When receiving machine learning services from the cloud, the provider does not need to receive all features; in fact, only a subset of the features are necessary for the target prediction task. Discerning this subset is the key problem of this work. We formulate this problem as a gradient-based perturbation maximization method that discovers this subset in the input feature space with respect to the functionality of the prediction model used by the provider. After identifying the subset, our framework, Cloak, suppresses the rest of the features using utility-preserving constant values that are discovered through a separate gradient-based optimization process. We show that Cloak does not necessarily require collaboration from the service provider beyond its normal service, and can be applied in scenarios where we only have black-box access to the service provider’s model. We theoretically guarantee that Cloak’s optimizations reduce the upper bound of the Mutual Information (MI) between the data and the sifted representations that are sent out. Experimental results show that Cloak reduces the mutual information between the input and the sifted representations by 85.01% with only negligible reduction in utility (1.42%). In addition, we show that Cloak greatly diminishes adversaries’ ability to learn and infer non-conducive features.

A wide variety of deep neural applications increasingly rely on the cloud to perform their compute-heavy inference. This common practice requires sending private and privileged data over the network to remote servers, exposing it to the service provider and potentially compromising its privacy. Even if the provider is trusted, the data can still be vulnerable over communication channels or via side-channel attacks in the cloud. To that end, this paper aims to reduce the information content of the communicated data with as little as possible compromise on the inference accuracy by making the sent data noisy. An undisciplined addition of noise can significantly reduce the accuracy of inference, rendering the service unusable. To address this challenge, this paper devises Shredder, an end-to-end framework, that, without altering the topology or the weights of a pre-trained network, learns additive noise distributions that significantly reduce the information content of communicated data while maintaining the inference accuracy. The key idea is finding the additive noise distributions by casting it as a disjoint offline learning process with a loss function that strikes a balance between accuracy and information degradation. The loss function also exposes a knob for a disciplined and controlled asymmetric trade-off between privacy and accuracy. While keeping the DNN intact, Shredder divides inference between the cloud and the edge device, striking a balance between computation and communication. In the separate phase of inference, the edge device takes samples from the Laplace distributions that were collected during the proposed offline learning phase and populates a noise tensor with these sampled elements. Then, the edge device merely adds this populated noise tensor to the intermediate results to be sent to the cloud. As such, Shredder enables accurate inference on noisy intermediate data without the need to update the model or the cloud, or any training process during inference. We also formally show that Shredder maximizes privacy with minimal impact on DNN accuracy while the tradeoff between privacy and accuracy is controlled through a mathematical knob. Experimentation with six real-world DNNs from text processing and image classification shows that Shredder reduces the mutual information between the input and the communicated data to the cloud by 74.70% compared to the original execution while only sacrificing 1.58% loss in accuracy. On average, Shredder also offers a speedup of 1.79x over Wi-Fi and 2.17x over LTE compared to cloud-only execution when using an off-the-shelf mobile GPU (Tegra X2) on the edge.

In this paper, the impacts of the simultaneous switching noise (SSN) in carbon nanotube field effect transistor-based ternary circuits are investigated. These effects, including the peak noise on the VDD and ground rails and the SSN-induced delay and output noise are compared between traditional Cu and multiwall carbon nanotube bundle power interconnects in ternary circuits. Simulations are performed using HSPICE for global power interconnects at 14- and 7-nm technology nodes. The results indicate that for interconnects with 200 μm length, the peak SSN voltage on the VDD and ground rails for a power distribution network, including ten ternary buffers, using multi-walled carbon nanotube (MWCNT) bundle power interconnects is 53% and 40% lower, respectively, compared to Cu power interconnects in the last stage at the 14-nm node. Also, with scaling down the technology to 7 nm, these improvements increase to 60% and 59%, respectively. Moreover, MWCNT bundle power interconnects reduce the SSN-induced delay at the output of the tenth stage for interconnects with 200-μm length on average by 82% as compared to the Cu interconnects at the 14-nm node. This improvement is 73% for the 7-nm technology node.

