Better Concrete Security for Half-Gates Garbling (in the Multi-instance Setting)

Guo, Chun; Katz, Jonathan; Wang, Xiao; Weng, Chenkai; Yu, Yu

doi:10.1007/978-3-030-56880-1_28

Cited by 18 publications

(9 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, DeepSecure reports the number of non-XOR gates that can be used for performance estimates. We used state-of-the-art for GC implementation, i.e., EMP-Toolkit [1], [52], [53], to obtain these performance estimates that are better than the performance reported by DeepSecure. The communication of our protocols is 25× lower (4 th row of Table I).…”

Section: Discussionmentioning

confidence: 99%

SIRNN: A Math Library for Secure RNN Inference

Rathee¹,

Rathee²,

Goli³

et al. 2021

Preprint

View full text Add to dashboard Cite

Complex machine learning (ML) inference algorithms like recurrent neural networks (RNNs) use standard functions from math libraries like exponentiation, sigmoid, tanh, and reciprocal of square root. Although prior work on secure 2party inference provides specialized protocols for convolutional neural networks (CNNs), existing secure implementations of these math operators rely on generic 2-party computation (2PC) protocols that suffer from high communication. We provide new specialized 2PC protocols for math functions that crucially rely on lookup-tables and mixed-bitwidths to address this performance overhead; our protocols for math functions communicate up to 423× less data than prior work. Some of the mixed bitwidth operations used by our math implementations are (zero and signed) extensions, different forms of truncations, multiplication of operands of mixed-bitwidths, and digit decomposition (a generalization of bit decomposition to larger digits). For each of these primitive operations, we construct specialized 2PC protocols that are more communication efficient than generic 2PC, and can be of independent interest. Furthermore, our math implementations are numerically precise, which ensures that the secure implementations preserve model accuracy of cleartext. We build on top of our novel protocols to build SIRNN, a library for end-to-end secure 2-party DNN inference, that provides the first secure implementations of an RNN operating on time series sensor data, an RNN operating on speech data, and a stateof-the-art ML architecture that combines CNNs and RNNs for identifying all heads present in images. Our evaluation shows that SIRNN achieves up to three orders of magnitude of performance improvement when compared to inference of these models using an existing state-of-the-art 2PC framework.Index Terms-privacy-preserving machine learning; secure two-party computation; recurrent neural networks; math functions; mixed-bitwidths; secure inference 1 We relegate comparisons with works that need additional parties for security, e.g., 3-party computation (3PC) to Section VII.

show abstract

Section: Discussionmentioning

confidence: 99%

SIRNN: A Math Library for Secure RNN Inference

Rathee¹,

Rathee²,

Goli³

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…Recent algorithmic optimizations are used to construct high-performance GCs [24,45]. More details can be found in [15,31,43].…”

Section: Private Nonlinear Computation (Gcs)mentioning

confidence: 99%

“…GCs: There are other prior works to accelerate GCs with GPU [9,19] and FPGA [7,17,18,39]. We note that prior work uses the less secure fixed-key GCs setup [15], or uses SHA-1 instead of AES, which is simpler and less secure [7]. Moreover, most prior work uses small benchmarks that do not stress off-chip bandwidth, which is one of HAAC's primary contributions.…”

Section: Related Workmentioning

confidence: 99%

Towards Fast and Scalable Private Inference

Garimella

Neda

et al. 2023

Proceedings of the 20th ACM International Conference on Computing Frontiers

View full text Add to dashboard Cite

Privacy and security have rapidly emerged as first order design constraints. Users now demand more protection over who can see their data (confidentiality) as well as how it is used (control). Here, existing cryptographic techniques for security fall short: they secure data when stored or communicated but must decrypt it for computation. Fortunately, a new paradigm of computing exists, which we refer to as privacy-preserving computation (PPC). Emerging PPC technologies can be leveraged for secure outsourced computation or to enable two parties to compute without revealing either users' secret data. Despite their phenomenal potential to revolutionize user protection in the digital age, the realization has been limited due to exorbitant computational, communication, and storage overheads.This paper reviews recent efforts on addressing various PPC overheads using private inference (PI) in neural network as a motivating application. First, the problem and various technologies, including homomorphic encryption (HE), secret sharing (SS), garbled circuits (GCs), and oblivious transfer (OT), are introduced. Next, a characterization of their overheads when used to implement PI is covered. The characterization motivates the need for both GCs and HE accelerators. Then two solutions are presented: HAAC for accelerating GCs and RPU for accelerating HE. To conclude, results and effects are shown with a discussion on what future work is needed to overcome the remaining overheads of PI. CCS CONCEPTS• Security and privacy → Privacy-preserving protocols; • Computer systems organization → Architectures.

show abstract

“…Thus, a key parameter is the expansion factor of the encryption. In garbled circuits using a 128-bit block cipher, including state-of-the-art optimizations (Point-and-Permute [2], Free XOR [47], Half Gates [90], and Fixed-Key Block Cipher [5,31]), each wire value is 16 bytes. Each wire represents only 1 bit of plaintext, so this is a 128× expansion factor.…”

Section: Analysis Of the Memory Demandmentioning

confidence: 99%

“…We implement oblivious transfer (OT) using multiple background threads. Concurrently with our work, EMP-toolkit was updated to use the MiTCCRH hash function [31]; our implementation is based on an older version of EMP-toolkit based on fixed-key AES [5]. When we compare MAGE to EMP-toolkit in §8, we use the older version of EMP-toolkit so the comparison is fair.…”

Section: Garbled Circuit Protocol Drivermentioning

confidence: 99%

MAGE: Nearly Zero-Cost Virtual Memory for Secure Computation

Kumar

Culler

Popa

2021

Preprint

View full text Add to dashboard Cite

Secure Computation (SC) is a family of cryptographic primitives for computing on encrypted data in single-party and multi-party settings. SC is being increasingly adopted by industry for a variety of applications. A significant obstacle to using SC for practical applications is the memory overhead of the underlying cryptography. We develop MAGE, an execution engine for SC that efficiently runs SC computations that do not fit in memory. We observe that, due to their intended security guarantees, SC schemes are inherently oblivioustheir memory access patterns are independent of the input data. Using this property, MAGE calculates the memory access pattern ahead of time and uses it to produce a memory management plan. This formulation of memory management, which we call memory programming, is a generalization of paging that allows MAGE to provide a highly efficient virtual memory abstraction for SC. MAGE outperforms the OS virtual memory system by up to an order of magnitude, and in many cases, runs SC computations that do not fit in memory at nearly the same speed as if the underlying machines had unbounded physical memory to fit the entire computation.

show abstract

Better Concrete Security for Half-Gates Garbling (in the Multi-instance Setting)

Cited by 18 publications

References 36 publications

SIRNN: A Math Library for Secure RNN Inference

SIRNN: A Math Library for Secure RNN Inference

Towards Fast and Scalable Private Inference

MAGE: Nearly Zero-Cost Virtual Memory for Secure Computation

Contact Info

Product

Resources

About