Anderson C. A. Nascimento

Sikha Pentyala, Nicola Neophytou, Anderson Nascimento et al. · uw

Group fairness ensures that the outcome of machine learning (ML) based decision making systems are not biased towards a certain group of people defined by a sensitive attribute such as gender or ethnicity. Achieving group fairness in Federated Learning (FL) is challenging because mitigating bias inherently requires using the sensitive attribute values of all clients, while FL is aimed precisely at protecting privacy by not giving access to the clients' data. As we show in this paper, this conflict between fairness and privacy in FL can be resolved by combining FL with Secure Multiparty Computation (MPC) and Differential Privacy (DP). In doing so, we propose a method for training group-fair ML models in cross-device FL under complete and formal privacy guarantees, without requiring the clients to disclose their sensitive attribute values.

Mayana Pereira, Sikha Pentyala, Anderson Nascimento et al. · uw

Legal and ethical restrictions on accessing relevant data inhibit data science research in critical domains such as health, finance, and education. Synthetic data generation algorithms with privacy guarantees are emerging as a paradigm to break this data logjam. Existing approaches, however, assume that the data holders supply their raw data to a trusted curator, who uses it as fuel for synthetic data generation. This severely limits the applicability, as much of the valuable data in the world is locked up in silos, controlled by entities who cannot show their data to each other or a central aggregator without raising privacy concerns. To overcome this roadblock, we propose the first solution in which data holders only share encrypted data for differentially private synthetic data generation. Data holders send shares to servers who perform Secure Multiparty Computation (MPC) computations while the original data stays encrypted. We instantiate this idea in an MPC protocol for the Multiplicative Weights with Exponential Mechanism (MWEM) algorithm to generate synthetic data based on real data originating from many data holders without reliance on a single point of failure.

Tucker Stewart, Bin Yu, Anderson Nascimento et al. · uw

For network administration and maintenance, it is critical to anticipate when networks will receive peak volumes of traffic so that adequate resources can be allocated to service requests made to servers. In the event that sufficient resources are not allocated to servers, they can become prone to failure and security breaches. On the contrary, we would waste a lot of resources if we always allocate the maximum amount of resources. Therefore, anticipating peak volumes in network traffic becomes an important problem. However, popular forecasting models such as Autoregressive Integrated Moving Average (ARIMA) forecast time-series data generally, thus lack in predicting peak volumes in these time-series. More than often, a time-series is a combination of different features, which may include but are not limited to 1) Trend, the general movement of the traffic volume, 2) Seasonality, the patterns repeated over some time periods (e.g. daily and monthly), and 3) Noise, the random changes in the data. Considering that the fluctuation of seasonality can be harmful for trend and peak prediction, we propose to extract seasonalities to facilitate the peak volume predictions in the time domain. The experiments on both synthetic and real network traffic data demonstrate the effectiveness of the proposed method.

Shan Jin, Sai Rahul Rachuri, Yizhen Wang et al.

We propose a novel end-to-end privacy-preserving framework, instantiated by three efficient protocols for different deployment scenarios, covering both input and output privacy, for the vertically split scenario in federated learning (FL), where features are split across clients and labels are not shared by all parties. We do so by distributing the role of the aggregator in FL into multiple servers and having them run secure multiparty computation (MPC) protocols to perform model and feature aggregation and apply differential privacy (DP) to the final released model. While a naive solution would have the clients delegating the entirety of training to run in MPC between the servers, our optimized solution, which supports purely global and also global-local models updates with privacy-preserving, drastically reduces the amount of computation and communication performed using multiparty computation. The experimental results also show the effectiveness of our protocols.

Samuel Adams, Chaitali Choudhary, Martine De Cock et al.

Most existing Secure Multi-Party Computation (MPC) protocols for privacy-preserving training of decision trees over distributed data assume that the features are categorical. In real-life applications, features are often numerical. The standard ``in the clear'' algorithm to grow decision trees on data with continuous values requires sorting of training examples for each feature in the quest for an optimal cut-point in the range of feature values in each node. Sorting is an expensive operation in MPC, hence finding secure protocols that avoid such an expensive step is a relevant problem in privacy-preserving machine learning. In this paper we propose three more efficient alternatives for secure training of decision tree based models on data with continuous features, namely: (1) secure discretization of the data, followed by secure training of a decision tree over the discretized data; (2) secure discretization of the data, followed by secure training of a random forest over the discretized data; and (3) secure training of extremely randomized trees (``extra-trees'') on the original data. Approaches (2) and (3) both involve randomizing feature choices. In addition, in approach (3) cut-points are chosen randomly as well, thereby alleviating the need to sort or to discretize the data up front. We implemented all proposed solutions in the semi-honest setting with additive secret sharing based MPC. In addition to mathematically proving that all proposed approaches are correct and secure, we experimentally evaluated and compared them in terms of classification accuracy and runtime. We privately train tree ensembles over data sets with 1000s of instances or features in a few minutes, with accuracies that are at par with those obtained in the clear. This makes our solution orders of magnitude more efficient than the existing approaches, which are based on oblivious sorting.