This paper is concerned with predicting the occurrence of periventricular leukomalacia (PVL) using vital and blood gas data which are collected over a period of 12 h after the neonatal cardiac surgery. A data mining approach has been employed to generate a set of rules for classification of subjects as healthy or PVL affected. In view of the fact that blood gas and vital data have different sampling rates, in this study we have divided the data into two categories: 1) high resolution (vital), and 2) low resolution (blood gas), and designed a separate classifier based on each data category. The developed algorithm is composed of several stages; first, a feature pool has been extracted from each data category and the extracted features have been ranked based on the data reliability and their mutual information content with the output. An optimal feature subset with the highest discriminative capability has been formed using simultaneous maximization of the class separability measure and mutual information of a set. Two separate decision trees (DTs) have been developed for the classification purpose and more importantly to discover hidden relationships that exist among the data to help us better understand PVL pathophysiology. The DT result shows that high amplitude 20 min variations and low sample entropy in the vital data and the defined out of range index as well as maximum rate of change in blood gas data are important factors for PVL prediction. Low sample entropy represents lack of variability in hemodynamic measurement, and constant blood pressure with small fluctuations is an important indicator of PVL occurrence. Finally, using the different time frames of data collection, we show that the first 6 h of data contain sufficient information for PVL occurrence prediction.

This paper considers the recovery of a low-rank matrix from an observed version that simultaneously contains both 1) erasures, most entries are not observed, and 2) errors, values at a constant fraction of (unknown) locations are arbitrarily corrupted. We provide a new unified performance guarantee on when minimizing nuclear norm plusl1norm succeeds in exact recovery. Our result allows for the simultaneous presence of random and deterministic components in both the error and erasure patterns. By specializing this one single result in different ways, we recover (up to poly-log factors) as corollaries all the existing results in exact matrix completion, and exact sparse and low-rank matrix decomposition. Our unified result also provides the first guarantees for 1) recovery when we observe a vanishing fraction of entries of a corrupted matrix, and 2) deterministic matrix completion.

The task of sparse linear regression consists of finding an unknown sparse vector from linear measurements. Solving this task even under “high-dimensional” settings, where the number of samples is fewer than the number of variables, is now known to be possible via methods such as the LASSO. We consider the multiple sparse linear regression problem, where the task consists of recovering several related sparse vectors at once. A simple approach to this task would involve solving independent sparse linear regression problems, but a natural question is whether one can reduce the overall number of samples required by leveraging partial sharing of the support sets, or nonzero patterns, of the signal vectors. A line of recent research has studied the use of ℓ1/ℓqnorm block-regularizations with q > 1 for such problems. However, depending on the level of sharing, these could actually perform worse in sample complexity when compared to solving each problem independently. We present a new “adaptive” method for multiple sparse linear regression that can leverage support and parameter overlap when it exists, but not pay a penalty when it does not. We show how to achieve this using a very simple idea: decompose the parameters into two components and regularize these differently. We show, theoretically and empirically, that our method strictly and noticeably outperforms both ℓ1or ℓ1/ℓqmethods, over the entire range of possible overlaps (except at boundary cases, where we match the best method), even under high-dimensional scaling.

Abstract The objective of this study is to develop an algorithm to detect and classify six types of electrocardiogram (ECG) signal beats including normal beats (N), atrial pre‐mature beats (A), right bundle branch block beats (R), left bundle branch block beats (L), paced beats (P), and pre‐mature ventricular contraction beats (PVC or V) using a neural network classifier. In order to prepare an appropriate input vector for the neural classifier several pre‐processing stages have been applied. Initially, a signal filtering method is used to remove the ECG signal baseline wandering. Continuous wavelet transform is then applied in order to extract features of the ECG signal. Next, principal component analysis is used to reduce the size of the data. A well‐known neural network architecture called the multi‐layered perceptron neural network is then utilized as the final classifier to classify each ECG beat as one of six groups of signals under study. Finally, the MIT‐BIH database is used to evaluate the proposed algorithm, resulting in 99.5% sensitivity, 99.66% positive predictive accuracy and 99.17% total accuracy.

