The field of secure and private computing is moving towards the development of more efficient and scalable solutions for privacy-preserving data processing and analysis. Researchers are exploring new architectures and frameworks that enable secure and private computation, such as zero-knowledge extensions and verifiable split learning. These innovations have the potential to transform the way data is processed and analyzed, enabling more secure and private collaboration and computation. Noteworthy papers in this area include:

• CryptoMoE, which proposes a novel framework for private and efficient inference for mixture-of-experts architectures, achieving significant latency and communication reductions.
• Verifiable Split Learning via zk-SNARKs, which integrates zero-knowledge proofs to ensure correctness and verifiability in split learning.
• Private Map-Secure Reduce, which introduces a network-native paradigm for efficient AI data markets, enabling verifiable privacy, efficient price discovery, and incentive alignment.