Rafael Dowsley, Caleb Horst, Anderson C. A. Nascimento

We propose protocols for obliviously evaluating finite-state machines, i.e., the evaluation is shared between the provider of the finite-state machine and the provider of the input string in such a manner that neither party learns the other's input, and the states being visited are hidden from both. For alphabet size $|Σ|$, number of states $|Q|$, and input length $n$, previous solutions have either required a number of rounds linear in $n$ or communication $Ω(n|Σ||Q|\log|Q|)$. Our solutions require 2 rounds with communication $O(n(|Σ|+|Q|\log|Q|))$. We present two different solutions to this problem, a two-party one and a setting with an untrusted but non-colluding helper.

Amanda Resende, Davis Railsback, Rafael Dowsley et al.

We propose a privacy-preserving Naive Bayes classifier and apply it to the problem of private text classification. In this setting, a party (Alice) holds a text message, while another party (Bob) holds a classifier. At the end of the protocol, Alice will only learn the result of the classifier applied to her text input and Bob learns nothing. Our solution is based on Secure Multiparty Computation (SMC). Our Rust implementation provides a fast and secure solution for the classification of unstructured text. Applying our solution to the case of spam detection (the solution is generic, and can be used in any other scenario in which the Naive Bayes classifier can be employed), we can classify an SMS as spam or ham in less than 340ms in the case where the dictionary size of Bob's model includes all words (n = 5200) and Alice's SMS has at most m = 160 unigrams. In the case with n = 369 and m = 8 (the average of a spam SMS in the database), our solution takes only 21ms.

Raaghavi Sivaguru, Jonathan Peck, Femi Olumofin et al.

Malware applications typically use a command and control (C&C) server to manage bots to perform malicious activities. Domain Generation Algorithms (DGAs) are popular methods for generating pseudo-random domain names that can be used to establish a communication between an infected bot and the C&C server. In recent years, machine learning based systems have been widely used to detect DGAs. There are several well known state-of-the-art classifiers in the literature that can detect DGA domain names in real-time applications with high predictive performance. However, these DGA classifiers are highly vulnerable to adversarial attacks in which adversaries purposely craft domain names to evade DGA detection classifiers. In our work, we focus on hardening DGA classifiers against adversarial attacks. To this end, we train and evaluate state-of-the-art deep learning and random forest (RF) classifiers for DGA detection using side information that is harder for adversaries to manipulate than the domain name itself. Additionally, the side information features are selected such that they are easily obtainable in practice to perform inline DGA detection. The performance and robustness of these models is assessed by exposing them to one day of real-traffic data as well as domains generated by adversarial attack algorithms. We found that the DGA classifiers that rely on both the domain name and side information have high performance and are more robust against adversaries.

Martine De Cock, Rafael Dowsley, Anderson C. A. Nascimento et al.

In this paper, we present a secure logistic regression training protocol and its implementation, with a new subprotocol to securely compute the activation function. To the best of our knowledge, we present the fastest existing secure Multi-Party Computation implementation for training logistic regression models on high dimensional genome data distributed across a local area network.

Anisha Agarwal, Rafael Dowsley, Nicholas D. McKinney et al.

Machine learning (ML) is revolutionizing research and industry. Many ML applications rely on the use of large amounts of personal data for training and inference. Among the most intimate exploited data sources is electroencephalogram (EEG) data, a kind of data that is so rich with information that application developers can easily gain knowledge beyond the professed scope from unprotected EEG signals, including passwords, ATM PINs, and other intimate data. The challenge we address is how to engage in meaningful ML with EEG data while protecting the privacy of users. Hence, we propose cryptographic protocols based on Secure Multiparty Computation (SMC) to perform linear regression over EEG signals from many users in a fully privacy-preserving (PP) fashion, i.e.~such that each individual's EEG signals are not revealed to anyone else. To illustrate the potential of our secure framework, we show how it allows estimating the drowsiness of drivers from their EEG signals as would be possible in the unencrypted case, and at a very reasonable computational cost. Our solution is the first application of commodity-based SMC to EEG data, as well as the largest documented experiment of secret sharing based SMC in general, namely with 15 players involved in all the computations.

