Vectorized secure evaluation of decision forests
Vectorized secure evaluation of decision forests
复制标题
决策森林的矢量化安全评估
DOI：
10.1145/3453483.3454094
复制
发表时间：
2021
期刊：
影响因子：
--
通讯作者：
Kulkarni, Milind
中科院分区：
文献类型：
--
作者：
Malik, Raghav;Singhal, Vidush;Gottfried, Benjamin;Kulkarni, Milind
随着对基于机器学习的推理的需求随着对隐私的担忧而增加，人们越来越认识到需要安全的机器学习，其中秘密模型可以用于对私有数据进行分类，而不会泄漏模型或数据。完全同态加密（FHE）允许对加密数据进行任意计算，为提供这种安全推断提供了一种有吸引力的方法。虽然这样的计算通常比明文计算慢几个数量级，但FHE密码系统对密文封装的能力（即加密明文的整个向量，以便在向量上按元素计算操作）有助于改善这种开销，有效地创建SIMD体系结构，其中可以对计算进行矢量化，以便更有效地进行计算。该领域的最新研究主要针对规则的、易于向量化的神经网络模型。将类似的技术应用于不规则的机器学习模型，如决策森林，由于其复杂的，难以矢量化的结构，仍然未被探索。在本文中，我们提出了COPSE，这是第一个利用密文打包来执行决策森林推理的系统。COPSE由一个分段编译器组成，该编译器可以自动地重组和编译决策林模型，将其简化为一组新的可向量化原语，以进行安全推理。我们发现，在一系列决策森林模型中，COPSE的编译模型的表现优于目前的技术水平，通常超过一个数量级，同时仍然可以很好地扩展。
As the demand for machine learning–based inference increases in tandem with concerns about privacy, there is a growing recognition of the need for secure machine learning, in which secret models can be used to classify private data without the model or data being leaked. Fully Homomorphic Encryption (FHE) allows arbitrary computation to be done over encrypted data, providing an attractive approach to providing such secure inference. While such computation is often orders of magnitude slower than its plaintext counterpart, the ability of FHE cryptosystems to dociphertext packing—that is, encrypting an entire vector of plaintexts such that operations are evaluated elementwise on the vector—helps ameliorate this overhead, effectively creating a SIMD architecture where computation can be vectorized for more efficient evaluation. Most recent research in this area has targeted regular, easily vectorizable neural network models. Applying similar techniques to irregular ML models such as decision forests remains unexplored, due to their complex, hard-to-vectorize structures.In this paper we present COPSE, the first system that exploits ciphertext packing to perform decision-forest inference. COPSE consists of a staging compiler that automatically restructures and compiles decision forest models down to a new set of vectorizable primitives for secure inference. We find that COPSE’s compiled models outperform the state of the art across a range of decision forest models, often by more than an order of magnitude, while still scaling well.
DOI：
10.1109/cgo.2013.6494989
发表时间：
2013-02
期刊：
Proceedings of the 2013 IEEE/ACM International Symposium on Code Generation and Optimization (CGO)
影响因子：
--
作者：
Bin Ren;G. Agrawal;J. Larus;Todd Mytkowicz;T. Poutanen;Wolfram Schulte
通讯作者：
Bin Ren;G. Agrawal;J. Larus;Todd Mytkowicz;T. Poutanen;Wolfram Schulte
ALCHEMY：一种轻松实现同态加密的语言和编译器
DOI：
--
发表时间：
2018
期刊：
Conference on Computer and Communications Security
影响因子：
--
作者：
E. Crockett;Chris Peikert;Chad Sharp
通讯作者：
Chad Sharp