CHET: Compiler and Runtime for Homomorphic Evaluation of Tensor Programs
This work addresses the problem of enabling secure, efficient computation on encrypted data for clients offloading to untrusted cloud providers, representing an incremental improvement by introducing a new interface to streamline FHE application development.
The paper tackles the challenge of developing practical applications for Fully Homomorphic Encryption (FHE) by proposing a Homomorphic Instruction Set Architecture (HISA) to decouple compiler optimizations from underlying FHE schemes, resulting in an end-to-end software stack that generates code faster than hand-optimized implementations for neural network evaluation on encrypted data.
Fully Homomorphic Encryption (FHE) refers to a set of encryption schemes that allow computations to be applied directly on encrypted data without requiring a secret key. This enables novel application scenarios where a client can safely offload storage and computation to a third-party cloud provider without having to trust the software and the hardware vendors with the decryption keys. Recent advances in both FHE schemes and implementations have moved such applications from theoretical possibilities into the realm of practicalities. This paper proposes a compact and well-reasoned interface called the Homomorphic Instruction Set Architecture (HISA) for developing FHE applications. Just as the hardware ISA interface enabled hardware advances to proceed independent of software advances in the compiler and language runtimes, HISA decouples compiler optimizations and runtimes for supporting FHE applications from advancements in the underlying FHE schemes. This paper demonstrates the capabilities of HISA by building an end-to-end software stack for evaluating neural network models on encrypted data. Our stack includes an end-to-end compiler, runtime, and a set of optimizations. Our approach shows generated code, on a set of popular neural network architectures, is faster than hand-optimized implementations.