Devin Reich, Ariel Todoki, Rafael Dowsley et al.

Classification of personal text messages has many useful applications in surveillance, e-commerce, and mental health care, to name a few. Giving applications access to personal texts can easily lead to (un)intentional privacy violations. We propose the first privacy-preserving solution for text classification that is provably secure. Our method, which is based on Secure Multiparty Computation (SMC), encompasses both feature extraction from texts, and subsequent classification with logistic regression and tree ensembles. We prove that when using our secure text classification method, the application does not learn anything about the text, and the author of the text does not learn anything about the text classification model used by the application beyond what is given by the classification result itself. We perform end-to-end experiments with an application for detecting hate speech against women and immigrants, demonstrating excellent runtime results without loss of accuracy.

Claude Crépeau, Rafael Dowsley, Anderson C. A. Nascimento

Noisy channels are a valuable resource from a cryptographic point of view. They can be used for exchanging secret-keys as well as realizing other cryptographic primitives such as commitment and oblivious transfer. To be really useful, noisy channels have to be consider in the scenario where a cheating party has some degree of control over the channel characteristics. Damgård et al. (EUROCRYPT 1999) proposed a more realistic model where such level of control is permitted to an adversary, the so called unfair noisy channels, and proved that they can be used to obtain commitment and oblivious transfer protocols. Given that noisy channels are a precious resource for cryptographic purposes, one important question is determining the optimal rate in which they can be used. The commitment capacity has already been determined for the cases of discrete memoryless channels and Gaussian channels. In this work we address the problem of determining the commitment capacity of unfair noisy channels. We compute a single-letter characterization of the commitment capacity of unfair noisy channels. In the case where an adversary has no control over the channel (the fair case) our capacity reduces to the well-known capacity of a discrete memoryless binary symmetric channel.

Jonathan Peck, Claire Nie, Raaghavi Sivaguru et al.

Domain generation algorithms (DGAs) are commonly leveraged by malware to create lists of domain names which can be used for command and control (C&C) purposes. Approaches based on machine learning have recently been developed to automatically detect generated domain names in real-time. In this work, we present a novel DGA called CharBot which is capable of producing large numbers of unregistered domain names that are not detected by state-of-the-art classifiers for real-time detection of DGAs, including the recently published methods FANCI (a random forest based on human-engineered features) and LSTM.MI (a deep learning approach). CharBot is very simple, effective and requires no knowledge of the targeted DGA classifiers. We show that retraining the classifiers on CharBot samples is not a viable defense strategy. We believe these findings show that DGA classifiers are inherently vulnerable to adversarial attacks if they rely only on the domain name string to make a decision. Designing a robust DGA classifier may, therefore, necessitate the use of additional information besides the domain name alone. To the best of our knowledge, CharBot is the simplest and most efficient black-box adversarial attack against DGA classifiers proposed to date.

Sisi Wang, Wing-Sea Poon, Golnoosh Farnadi et al.

User profiling from user generated content (UGC) is a common practice that supports the business models of many social media companies. Existing systems require that the UGC is fully exposed to the module that constructs the user profiles. In this paper we show that it is possible to build user profiles without ever accessing the user's original data, and without exposing the trained machine learning models for user profiling -- which are the intellectual property of the company -- to the users of the social media site. We present VirtualIdentity, an application that uses secure multi-party cryptographic protocols to detect the age, gender and personality traits of users by classifying their user-generated text and personal pictures with trained support vector machine models in a privacy-preserving manner.

Rafael Dowsley, Jörn Müller-Quade, Anderson C. A. Nascimento

We show that stand-alone statistically secure random oblivious transfer protocols based on two-party stateless primitives are statistically universally composable. I.e. they are simulatable secure with an unlimited adversary, an unlimited simulator and an unlimited environment machine. Our result implies that several previous oblivious transfer protocols in the literature which were proven secure under weaker, non-composable definitions of security can actually be used in arbitrary statistically secure applications without lowering the security.

Paulo Barreto, Anderson Nascimento, Glaucio Oliveira et al.

