Privacy-enhancing data science platform Duality Technologies today announced that it raised $30 million in a series B round, bringing its total raised to $49 million to date. LG Technology Ventures led the funding with participation from Euclidean Capital, Intel Capital, Hearst Ventures, Team8, and the National Bank of Canada’s corporate venture capital arm, NAventures, which Duality says will be used to expand its go-to-market operations and advance its partnerships with technology vendors.
Newark, New Jersey-based Duality was founded in 2016 by Turing Award winner Shafi Goldwasser, MIT professor Vinod Vaikuntanathan, Kurt Rohloff, Alon Kaufman, and Rina Shainski. Vaikuntanathan is the coinventor of the foundational BGV Homomorphic Encryption scheme, while Rohloff cocreated the Palisade Homomorphic Encryption library, the open source package on which Duality’s platform is based.
“Duality was launched to enable organizations, both public and private, to collaborate on sensitive data without compromising on data privacy and business confidentiality,” Kaufman told VentureBeat via email. “In this age of big data, the need to collaborate on sensitive information continues to rise exponentially, while at the same time, data privacy continues to take center stage, leading to a growing range of global data privacy regulations (GDPR, CCPA, and others) designed to protect personally identifiable information. Reconciling these two conflicting mega-trends requires advanced technological solutions.”
Secure data platform
Duality’s platform, SecurePlus, protects data through a form of encryption called homomorphic encryption (HE). HE enables computation on file contents encrypted using an algorithm known as cryptonets. Cryptonets generate an encrypted result that, when decrypted, exactly matches the result of operations that would have been performed on unencrypted text.
By leveraging advances in HE, Duality claims it can enable organizations to collaborate on sensitive data without ever exposing the data.
Companies including Enveil, Inpher, Cosmian, Intel, Microsoft, and IBM are in the process of investigating HE. HE libraries don’t yet fully leverage modern hardware and are at least an order of magnitude slower than conventional methods, but newer projects like the accelerated encryption library cuHE claim speedups of 12 to 50 times on various encrypted tasks over previous approaches. HE-Transformer — a backend for nGraph, Intel’s neural network compiler — also delivers leading performance on some cryptonets.
“Due to the increasing level of regulatory attention being paid to data privacy, startups in the field are attracting more and more venture capital investment, while cloud computing giants … are also investing in research. However, it remains an extremely complex field, requiring deep expertise in advanced cryptographic techniques and data science and machine learning,” Kaufman said. “Duality is unique, and sets itself apart from its competitors, in that it implements a broad range of data science capabilities optimized for HE, providing a large number of use cases for secure data collaboration.”
Duality says it’s engaged with several organizations in the field of financial services, including Scotiabank, the Cyber Defence Alliance (a consortium of British and European banks and regulatory bodies), and others. The platform was originally piloted by Scotiabank, and in collaboration with CDA implemented a proof-of-concept inter-bank information sharing system. Duality has also partnered with the Tel Aviv Sourasky Medical Center to build privacy-preserving HE solutions for cancer research. And the U.S. Defense Advanced Research Projects Agency (DARPA) recently contracted Duality to develop privacy-preserving machine learning and training tools for COVID-19 research and response, as well as for cybercrime detection.
Other partners include Oracle’s Financial Crime and Compliance Management and the aforementioned Intel. In collaboration with Intel, Duality aims to optimize HE-based privacy-enhanced data science and AI applications for Intel’s hardware platforms.
“HE presents a paradigm shift in how machine learning is utilized by making it possible to run machine learning models on encrypted data without ever exposing personally identifiable information or other sensitive features. For example, in financial crime investigations, banks can use AI- or machine learning-powered data analysis to yield far-reaching conclusions about suspicious activity, boosting the efficiency of such investigations,” Kaufman said. “Moreover, Duality’s recent advances in hardware acceleration have expanded the range of data set sizes for which HE is feasible, making AI analysis on encrypted data sets even more effective and practical. In 2009, multiplying two bits homomorphically took half an hour. Today, that same task can be performed in just nanoseconds, and experts expect that figure to drop to a fraction of a nanosecond over the next few years.”
Duality currently has 35 employees and expects to expand its workforce in the coming months.
VentureBeatVentureBeat's mission is to be a digital town square for technical decision-makers to gain knowledge about transformative technology and transact. Our site delivers essential information on data technologies and strategies to guide you as you lead your organizations. We invite you to become a member of our community, to access:
- up-to-date information on the subjects of interest to you
- our newsletters
- gated thought-leader content and discounted access to our prized events, such as Transform 2021: Learn More
- networking features, and more