Multiparty Computation with a Friend

Title of the Talk: Multiparty Computation with a Friend
Speakers: Dr. Protik Kumar Paul
Host Faculty: Dr.Maria Francis
Date: Dec 17, 2025
Time: 2:00 pm.
Venue: CSE-LH1

Abstract: Secure multiparty computation (MPC) enables collaborative computation over sensitive data but remains expensive and difficult to deploy in the dishonest-majority and asynchronous settings. Traditional dishonest-majority MPC suffers from high overheads and weak security guarantees such as lack of fairness, while asynchronous MPC (AMPC) protocols incur significant costs and require a two-thirds honest majority to ensure correctness and liveness. We explore the MPC with a friend (helper-aided) model, which introduces a semi-honest, non-colluding helper party (HP)—a realistic assumption in many real-world applications such as financial dark pools. Leveraging this model, we design Asterisk, a practically efficient MPC framework that achieves fairness even with a malicious majority, invokes the HP only a constant number of times, and scales to hundreds of parties. Asterisk outperforms all existing dishonest-majority MPC protocols and is competitive with honest-majority systems, achieving 228–288× faster preprocessing and evaluating circuits with 10⁶ multiplications among 100 parties in ~20 seconds. We also implement fast and scalable dark pool applications using Asterisk. Additionally, we develop new asynchronous helper-aided MPC protocols (Pollux and Castor) under the preprocessing–online paradigm. Our construction, Pollux, uses only lightweight cryptography in the honest-majority setting and shows that oblivious transfer during preprocessing is necessary in the dishonest-majority case. Our experiments demonstrate substantial performance and robustness improvements over prior helper-aided MPC and AMPC schemes, particularly under network delays.

Speaker Bio: Protik Kumar Paul is a postdoctoral researcher at TU Darmstadt, Germany, specializing in cryptography and privacy-preserving technologies. He completed his PhD from IISc Bangalore. His research is focused on Secure Multiparty Computation (MPC), Zero-Knowledge (ZK) proofs, and privacy-preserving frameworks for real-world applications. Protik has contributed to advancing both theory and practice, including designing efficient MPC protocols in challenging adversarial settings and developing privacy-preserving solutions for financial systems, secure machine learning, and digital services. He has published in leading venues such as IEEE S&P, PoPETS and Asiacrypt, and his work focuses on bridging foundational cryptographic research with practical, scalable applications.