Multiparty Computation (MPC) technology plays a central role in enabling the future of advertising at Facebook. Through MPC and other Privacy Enhancing Technologies, users’ data can remain private while being used to create the best advertising experience for businesses.

Featuring:
Benjamin Case, Research Scientist, Facebook

Topic Category: Privacy

Within the fractured value chains of the data-driven marketing and advertising spaces, companies engage in data transactions on a daily basis. Even when a firm is evaluating another vendor’s product or service, a transfer of sensitive data about consumers may be involved. Conducting these transactions in a safe and compliant manner can require lengthy legal negotiations and use of complex data governance platforms. Secure MPC eliminates these issues at the root: only encrypted data is ever transferred, and no one except the original data owner has a decryption key. We review a real-world use case in which MPC simultaneously secures and accelerates workflows in this way.

Featuring:
Shereen Shermak, CEO, Nth Party, Ltd.
Frederick Jansen, CTO, Nth Party, Ltd.

Topic Category: Privacy

With the explosion and proliferation of data that comes from new digital initiatives and increased mobility, most organizations are expanding their use of encryption. And for the first time protecting customer data has become the number one priority. There is an expression often used in crypto “Not your private keys, not your bitcoin”. In fact, we can use the same expression for cloud services “Not your encryption keys, not your data” clearly leads to a trust issue. Microsoft introduced support for Double Key Encryption (DKE) of Office 365 earlier this year. The objective was to make it easy for enterprise customers to provide an extra level of security over critical content stored in Azure by supporting the second level of encryption with a customer-controlled key. For many enterprises, the cost and complexity of procuring and operating their own key management system and hardware security modules (HSMs) to protect the customer-controlled keys were beyond their capacity. DuoKey resolves this challenge with its DuoKey Key Service based on MPC, we will showcase how organization leverages on MPC to migrate to Microsoft 365.

Featuring:
Nagib Aouini, CEO and founder of DuoKey

Topic Category: Security

Fighting financial crimes and delivering faster transactions has actively pushed the data science teams to explore innovative ways to detect fraud accurately. While ‘bad actors’ today have access to the same technology as the financial organizations, they don’t have the scale of training data that these orgs do. The datasets are, however, locked away from sharing due to security and compliance reasons. Using Secure Multi-Party Computation (MPC), data scientists at Inpher and BNY Mellon have built a novel collaborative fraud detection framework, where one can train the AI models without revealing partners’ data. This talk shares the solution framework and presents our initial findings, including up to 20% better fraud predictions due to access to more ‘features’ in the data.

Featuring:
Kevin McCarthy, Vice President at Inpher

Topic Category: Privacy

The presentation outline a public-private project that aim at enhancing the use of health data for improved decision making across all aspects of health from preventing diseases to developing and testing treatments. The core part of the project is an MPC based infrastructure that allows for commercial use of Danish register data as well as patient generated behavior data.

The technical solution is a GDPR compliant IT infrastructure for Health Data Exchange that both combine independent registers as well as facilitating sharing of health data through individual consent. The solution consists of two separate MPC based systems – the VPR and DFG Platforms:

The Virtual Public Register (VPR) Platform is a system that facilitates confidential collaboration between independent public registers i.e. a system to virtually combine and uses public registers. The governance of the VPR Platform is linked directly to the technical platform based on MPC. The initial operational model consists of two public registers; Statistics Denmark and the Danish Health Data. The MPC model is a two-party protocol, where the registers themself run a MPC node each.

The Data for Good Platform (DFG) Platform is a citizen-centric system that both facilitates decentralized control of personal data and orchestrates privacy-preserving use of personal data. MPC is the secure infrastructure used to combine data and produce the anonymous output. The infrastructure is governed by the not for profit Data for Good Foundation and the Computing Parties that run MPC. The MPC protocols are orchestrated via a blockchain infrastructure that ensures a flexible and robust operation.

