Wen-jie Lu (Ant Group), Zhicong Huang (Ant Group), Zhen Gu (Alibaba Group), Jingyu Li (Ant Group & Zhejiang University), Jian Liu (Zhejiang University), Cheng Hong (Ant Group), Kui Ren (Zhejiang University), Tao Wei (Ant Group), WenGuang Chen (Ant Group)

Large transformer-based models have realized state-of-the-art performance on lots of real-world tasks such as natural language processing and computer vision.
However, with the increasing sensitivity of the data and tasks they handle, privacy has become a major concern during model deployment.
In this work, we focus on private inference in two-party settings, where one party holds private inputs and the other holds the model.
We introduce BumbleBee, a fast and communication-friendly two-party private transformer inference system.
Our contributions are three-fold:
First, we propose optimized protocols for matrix multiplication, which significantly reduce communication costs by 80% -- 90% compared to previous techniques.
Secondly, we develop a methodology for constructing efficient protocols tailored to the non-linear activation functions employed in transformer models.
The proposed activation protocols have realized a significant enhancement in processing speed, alongside a remarkable reduction in communication costs by 80% -- 95% compared with two prior methods.
Lastly, we have performed extensive benchmarks on five transformer models.
BumbleBee demonstrates its capability by evaluating the LLaMA-7B model, generating one token in approximately 8 minutes using CPUs.
Our results further reveal that BumbleBee outperforms Iron (NeurIPS22) by over an order of magnitude and is three times faster than BOLT (Oakland24) with one-tenth communication.

View More Papers

LADDER: Multi-Objective Backdoor Attack via Evolutionary Algorithm

Dazhuang Liu (Delft University of Technology), Yanqi Qiao (Delft University of Technology), Rui Wang (Delft University of Technology), Kaitai Liang (Delft University of Technology), Georgios Smaragdakis (Delft University of Technology)

Read More

RAIFLE: Reconstruction Attacks on Interaction-based Federated Learning with Adversarial...

Dzung Pham (University of Massachusetts Amherst), Shreyas Kulkarni (University of Massachusetts Amherst), Amir Houmansadr (University of Massachusetts Amherst)

Read More

L-HAWK: A Controllable Physical Adversarial Patch Against a Long-Distance...

Taifeng Liu (Xidian University), Yang Liu (Xidian University), Zhuo Ma (Xidian University), Tong Yang (Peking University), Xinjing Liu (Xidian University), Teng Li (Xidian University), Jianfeng Ma (Xidian University)

Read More

Rethink Custom Transformers for Binary Analysis

Heng Yin, Professor, Department of Computer Science and Engineering, University of California, Riverside

Read More