We study the problem of searching on data that is encrypted using a public key system. Consider user Bob who sends email to user Alice encrypted under Alice’s public key. An email gateway wants to test whether the email contains the keyword “urgent” so that it could route the email accordingly. Alice, on the other hand does not wish to give the gateway the ability to decrypt all her messages. We define and construct a mechanism that enables Alice to provide a key to the gateway that enables the gateway to test whether the word “urgent” is a keyword in the email without learning anything else about the email. We refer to this mechanism as Public Key Encryption with keyword Search. As another example, consider a mail server that stores various messages publicly encrypted for Alice by others. Using our mechanism Alice can send the mail server a key that will enable the server to identify all messages containing some specific keyword, but learn nothing else. We define the concept of public key encryption with keyword search and give several constructions.

https://link.springer.com/content/pdf/10.1007%2F978-3-540-24676-3_30.pdf

Public Key Encryption with Keyword Search

Machine learning techniques based on neural networks are achieving remarkable results in a wide variety of domains. Often, the training of models requires large, representative datasets, which may be crowdsourced and contain sensitive information. The models should not expose private information in these datasets. Addressing this goal, we develop new algorithmic techniques for learning and a refined analysis of privacy costs within the framework of differential privacy. Our implementation and experiments demonstrate that we can train deep neural networks with non-convex objectives, under a modest privacy budget, and at a manageable cost in software complexity, training efficiency, and model quality.

Deep Learning with Differential Privacy

In this paper we address the issue of privacy preserving data mining. Specifically, we consider a scenario in which two parties owning confidential databases wish to run a data mining algorithm on the union of their databases, without revealing any unnecessary information. Our work is motivated by the need both to protect privileged information and to enable its use for research or other purposes.

The above problem is a specific example of secure multi-party computation and, as such, can be solved using known generic protocols. However, data mining algorithms are typically complex and, furthermore, the input usually consists of massive data sets. The generic protocols in such a case are of no practical use and therefore more efficient protocols are required. We focus on the problem of decision tree learning with the popular ID3 algorithm. Our protocol is considerably more efficient than generic solutions and demands both very few rounds of communication and reasonable bandwidth.

https://link.springer.com/content/pdf/10.1007/s00145-001-0019-2.pdf

Privacy Preserving Data Mining

We quantitatively investigate how machine learning models leak information about the individual data records on which they were trained. We focus on the basic membership inference attack: given a data record and black-box access to a model, determine if the record was in the model's training dataset. To perform membership inference against a target model, we make adversarial use of machine learning and train our own inference model to recognize differences in the target model's predictions on the inputs that it trained on versus the inputs that it did not train on. We empirically evaluate our inference techniques on classification models trained by commercial "machine learning as a service" providers such as Google and Amazon. Using realistic datasets and classification tasks, including a hospital discharge dataset whose membership is sensitive from the privacy perspective, we show that these models can be vulnerable to membership inference attacks. We then investigate the factors that influence this leakage and evaluate mitigation strategies.

/pdf/membership-inference-attacks-against-machine-learning-models-2tzchejf91.pdf

Membership Inference Attacks Against Machine Learning Models

Usually, a proof of a theorem contains more knowledge than the mere fact that the theorem is true. For instance, to prove that a graph is Hamiltonian it suffices to exhibit a Hamiltonian tour in it; however, this seems to contain more knowledge than the single bit Hamiltonian/non-Hamiltonian.In this paper a computational complexity theory of the “knowledge” contained in a proof is developed. Zero-knowledge proofs are defined as those proofs that convey no additional knowledge other than the correctness of the proposition in question. Examples of zero-knowledge proof systems are given for the languages of quadratic residuosity and 'quadratic nonresiduosity. These are the first examples of zero-knowledge proofs for languages not known to be efficiently recognizable.

