We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.CL

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Machine Learning

Title: Differentially Private Bias-Term only Fine-tuning of Foundation Models

Abstract: We study the problem of differentially private (DP) fine-tuning of large pre-trained models -- a recent privacy-preserving approach suitable for solving downstream tasks with sensitive data. Existing work has demonstrated that high accuracy is possible under strong privacy constraint, yet requires significant computational overhead or modifications to the network architecture.
We propose differentially private bias-term fine-tuning (DP-BiTFiT), which matches the state-of-the-art accuracy for DP algorithms and the efficiency of the standard BiTFiT. DP-BiTFiT is model agnostic (not modifying the network architecture), parameter efficient (only training about $0.1\%$ of the parameters), and computation efficient (almost removing the overhead caused by DP, in both the time and space complexity). On a wide range of tasks, DP-BiTFiT is $2\sim 30\times$ faster and uses $2\sim 8\times$ less memory than DP full fine-tuning, even faster than the standard full fine-tuning. This amazing efficiency enables us to conduct DP fine-tuning on language and vision tasks with long-sequence texts and high-resolution images, which were computationally difficult using existing methods.
Subjects: Machine Learning (cs.LG); Computation and Language (cs.CL); Cryptography and Security (cs.CR); Computer Vision and Pattern Recognition (cs.CV)
Cite as: arXiv:2210.00036 [cs.LG]
  (or arXiv:2210.00036v2 [cs.LG] for this version)

Submission history

From: Zhiqi Bu [view email]
[v1] Fri, 30 Sep 2022 18:30:48 GMT (240kb,D)
[v2] Tue, 4 Oct 2022 17:51:09 GMT (243kb,D)

Link back to: arXiv, form interface, contact.