This paper considers the recovery of a low-rank matrix from an observed version that simultaneously contains both (a) erasures: most entries are not observed, and (b) errors: values at a constant fraction of (unknown) locations are arbitrarily corrupted. We provide a new unified performance guarantee on when a (natural) recently proposed method, based on convex optimization, succeeds in exact recovery. Our result allows for the simultaneous presence of random and deterministic components in both the error and erasure patterns. On the one hand, corollaries obtained by specializing this one single result in different ways recovers (upto poly-log factors) all the existing works in matrix completion, and sparse and low-rank matrix recovery. On the other hand, our results also provide the first guarantees for (a) deterministic matrix completion, and (b) recovery when we observe a vanishing fraction of entries of a corrupted matrix.

In this paper, we address the problem of learning the structure of a pairwise graphical model from samples in a high-dimensional setting. Our first main result studies the sparsistency, or consistency in sparsity pattern recovery, properties of a forward-backward greedy algorithm as applied to general statistical models. As a special case, we then apply this algorithm to learn the structure of a discrete graphical model via neighborhood estimation. As a corollary of our general result, we derive sufficient conditions on the number of samples n, the maximum node-degree d and the problem size p, as well as other conditions on the model parameters, so that the algorithm recovers all the edges with high probability. Our result guarantees graph selection for samples scaling as n = Omega(d log(p)), in contrast to existing convex-optimization based algorithms that require a sample complexity of Omega(d^2 log(p)). Further, the greedy algorithm only requires a restricted strong convexity condition which is typically milder than irrepresentability assumptions. We corroborate these results using numerical simulations at the end.

We consider the multiple linear regression problem, in a setting where some of the set of relevant features could be shared across the tasks. A lot of recent research has studied the use of $\ell_1/\ell_q$ norm block-regularizations with $q > 1$ for such (possibly) block-structured problems, establishing strong guarantees on recovery even under high-dimensional scaling where the number of features scale with the number of observations. However, these papers also caution that the performance of such block-regularized methods are very dependent on the {\em extent} to which the features are shared across tasks. Indeed they show~\citep{NWJoint} that if the extent of overlap is less than a threshold, or even if parameter {\em values} in the shared features are highly uneven, then block $\ell_1/\ell_q$ regularization could actually perform {\em worse} than simple separate elementwise $\ell_1$ regularization. We are far away from a realistic multi-task setting: not only do the set of relevant features have to be exactly the same across tasks, but their values have to as well. Here, we ask the question: can we leverage support and parameter overlap when it exists, but not pay a penalty when it does not? Indeed, this falls under a more general question of whether we can model such \emph{dirty data} which may not fall into a single neat structural bracket (all block-sparse, or all low-rank and so on). Here, we take a first step, focusing on developing a dirty model for the multiple regression problem. Our method uses a very simple idea: we decompose the parameters into two components and {\em regularize these differently.} We show both theoretically and empirically, our method strictly and noticeably outperforms both $\ell_1$ and $\ell_1/\ell_q$ methods, over the entire range of possible overlaps. We also provide theoretical guarantees that the method performs well under high-dimensional scaling.

In this paper, a new graph product, called Edge Graph Product (EGP) is proposed by replacing each edge in the multiplicand graph by a copy of the multiplier graph via two candidate nodes. The edge product, unlike other products already proposed, results in a graph whose number of edges is numerical product of the number of the edges in the multiplicand and multiplier graphs, and the number of vertices is not equal to the numerical product of the number of vertices in the multiplicand and multiplier graphs. After formal definition of the new product, some basic properties of the product operator are studied. We then address Hamiltonian, Eulerian and routing properties of the new product, and we show that some of the recently proposed topologies fall within the family of edge product graphs.