Showing papers by "Michael Naehrig published in 2016"

CryptoNets: applying neural networks to encrypted data with high throughput and accuracy

Abstract: Applying machine learning to a problem which involves medical, financial, or other types of sensitive data, not only requires accurate predictions but also careful attention to maintaining data privacy and security. Legal and ethical requirements may prevent the use of cloud-based machine learning solutions for such tasks. In this work, we will present a method to convert learned neural networks to CryptoNets, neural networks that can be applied to encrypted data. This allows a data owner to send their data in an encrypted form to a cloud service that hosts the network. The encryption ensures that the data remains confidential since the cloud does not have access to the keys needed to decrypt it. Nevertheless, we will show that the cloud service is capable of applying the neural network to the encrypted data to make encrypted predictions, and also return them in encrypted form. These encrypted predictions can be sent back to the owner of the secret key who can decrypt them. Therefore, the cloud service does not gain any information about the raw data nor about the prediction it made. We demonstrate CryptoNets on the MNIST optical character recognition tasks. CryptoNets achieve 99% accuracy and can make around 59000 predictions per hour on a single PC. Therefore, they allow high throughput, accurate, and private predictions.

Frodo: Take off the Ring! Practical, Quantum-Secure Key Exchange from LWE

Abstract: Lattice-based cryptography offers some of the most attractive primitives believed to be resistant to quantum computers. Following increasing interest from both companies and government agencies in building quantum computers, a number of works have proposed instantiations of practical post-quantum key exchange protocols based on hard problems in ideal lattices, mainly based on the Ring Learning With Errors (R-LWE) problem. While ideal lattices facilitate major efficiency and storage benefits over their non-ideal counterparts, the additional ring structure that enables these advantages also raises concerns about the assumed difficulty of the underlying problems. Thus, a question of significant interest to cryptographers, and especially to those currently placing bets on primitives that will withstand quantum adversaries, is how much of an advantage the additional ring structure actually gives in practice. Despite conventional wisdom that generic lattices might be too slow and unwieldy, we demonstrate that LWE-based key exchange is quite practical: our constant time implementation requires around 1.3ms computation time for each party; compared to the recent NewHope R-LWE scheme, communication sizes increase by a factor of 4.7x, but remain under 12 KiB in each direction. Our protocol is competitive when used for serving web pages over TLS; when partnered with ECDSA signatures, latencies increase by less than a factor of 1.6x, and (even under heavy load) server throughput only decreases by factors of 1.5x and 1.2x when serving typical 1 KiB and 100 KiB pages, respectively. To achieve these practical results, our protocol takes advantage of several innovations. These include techniques to optimize communication bandwidth, dynamic generation of public parameters (which also offers additional security against backdoors), carefully chosen error distributions, and tight security parameters.

Efficient Algorithms for Supersingular Isogeny Diffie-Hellman

Abstract: We propose a new suite of algorithms that significantly improve the performance of supersingular isogeny Diffie-Hellman SIDH key exchange. Subsequently, we present a full-fledged implementation of SIDH that is geared towards the 128-bit quantum and 192-bit classical security levels. Our library is the first constant-time SIDH implementation and is upi¾?to 2.9 times faster than the previous best non-constant-time SIDH software. The high speeds in this paper are driven by compact, inversion-free point and isogeny arithmetic and fast SIDH-tailored field arithmetic: on an Intel Haswell processor, generating ephemeral public keys takes 46 million cycles for Alice and 52 million cycles for Bob, while computing the shared secret takes 44 million and 50 million cycles, respectively. The size of public keys is only 564 bytes, which is significantly smaller than most of the popular post-quantum key exchange alternatives. Ultimately, the size and speed of our software illustrates the strong potential of SIDH as a post-quantum key exchange candidate and we hope that these results encourage a wider cryptanalytic effort.

Privately Evaluating Decision Trees and Random Forests

Abstract: Decision trees and random forests are com- mon classifiers with widespread use. In this paper, we develop two protocols for privately evaluating decision trees and random forests. We operate in the standard two-party setting where the server holds a model (ei- ther a tree or a forest), and the client holds an input (a feature vector). At the conclusion of the protocol, the client learns only the model's output on its input and a few generic parameters concerning the model; the server learns nothing. The first protocol we develop provides security against semi-honest adversaries. We then give an extension of the semi-honest protocol that is robust against malicious adversaries. We implement both pro- tocols and show that both variants are able to process trees with several hundred decision nodes in just a few seconds and a modest amount of bandwidth. Compared to previous semi-honest protocols for private decision tree evaluation, we demonstrate a tenfold improvement in computation and bandwidth.

Speeding up the Number Theoretic Transform for Faster Ideal Lattice-Based Cryptography.

Abstract: The Number Theoretic Transform (NTT) provides efficient algorithms for cyclic and nega-cyclic convolutions, which have many applications in computer arithmetic, e.g., for multiplying large integers and large degree polynomials. It is commonly used in cryptographic schemes that are based on the hardness of the Ring Learning With Errors (R-LWE) problem to efficiently implement modular polynomial multiplication.

Speeding up the Number Theoretic Transform for Faster Ideal Lattice-Based Cryptography

Abstract: The Number Theoretic Transform (NTT) provides efficient algorithms for cyclic and nega-cyclic convolutions, which have many applications in computer arithmetic, e.g., for multiplying large integers and large degree polynomials. It is commonly used in cryptographic schemes that are based on the hardness of the Ring Learning With Errors (R-LWE) problem to efficiently implement modular polynomial multiplication.

Selecting elliptic curves for cryptography: an efficiency and security analysis

Abstract: We select a set of elliptic curves for cryptography and analyze our selection from a performance and security perspective. This analysis complements recent curve proposals that suggest (twisted) Edwards curves by also considering the Weierstrass model. Working with both Montgomery-friendly and pseudo-Mersenne primes allows us to consider more possibilities which help to improve the overall efficiency of base field arithmetic. Our Weierstrass curves are backwards compatible with current implementations of prime order NIST curves, while providing improved efficiency and stronger security properties. We choose algorithms and explicit formulas to demonstrate that our curves support constant-time, exception-free scalar multiplications, thereby offering high practical security in cryptographic applications. Our implementation shows that variable-base scalar multiplication on the new Weierstrass curves at the 128-bit security level is about 1.4 times faster than the recent implementation record on the corresponding NIST curve. For practitioners who are willing to use a different curve model and sacrifice a few bits of security, we present a collection of twisted Edwards curves with particularly efficient arithmetic that are up to 1.42, 1.26 and 1.24 times faster than the new Weierstrass curves at the 128-, 192- and 256-bit security levels, respectively. Finally, we discuss how these curves behave in a real-world protocol by considering different scalar multiplication scenarios in the transport layer security protocol. The proposed curves and the results of the analysis are intended to contribute to the recent efforts towards recommending new elliptic curves for Internet standards.

Efficient compression of SIDH public keys.

Abstract: Supersingular isogeny Diffie-Hellman (SIDH) is an attractive candidate for post-quantum key exchange, in large part due to its relatively small public key sizes. A recent paper by Azarderakhsh, Jao, Kalach, Koziel and Leonardi showed that the public keys defined in Jao and De Feo’s original SIDH scheme can be further compressed by around a factor of two, but reported that the performance penalty in utilizing this compression blew the overall SIDH runtime out by more than an order of magnitude. Given that the runtime of SIDH key exchange is currently its main drawback in relation to its lattice- and code-based post-quantum alternatives, an order of magnitude performance penalty for a factor of two improvement in bandwidth presents a trade-off that is unlikely to favor public-key compression in many scenarios.