Optimizing Privacy-Preserving Primitives to Support LLM-Scale Applications
This addresses the scalability problem for practitioners needing privacy-preserving learning systems, though it appears incremental as an overview of optimization efforts.
The paper tackles the computational and communication overhead that hinders practical adoption of privacy-preserving technologies at scale, showing progress through hardware/software/algorithm co-design to enable LLM-scale applications in contexts like DNN IP ownership and transformer inference.
Privacy-preserving technologies have introduced a paradigm shift that allows for realizable secure computing in real-world systems. The significant barrier to the practical adoption of these primitives is the computational and communication overhead that is incurred when applied at scale. In this paper, we present an overview of our efforts to bridge the gap between this overhead and practicality for privacy-preserving learning systems using multi-party computation (MPC), zero-knowledge proofs (ZKPs), and fully homomorphic encryption (FHE). Through meticulous hardware/software/algorithm co-design, we show progress towards enabling LLM-scale applications in privacy-preserving settings. We demonstrate the efficacy of our solutions in several contexts, including DNN IP ownership, ethical LLM usage enforcement, and transformer inference.