The recent developments in the field of privacy-preserving computation and secure multi-party computation have been marked by significant advancements in efficiency, security, and applicability. A notable trend is the enhancement of fully homomorphic encryption (FHE) schemes, which allow computations on encrypted data without needing to decrypt it first. These improvements are crucial for applications requiring high levels of privacy and security, such as in healthcare and finance. Another key area of progress is in federated learning (FL), where techniques are being developed to protect data privacy while enabling collaborative model training across decentralized devices. Innovations in this space focus on reducing computational overhead and enhancing security against both server-side and client-side threats, without compromising model accuracy. Additionally, advancements in asynchronous Byzantine agreement protocols are addressing the challenges of achieving consensus in distributed systems with potential malicious actors, offering solutions with improved communication and round complexity.

Noteworthy Papers

• Arbitrary-Threshold Fully Homomorphic Encryption with Lower Complexity: Introduces a novel AThFHE scheme that significantly reduces complexity, enabling efficient and secure computations with a large number of parties and data sizes.
• BlindFL: Segmented Federated Learning with Fully Homomorphic Encryption: Presents a framework that enhances the efficiency and security of FL by allowing clients to encrypt and send only a subset of their model updates, offering protection against both server-side and client-side attacks.
• OciorABA: Improved Error-Free Asynchronous Byzantine Agreement via Partial Vector Agreement: Proposes an ABA protocol with optimal resilience and reduced communication and round complexity, leveraging a new primitive called asynchronous partial vector agreement.
• A Selective Homomorphic Encryption Approach for Faster Privacy-Preserving Federated Learning: Describes a technique that combines selective encryption, homomorphic encryption, and differential privacy to achieve faster and secure FL, particularly effective in healthcare applications.