In combination, the VPR and the DFG Platform constitute a generic public-private infrastructure designed for privacy-preserving health data exchange.

The project was initially matured through a public funded R&D project, which has turned into a larger public-private collaboration and currently funded with around USD 18,5 millions.

Featuring:
Kurt Nielsen, CEO, Partisia

Topic Category: Privacy

Learn how CYBAVO uses MPC as a core technology to generate and protect crypto wallets. MPC’s key-share approach to key generation resolves a main issue found in existing private key management systems: it removes the existence of a single-private key. This, in turn, removes the single-point of failure that has led to so many of the multi-million dollar breaches in blockchain’s short history. CYBAVO applies MPC technology, along with a triple-lock private key protection model, to help enterprises enhance the protection and security of private keys and digital transactions.

Featuring:
Jonathan Cowling, Sales Director at CYBAVO

Topic Category: Security

Financial crime is a huge, world-wide problem. In 2018, an estimated 5.8 trillion dollars (6.7% of global GDP) worth of financial crime was perpetrated. Therefore, collaboration and data sharing is key in detecting financial crime and specifically money laundering. However, legislation such as GDPR and competition law regulate and can restrict data sharing. Fortunately, privacy-enhancing technologies can enable such a collaboration without having to share sensitive data.

Banks have data associated with money laundering for their customers (e.g. transaction networks), for example based on receival of large amounts of cash, cryptocurrencies or money from or to certain high-risk countries. The value of these data is often limited as long as they are based on only the local network of a single bank. Criminals try to obfuscate money laundering flows and often will not perform such an act within one bank. By using MPC, information about bank accounts and/or clients can securely be “propagated” on the entire transaction network of participating banks while preserving privacy. This way, more information could be gathered on criminal flows in the network.

TNO, in collaboration with ABN AMRO and Rabobank, built a Proof of Concept of the so-called secure risk propagation using synthetic data that includes money laundering patterns. We have shown that it is possible to securely perform the risk propagation algorithm among different banks and that this is scalable. This demonstrates the value of collaborative transaction analysis.

Featuring:
Marie Beth van Egmond, Scientist, TNO

Topic Category: Privacy

Real-world data science pipelines are 90% data cleaning and preparation and 10% the fun stuff – whether statistics, data mining or machine learning. Rmind allows the user to run the whole pipeline with the end-to-end privacy using the Sharemind MPC platform.

The data owners will secret-share/encrypt data on premises and upload their tables for use by Rmind. Rmind allows the tables to pre-processed, exceptional data filtered, linked, sorted or aggregated, in any order. And then, you can go on to apply machine learning, statistics or custom algorithms, e.g., for fraud detection. All results can be reused in further processing.

At all stages, Rmind guarantees cryptographic privacy for all input, intermediate and output data. It also prevents a curious analyst from accessing raw records through configurable controls. The 5-minute pitch will show a demonstration of the tool in a credit scoring context.

Featuring:
Dr. Dan Bogdanov, Head of R&D, Cybernetica

Topic Category: Privacy

The transition towards more sustainable energy leads to challenges for grid operators. Electric vehicle chargers, solar panels and windmills all have an impact on the available capacity of the grid and can lead to congestion. Smart energy meters have the potential to provide insights to accelerate congestion resolution. However, smart meter data is very sensitive and is subject to strict regulation.

MPC offers a solution that can be applied within existing regulations, that enables the grid operators to use this data for network optimization. With this solution, an electricity grid operator can privately aggregate data over multiple meters, without the grid operator being able to see the individual meter data.

When meter data has been aggregated in this way, the grid operator may use the data in accordance with the Electricity Act to further analyse the local performance of the grid. This enables the operator to manage the network more precisely and plan investments more closely.

Together, with a large Dutch grid operator, the technology has been tested in a laboratory environment and now the solution is being tested live on 200 households.

Featuring:
Niek Bouman, CTO & Founder, Roseman Labs

Topic Category: Privacy