Timezone: »

Cryptographic Auditing for Collaborative Learning
Hidde Lycklama · Nicolas Küchler · Alexander Viand · Emanuel Opel · Lukas Burkhalter · Anwar Hithnawi

Collaborative machine learning paradigms based on secure multi-party computation have emerged as a compelling alternative for sensitive applications in the last few years. These paradigms promise to unlock the potential of important data silos that are currently hard to access and compute across due to privacy concerns and regulatory policies (e.g., health and financial sectors). Although collaborative machine learning provides many privacy benefits, it makes sacrifices in terms of robustness. It opens the learning process to the possibility of an active malicious participant who can covertly influence the model’s behavior. As these systems are being deployed for a range of sensitive applications, their robustness is increasingly important. To date, no compelling solution exists that fully addresses the robustness of secure collaborative learning paradigms. As the robustness of these learning paradigms remains an open challenge, it is necessary to augment these systems with measures that strengthen their reliability at deployment time. This paper describes our efforts in developing privacy-preserving auditing mechanisms for secure collaborative learning. We focus on audits that allow tracing the source of integrity issues back to the responsible party, providing a technical path towards accountability in these systems.

Author Information

Hidde Lycklama (ETH Zürich)
Nicolas Küchler (ETHZ)
Alexander Viand (ETH Zurich)
Alexander Viand

I am a doctoral student & research assistant in the Applied Cryptography Group at ETH Zürich and a member of the Privacy Preserving Systems Lab. I've received both my MSc and BSc in Computer Science from ETH Zürich. My interests include useable security and privacy, privacy enhancing technologies, and the interactions between these technologies and society. In my research, I work with secure computation technologies including Fully Homomorphic Encryption, Secure Multi-Party Computation and Zero-Knowledge Proofs, trying to make these techniques more accessible to non-experts by developing new systems, tools and abstractions.

Emanuel Opel (ETH Zurich)
Lukas Burkhalter (ETH Zurich)
Anwar Hithnawi (ETHZ - ETH Zurich)

More from the Same Authors