Information Flow Control in Machine Learning through Modular Model Architecture

Authors: 

Trishita Tiwari, Cornell University; Suchin Gururangan, University of Washington; Chuan Guo, FAIR at Meta; Weizhe Hua, Google DeepMind; Sanjay Kariyappa, Georgia Institute of Technology; Udit Gupta, Cornell University; Wenjie Xiong, Virginia Tech; Kiwan Maeng, Pennsylvania State University; Hsien-Hsin S. Lee, Intel; G. Edward Suh, NVIDIA/Cornell University

Abstract: 

In today's machine learning (ML) models, any part of the training data can affect the model output. This lack of control for information flow from training data to model output is a major obstacle in training models on sensitive data when access control only allows individual users to access a subset of data. To enable secure machine learning for access-controlled data, we propose the notion of information flow control for machine learning, and develop an extension to the Transformer language model architecture that strictly adheres to the IFC definition we propose. Our architecture controls information flow by limiting the influence of training data from each security domain to a single expert module, and only enables a subset of experts at inference time based on the access control policy. The evaluation using large text and code datasets show that our proposed parametric IFC architecture has minimal (1.9%) performance overhead and can significantly improve model accuracy (by 38% for the text dataset, and between 44%–62% for the code datasets) by enabling training on access-controlled data.

Open Access Media

USENIX is committed to Open Access to the research presented at our events. Papers and proceedings are freely available to everyone once the event begins. Any video, audio, and/or slides that are posted after the event are also free and open to everyone. Support USENIX and our commitment to Open Access.

BibTeX
@inproceedings {299770,
author = {Trishita Tiwari and Suchin Gururangan and Chuan Guo and Weizhe Hua and Sanjay Kariyappa and Udit Gupta and Wenjie Xiong and Kiwan Maeng and Hsien-Hsin S. Lee and G. Edward Suh},
title = {Information Flow Control in Machine Learning through Modular Model Architecture},
booktitle = {33rd USENIX Security Symposium (USENIX Security 24)},
year = {2024},
isbn = {978-1-939133-44-1},
address = {Philadelphia, PA},
pages = {6921--6938},
url = {https://www.usenix.org/conference/usenixsecurity24/presentation/tiwari},
publisher = {USENIX Association},
month = aug
}