Private Transformer Inference in MLaaS: A Survey

TL;DR

This survey reviews recent advances in Private Transformer Inference (PTI) techniques that enable privacy-preserving AI model deployment in MLaaS, emphasizing cryptographic methods, challenges, and evaluation frameworks.

Contribution

It introduces a structured taxonomy and evaluation framework for PTI, highlighting recent solutions and addressing the balance between privacy, efficiency, and high-performance inference.

Findings

Overview of cryptographic techniques like secure multi-party computation and homomorphic encryption.

Identification of key challenges in resource efficiency and privacy trade-offs.

Proposed evaluation framework for PTI solutions.

Abstract

Transformer models have revolutionized AI, powering applications like content generation and sentiment analysis. However, their deployment in Machine Learning as a Service (MLaaS) raises significant privacy concerns, primarily due to the centralized processing of sensitive user data. Private Transformer Inference (PTI) offers a solution by utilizing cryptographic techniques such as secure multi-party computation and homomorphic encryption, enabling inference while preserving both user data and model privacy. This paper reviews recent PTI advancements, highlighting state-of-the-art solutions and challenges. We also introduce a structured taxonomy and evaluation framework for PTI, focusing on balancing resource efficiency with privacy and bridging the gap between high-performance inference and data privacy.