/pdf/the-knowledge-complexity-of-interactive-proof-systems-31apre0ecf.pdf

The knowledge complexity of interactive proof-systems

The pseudo-random number generator (PRNG) used by the Windows operating system is the most commonly used PRNG. The pseudo-randomness of the output of this generator is crucial for the security of almost any application running in Windows. Nevertheless, its exact algorithm was never published.We examined the binary code of a distribution of Windows 2000, which is still the second most popular operating system after Windows XP. (This investigation was done without any help from Microsoft.) We reconstructed, for the first time, the algorithm used by the pseudo-random number generator (namely, the function CryptGenRandom). We analyzed the security of the algorithm and found a on-trivial attack: given the internal state of the generator, the previous state can be computed in O(223) work (this is an attack on the forward-security of the generator, an O(1) attack on backward security is trivial). The attack on the forward-security demonstrates that the design of the generator is flawed, since it is well known how to prevent such attacks.We also analyzed the way in which the generator is run by the operating system, and found that it amplifies the effect of the attacks. As a result, learning a single state may reveal 128 Kbytes of the past and future output of the generator. The implication of these findings is that a buffer overflow attack or a similar attack can be used to learn a single state of the generator, which can then be used to predict all random values, such as SSL keys, used by a process in all its past and future operation. This attack is more severe and more efficient than known attacks, in which an attacker can only learn SSL keys if it is controlling the attacked machine at the time the keys are used.

/pdf/cryptanalysis-of-the-windows-random-number-generator-3u6cz246vq.pdf

Cryptanalysis of the windows random number generator

This paper examines secure two-party computation of functions which depend only on the Hamming distance of the inputs of the two parties. We present efficient protocols for computing these functions. In particular, we present protocols which are secure in the sense of full simulatability against malicious adversaries.

We show different applications of this family of functions, including a protocol we call m -point-SPIR, which is an efficient variant of symmetric private information retrieval (SPIR). It can be used if the server's database contains N entries, at most N /logN of which have individual values, and the rest are set to some default value. This variant of PIR is unique since it can be based on the existence of OT alone.

/pdf/secure-hamming-distance-based-computation-and-its-kx11iy9h66.pdf

Secure Hamming Distance Based Computation and Its Applications

Recently, there has been huge progress in the field of concretely efficient secure computation, even while providing security in the presence of malicious adversaries. This is especially the case in the two-party setting, where constant-round protocols exist that remain fast even over slow networks. However, in the multi-party setting, all concretely efficient fully-secure protocols, such as SPDZ, require many rounds of communication.

/pdf/efficient-constant-round-multi-party-computation-combining-2rpmil7cui.pdf

Efficient Constant Round Multi-party Computation Combining BMR and SPDZ

The properties of weak sources of randomness have been investigated in many contexts and using several models of weakly random behaviour For two such models, developed by Santha and Vazirani, and Chor and Goldreich, it is known that the output from one such source cannot be "compressed" to produce nearly random bits At the same time, however, a single source is sufficient to solve problems in the randomized complexity classes BPP and RP It is natural to ask exactly which tasks can be done using a single, weak source of randomness and which cannot The present work begins to answer this question by establishing that a single weakly random source of either model cannot be used to obtain a secure "one-time-pad" type of cryptosystem

/pdf/on-the-impossibility-of-private-key-cryptography-with-weakly-2aes4kikdn.pdf

On the Impossibility of Private Key Cryptography with Weakly Random Keys

Protocols for secure computation enable mutually distrustful parties to jointly compute on their private inputs without revealing anything but the result. Over recent years, secure computation has become practical and considerable effort has been made to make it more and more efficient. A highly important tool in the design of two-party protocols is Yao's garbled circuit construction (Yao 1986), and multiple optimizations on this primitive have led to performance improvements of orders of magnitude over the last years. However, many of these improvements come at the price of making very strong assumptions on the underlying cryptographic primitives being used (e.g., that AES is secure for related keys, that it is circular secure, and even that it behaves like a random permutation when keyed with a public fixed key). The justification behind making these strong assumptions has been that otherwise it is not possible to achieve fast garbling and thus fast secure computation. In this paper, we take a step back and examine whether it is really the case that such strong assumptions are needed. We provide new methods for garbling that are secure solely under the assumption that the primitive used (e.g., AES) is a pseudorandom function. Our results show that in many cases, the penalty incurred is not significant, and so a more conservative approach to the assumptions being used can be adopted.

/pdf/fast-garbling-of-circuits-under-standard-assumptions-42iexzpgqk.pdf

Benny Pinkas

Papers

Cryptanalysis of the windows random number generator

Secure Hamming Distance Based Computation and Its Applications

Efficient Constant Round Multi-party Computation Combining BMR and SPDZ

On the Impossibility of Private Key Cryptography with Weakly Random Keys

Fast Garbling of Circuits Under Standard Assumptions