Meeting in the Middle: A Co-Design Paradigm for FHE and AI Inference
This addresses privacy concerns for users and providers in cloud AI, but it is incremental as it builds on existing FHE and inference methods.
The paper tackles the privacy problem in cloud inference by proposing a co-design paradigm that specializes FHE schemes for inference circuits and constrains architectures to reduce costs, aiming to make FHE practical for AI applications.
Modern cloud inference creates a two sided privacy problem where users reveal sensitive inputs to providers, while providers must execute proprietary model weights inside potentially leaky execution environments. Fully homomorphic encryption (FHE) offers cryptographic guarantees but remains prohibitively expensive for modern architectures. We argue that progress requires co-design where specializing FHE schemes/compilers for the static structure of inference circuits, while simultaneously constraining inference architectures to reduce dominant homomorphic cost drivers. We outline a meet in the middle agenda and concrete optimization targets on both axes.