{"id":5763,"date":"2024-11-25T14:00:27","date_gmt":"2024-11-25T06:00:27","guid":{"rendered":"https:\/\/changjinyuan.com\/?p=5763"},"modified":"2026-05-02T19:26:47","modified_gmt":"2026-05-02T11:26:47","slug":"%e5%88%98%e5%8f%b2%e6%af%93-luo-j-zhang-y-wang-h-yu-y-xu-z-2024-efficient-privacy-preserving-gaussian-process-via-secure-multi-party-computation-journal-of-systems-architecture-1-2","status":"publish","type":"post","link":"https:\/\/changjinyuan.com\/index.php\/publications\/publications-all\/5763\/","title":{"rendered":"Luo, J., Chen, G., Zhang, Y., \u5218\u53f2\u6bd3, Wang, H., Yu, Y., \u2026 &#038; Xu, Z. (2025). Centaur: bridging the impossible trinity of privacy, efficiency, and performance in privacy-preserving transformer inference. Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (ACL), 22751-22770."},"content":{"rendered":"<p>With the growing deployment of pre-trained models like Transformers on cloud platforms, privacy concerns about model parameters and inference data are intensifying. Existing Privacy-Preserving Transformer Inference (PPTI) frameworks face the \u201cimpossible trinity\u201d of balancing privacy, efficiency, and performance: Secure Multi-Party Computation (SMPC)-based approaches ensure strong privacy but suffer from high computational overhead and performance losses; Conversely, permutation-based methods achieve near-plaintext efficiency and accuracy but compromise privacy by exposing sensitive model parameters and intermediate results. Bridging this gap with a single approach presents substantial challenges, motivating the introduction of CENTAUR, a groundbreaking PPTI framework that seamlessly integrates random permutations and SMPC to address the \u201cimpossible trinity\u201d. By designing efficient PPTI algorithms tailored to the structural properties of Transformer models, CENTAUR achieves an unprecedented balance among privacy, efficiency, and performance. Our experiments demonstrate CENTAUR\u2019s ability to resist diverse data reconstruction attacks, achieve plaintext-level inference accuracy, and boost inference speed by 5.0\u223c30.4 times, unlocking new possibilities for secure and efficient AI deployment.<\/p>\r\n\r\n<div data-wp-interactive=\"core\/file\" class=\"wp-block-file\"><object data-wp-bind--hidden=\"!state.hasPdfPreview\" hidden class=\"wp-block-file__embed\" style=\"width: 100%; height: 600px;\" data=\"https:\/\/changjinyuan.com\/wp-content\/uploads\/2026\/05\/Centaur-bridging-the-impossible-trinity-of-privacy-efficiency-and-performance-in-privacy-preserving-transformer-inference.pdf\" type=\"application\/pdf\" width=\"300\" height=\"150\" aria-label=\"\u5d4c\u5165 Efficient privacy-preserving Gaussian process via secure multi-party computation\"><\/object><a id=\"wp-block-file--media-bea85d14-741f-478a-9ce3-b11f96ac753d\" href=\"https:\/\/changjinyuan.com\/wp-content\/uploads\/2026\/05\/Centaur-bridging-the-impossible-trinity-of-privacy-efficiency-and-performance-in-privacy-preserving-transformer-inference.pdf\">Efficient privacy-preserving Gaussian process via secure multi-party computation<\/a><a class=\"wp-block-file__button wp-element-button\" href=\"https:\/\/changjinyuan.com\/wp-content\/uploads\/2026\/05\/Centaur-bridging-the-impossible-trinity-of-privacy-efficiency-and-performance-in-privacy-preserving-transformer-inference.pdf\" download=\"\" aria-describedby=\"wp-block-file--media-bea85d14-741f-478a-9ce3-b11f96ac753d\">\u4e0b\u8f7d<\/a><\/div>\r\n","protected":false},"excerpt":{"rendered":"<p>Gaussian processes (GPs), known for their flexibility as non-parametric models, have been widely used inpractice involving sensitive data (e.g., healthcare, finance) from multiple sources. With the challenge of dataisolation and the need for high-performance models, how to jointly develop privacy-preserving GP for multipleparties has emerged as a crucial topic, In this paper, we propose a new privacy-preserving GP algorithm, namelyPP-GP, which employs secret sharing ($$) techniques, Specifically, we introduce a new ss-based exponentiationoperation (PP-Exp) through confusion correction and an SS-based matrix inversion operation (PP-Ml) basedon Cholesky decomposition. However, the advantages of the GP come with a great computational burden andspace cost. To further enhance the efficiency, we propose an efficient split learning framework for privacy.preserving GP, named Split-GP, which demonstrably improves performance on large-scale data. We leave theDrivate data-related and SMPC-hostile computations (i.., random features) on data holders, and delegate therest of SMPC-friendly computations (i.e., low-rank approximation, model construction, and prediction) to semihonest servers. The resulting algorithm significantly reduces computational and communication costs comparedto Pp-GPp, making it well-suited for application to large-scale datasets. We provide a theoretical analysis interms of the correctness and security of the proposed Ss-based operations. Extensive experiments show thatour methods can achieve competitive performance and efficiency under the premise of preserving privacy.<\/p>\n","protected":false},"author":1,"featured_media":5767,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":""},"categories":[15],"tags":[],"class_list":["post-5763","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-publications-all"],"acf":[],"lang":"cn","translations":{"cn":5763,"en":5773},"pll_sync_post":[],"_links":{"self":[{"href":"https:\/\/changjinyuan.com\/index.php\/wp-json\/wp\/v2\/posts\/5763","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/changjinyuan.com\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/changjinyuan.com\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/changjinyuan.com\/index.php\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/changjinyuan.com\/index.php\/wp-json\/wp\/v2\/comments?post=5763"}],"version-history":[{"count":5,"href":"https:\/\/changjinyuan.com\/index.php\/wp-json\/wp\/v2\/posts\/5763\/revisions"}],"predecessor-version":[{"id":5772,"href":"https:\/\/changjinyuan.com\/index.php\/wp-json\/wp\/v2\/posts\/5763\/revisions\/5772"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/changjinyuan.com\/index.php\/wp-json\/wp\/v2\/media\/5767"}],"wp:attachment":[{"href":"https:\/\/changjinyuan.com\/index.php\/wp-json\/wp\/v2\/media?parent=5763"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/changjinyuan.com\/index.php\/wp-json\/wp\/v2\/categories?post=5763"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/changjinyuan.com\/index.php\/wp-json\/wp\/v2\/tags?post=5763"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}