Periodic Reporting for period 4 - SPEC (Secure, Private, Efficient Multiparty Computation)
Reporting period: 2023-07-01 to 2024-10-31
The SPEC project has successfully developed innovative techniques and methods for secure multiparty computation (MPC) protocols with enhanced security, privacy, and efficiency. This achievement represents a significant step forward in the field, addressing critical limitations in previous MPC approaches and paving the way for more practical and impactful solutions.
MPC is a cryptographic technology that allows groups of individuals or organizations to compute shared results from their private data without revealing the data itself. This has wide-ranging applications, including secure auctions, privacy-preserving data analysis, and collaborative research. Over the past decade, MPC efficiency has improved considerably, but existing solutions still had several shortcomings.
The SPEC project addressed these challenge by rethinking the “MPC Stack”—the layers comprising the system, cryptographic, and application components. By refining theoretical models and integrating insights from other areas of computer science, the project achieved several key results:
1. Real-World Security: The project identified and addressed limitations in existing protocols and models for MPC and other cryptographic primitives, discovering vulnerabilities and designing frameworks that better reflect real-world requirements for security, privacy, and efficiency.
2. Next-Generation Protocols: Using these refined models, the team developed advanced MPC protocols that overcome existing performance barriers, thanks to the development of new and unexpected mathematical tools.
3. Balancing Privacy and Utility: The project explored the trade-offs participants face between privacy and usability when sharing data. This led to the design of MPC functionalities that encourage rational cooperation, as well as investigating the interaction between MPC and output privacy, which is a very timely topic also due to advances in machine-learning.
The outcomes of the SPEC project are not only academically significant but also have potential for real world impact. By enhancing privacy-preserving computation, the project supports secure and trustworthy data-sharing solutions. Startups in the field are already beginning to exploit the results of the project towards enhancing privacy and security, with potential applications to sectors such as healthcare, finance, and public administration. These advances have the potential to help protect individual and organizational privacy while fostering collaboration on a global scale.
MPC is a cryptographic technology that allows groups of individuals or organizations to compute shared results from their private data without revealing the data itself. This has wide-ranging applications, including secure auctions, privacy-preserving data analysis, and collaborative research. Over the past decade, MPC efficiency has improved considerably, but existing solutions still had several shortcomings.
The SPEC project addressed these challenge by rethinking the “MPC Stack”—the layers comprising the system, cryptographic, and application components. By refining theoretical models and integrating insights from other areas of computer science, the project achieved several key results:
1. Real-World Security: The project identified and addressed limitations in existing protocols and models for MPC and other cryptographic primitives, discovering vulnerabilities and designing frameworks that better reflect real-world requirements for security, privacy, and efficiency.
2. Next-Generation Protocols: Using these refined models, the team developed advanced MPC protocols that overcome existing performance barriers, thanks to the development of new and unexpected mathematical tools.
3. Balancing Privacy and Utility: The project explored the trade-offs participants face between privacy and usability when sharing data. This led to the design of MPC functionalities that encourage rational cooperation, as well as investigating the interaction between MPC and output privacy, which is a very timely topic also due to advances in machine-learning.
The outcomes of the SPEC project are not only academically significant but also have potential for real world impact. By enhancing privacy-preserving computation, the project supports secure and trustworthy data-sharing solutions. Startups in the field are already beginning to exploit the results of the project towards enhancing privacy and security, with potential applications to sectors such as healthcare, finance, and public administration. These advances have the potential to help protect individual and organizational privacy while fostering collaboration on a global scale.
From the beginning of the project, the SPEC team focused on advancing secure multiparty computation (MPC) by developing novel techniques to enhance the security, privacy, and efficiency of cryptographic protocols. Key achievements include breakthroughs in real-world security, such as identifying the LadderLeak vulnerability in widely-used signature schemes, and advancing the theoretical foundation of MPC with the discovery of efficient methods for distributed discrete logarithms (EUROCRYPT 2021). This foundational work enabled the development of homomorphic secret-sharing techniques and innovative cryptographic protocols, which have been widely recognized and cited in the field. Additionally, the project contributed to exploring the privacy-utility trade-offs in MPC, resulting in practical models for incentivizing collaboration and secure distributed machine learning frameworks (WWW 2024).
The dissemination of results has been impactful, with publications in top-tier venues, collaborations with industry leaders like Partisia, and contributions to international standardization efforts, such as FAEST for quantum-resistant digital signatures. While some research avenues did not yield the expected results, others led to unanticipated discoveries that significantly advanced the state of the art. These achievements demonstrate the project’s success in addressing the limitations of existing MPC models and providing tools that are not only academically significant but also have real-world applications across sectors like privacy-preserving data analysis and secure computation.
The dissemination of results has been impactful, with publications in top-tier venues, collaborations with industry leaders like Partisia, and contributions to international standardization efforts, such as FAEST for quantum-resistant digital signatures. While some research avenues did not yield the expected results, others led to unanticipated discoveries that significantly advanced the state of the art. These achievements demonstrate the project’s success in addressing the limitations of existing MPC models and providing tools that are not only academically significant but also have real-world applications across sectors like privacy-preserving data analysis and secure computation.
The SPEC project has pushed the boundaries of secure multiparty computation (MPC) by introducing innovative tools and methodologies that address critical limitations in existing protocols. Notable progress includes the development of succinct, universally composable SNARKs, demonstrating for the first time the feasibility of achieving the highest security standards in concurrent environments without sacrificing (asymptotic) efficiency (EUROCRYPT 2023). Additionally, the project advanced the field of arithmetic garbling, achieving unprecedented efficiency and for computations expressed as arithmetic circuits. These results open new avenues for secure computation in novel application domains. As the project concludes, its contributions are expected to to play an important role for the advancements of cryptographic research, with follow-up work already underway by the research community to build on these novel techniques.