy of individuals or entities. Overcoming this conflict by creating a data collaboration strategy – when possible – has the potential to bring exponential impact to both organizations and their customers. For example, medical facilities can accelerate discovery, personalize healthcare, and improve health outcomes. Governments can perform confidential investigations and inter-agency collaborations.
The Duality Platform was built to solve this conflict, making collaborative data analytics safe and utile. It allows collaborative projects to be quickly and easily set up in any environment. It coalesces cutting-edge cryptography and data science techniques, expertly enabling secure data collaboration. Read below to understand how.
Each organization has a different need when it comes to data collaboration. The Duality Platform covers a wide range of secured data sharing to enable many forms of collaboration and analysis. We break these options down into three main models.
The first deals with analytics of a single data set which enables organizations to provide internal or third-party users the option to securely analyze data remotely. The second addresses analytics of union datasets (aka horizontal data analysis), enabling analysis on disparate data sets that share the same schema. Finally, the Duality Platform supports analytics of join datasets (aka vertical data analysis) which lets organizations link between different datasets with different schemas using a share join key.
With these three options for dataset analytics, an organization can do analysis on any connected or separate datasets that they please.
The Duality Platform offers 3 models for collaboration enabling an organization to secure sensitive data while still allowing different types of sharing:
Each of these models is useful for different types of data collaboration, and allows flexibility for our users.
The Platform supports multiple PET technologies and adjusts the right technology to the various use cases. The computation supports machine learning for both training and inference as well as SQL queries, and statistics. A new capability that was recently added to the Platform is secure federated learning. Duality is using a federated learning approach with an additional layer of security that makes it more secure than a typical federated learning framework.
To help users organize and manage their data in a secured collaboration, the Duality Platform offers a wide set of capabilities that simplify the data collaboration between multiple parties.. We now support schema collaboration and schema mapping, to aid in connecting datasets for joint analysis. Also, we help users in data preprocessing which is an essential step for any data science workflow. The Platform also helps users find matches between datasets using an advanced entity resolution capabilities allowing various parties to find join data for collaboration. Duality does all of this while making sure that the data being stored is consistent, trustworthy, and doesn’t get misused.
With Duality, organizations can start sharing data and getting new insight while keeping their sensitive data protected and applying a wide range of computations including complex data science models in a way that wasn’t possible in the past. In a nutshell, Duality makes the complexity of privacy preserving data collaboration seamless, allowing enterprises to harness the full value of their data.
Read the full white paper: “Building a Privacy Enabled Data Collaboration Strategy.”