We present an oblivious transfer (OT) protocol that combines the OT scheme of Chou and Orlandi together with thesupersingular isogeny Diffie-Hellman (SIDH) primitive of De Feo, Jao, and Plût. Our construction is a candidate for post-quantum secure OT and demonstrates that SIDH naturally supports OT functionality. We consider the protocol in the simplest configuration of $\binom{2}{1}$-OT and analyze the protocol to verify its security.

Paulo S. L. M. Barreto, Bernardo David, Rafael Dowsley et al.

Oblivious Transfer (OT) is a fundamental cryptographic protocol that finds a number of applications, in particular, as an essential building block for two-party and multi-party computation. We construct a round-optimal (2 rounds) universally composable (UC) protocol for oblivious transfer secure against active adaptive adversaries from any OW-CPA secure public-key encryption scheme with certain properties in the random oracle model (ROM). In terms of computation, our protocol only requires the generation of a public/secret-key pair, two encryption operations and one decryption operation, apart from a few calls to the random oracle. In~terms of communication, our protocol only requires the transfer of one public-key, two ciphertexts, and three binary strings of roughly the same size as the message. Next, we show how to instantiate our construction under the low noise LPN, McEliece, QC-MDPC, LWE, and CDH assumptions. Our instantiations based on the low noise LPN, McEliece, and QC-MDPC assumptions are the first UC-secure OT protocols based on coding assumptions to achieve: 1) adaptive security, 2) optimal round complexity, 3) low communication and computational complexities. Previous results in this setting only achieved static security and used costly cut-and-choose techniques.Our instantiation based on CDH achieves adaptive security at the small cost of communicating only two more group elements as compared to the gap-DH based Simplest OT protocol of Chou and Orlandi (Latincrypt 15), which only achieves static security in the ROM.

Rafael Dowsley, Felipe Lacerda, Anderson C. A. Nascimento

The bounded storage model restricts the memory of an adversary in a cryptographic protocol, rather than restricting its computational power, making information theoretically secure protocols feasible. We present the first protocols for commitment and oblivious transfer in the bounded storage model with errors, i.e., the model where the public random sources available to the two parties are not exactly the same, but instead are only required to have a small Hamming distance between themselves. Commitment and oblivious transfer protocols were known previously only for the error-free variant of the bounded storage model, which is harder to realize.

Rafael Dowsley, Anderson C. A. Nascimento

Noisy channels are a powerful resource for cryptography as they can be used to obtain information-theoretically secure key agreement, commitment and oblivious transfer protocols, among others. Oblivious transfer (OT) is a fundamental primitive since it is complete for secure multi-party computation, and the OT capacity characterizes how efficiently a channel can be used for obtaining string oblivious transfer. Ahlswede and Csiszár (\emph{ISIT'07}) presented upper and lower bounds on the OT capacity of generalized erasure channels (GEC) against passive adversaries. In the case of GEC with erasure probability at least 1/2, the upper and lower bounds match and therefore the OT capacity was determined. It was later proved by Pinto et al. (\emph{IEEE Trans. Inf. Theory 57(8)}) that in this case there is also a protocol against malicious adversaries achieving the same lower bound, and hence the OT capacity is identical for passive and malicious adversaries. In the case of GEC with erasure probability smaller than 1/2, the known lower bound against passive adversaries that was established by Ahlswede and Csiszár does not match their upper bound and it was unknown whether this OT rate could be achieved against malicious adversaries as well. In this work we show that there is a protocol against malicious adversaries achieving the same OT rate that was obtained against passive adversaries. In order to obtain our results we introduce a novel use of interactive hashing that is suitable for dealing with the case of low erasure probability ($p^* <1/2$).

Nico Döttling, Rafael Dowsley, Jörn Müller-Quade et al.

The McEliece public-key encryption scheme has become an interesting alternative to cryptosystems based on number-theoretical problems. Differently from RSA and ElGa- mal, McEliece PKC is not known to be broken by a quantum computer. Moreover, even tough McEliece PKC has a relatively big key size, encryption and decryption operations are rather efficient. In spite of all the recent results in coding theory based cryptosystems, to the date, there are no constructions secure against chosen ciphertext attacks in the standard model - the de facto security notion for public-key cryptosystems. In this work, we show the first construction of a McEliece based public-key cryptosystem secure against chosen ciphertext attacks in the standard model. Our construction is inspired by a recently proposed technique by Rosen and Segev.