Noisy Version Research Articles

Mobility-Assisted Over-the-Air Computation for Backscatter Sensor Networks

Future intelligent systems will consist of a massive number of battery-less sensors, where quick and accurate aggregation of sensor data will be of paramount importance. Over-the-air computation (AirComp) is a promising technology wherein sensors concurrently transmit their measurements over the wireless channel, and a reader receives the noisy version of a function of measurements due to the superposition property. A key challenge in AirComp is the accurate power alignment of individual transmissions, addressed previously by using conventional precoding methods. In this paper, we investigate a UAVenabled backscatter communication framework, wherein UAV acts both as a power emitter and reader. The mobility of the reader is leveraged to replace the complicated precoding at sensors, where UAV first collects sum channel gains in the first flyover, and then, use these to estimate the actual aggregated sensor data in the second flyover. Our results demonstrate improvements of up to 10 dB in MSE compared to that of a benchmark case where UAV is incognizant of sum channel gains.

Pathwise asymptotics for Volterra processes conditioned to a noisy version of the Brownian motion

In this paper we investigate a problem of large deviations for continuous Volterra processes under the influence of model disturbances. More precisely, we study the behavior, in the near future after $T$, of a Volterra process driven by a Brownian motion in a case where the Brownian motion is not directly observable, but only a noisy version is observed or some linear functionals of the noisy version are observed. Some examples are discussed in both cases.

Robust Neural Networks Learning via a Minimization of Stochastic Output Sensitivity

this article, we propose Sensitivity Minimization Learning (SML) to overcome the performance degradation problem caused by features corruption at the testing phase by using the stochastic sensitivity measure (STSM) as a regularizer. The STSM measures output deviations between each training sample and its noisy versions generated by a feature perturbation strategy. The feature perturbation strategy is user defined to simulate noises that our model target to defense at the testing phase. The SML is computational efficiency in training and testing phase and minimizes the generalization error on both training samples and its noise version set with feature perturbations. Models regularized by the STSM can be trained by the stochastic gradient descent algorithm efficiently and applied to very large scale applications. Experiment results on eight grayscale image, one color image, and two face databases show that the SML significantly outperforms several regularization techniques and yields much lower classification error when testing sets are contaminated with noises.

Group-Teaching: Learning Robust CNNs From Extremely Noisy Labels

Deep convolutional neural networks have achieved tremendous success in a variety of applications across many disciplines. However, their superior performance relies on correctly annotated large-scale datasets. It is very expensive and time-consuming to get the annotated large-scale datasets, especially in the medical field. While collecting a large amount of data is relatively easy, given the amount of data available on the web, but these data are highly unreliable, and they often include a massive amount of noisy labels. The past research works have shown that these noisy labels could significantly affect the performance of the deep convolutional neural networks on image classification. However, training a robust deep convolutional neural network with extremely noisy labels is a very challenging task. Inspired by the co-teaching concept, this paper proposes a novel method for training a robust convolutional neural network with extremely noisy labels, which is called group-teaching. Specifically, we train a group of convolutional neural networks simultaneously, and let them teach each other by selecting possibly clean samples for each network in each mini-batch. Each network back propagates the samples selected by other networks except itself and then it updates itself. The empirical results on noisy versions of CIFAR-10 and CIFAR-100 datasets demonstrate that our method is superior to the state-of-the-art methods in the robustness for noisy labels. Particularly, to verify the efficacy of our group-teaching in real-world noisy labels distribution, we have also validated the effectiveness of our method on the real-world noisy WebVision1000-100 dataset. The results show that our method has achieved higher performance than the state-of-the-art methods.

Dynamic predictive probabilities to monitor rapid cystic fibrosis disease progression.

Cystic fibrosis (CF) is a progressive, genetic disease characterized by frequent, prolonged drops in lung function. Accurately predicting rapid underlying lung‐function decline is essential for clinical decision support and timely intervention. Determining whether an individual is experiencing a period of rapid decline is complicated due to its heterogeneous timing and extent, and error component of the measured lung function. We construct individualized predictive probabilities for “nowcasting” rapid decline. We assume each patient's true longitudinal lung function, , follows a nonlinear, nonstationary stochastic process, and accommodate between‐patient heterogeneity through random effects. Corresponding lung‐function decline at time t is defined as the rate of change, . We predict conditional on observed covariate and measurement history by modeling a measured lung function as a noisy version of . The method is applied to data on 30 879 US CF Registry patients. Results are contrasted with a currently employed decision rule using single‐center data on 212 individuals. Rapid decline is identified earlier using predictive probabilities than the center's currently employed decision rule (mean difference: 0.65 years; 95% confidence interval (CI): 0.41, 0.89). We constructed a bootstrapping algorithm to obtain CIs for predictive probabilities. We illustrate real‐time implementation with R Shiny. Predictive accuracy is investigated using empirical simulations, which suggest this approach more accurately detects peak decline, compared with a uniform threshold of rapid decline. Median area under the ROC curve estimates (Q1‐Q3) were 0.817 (0.814‐0.822) and 0.745 (0.741‐0.747), respectively, implying reasonable accuracy for both. This article demonstrates how individualized rate of change estimates can be coupled with probabilistic predictive inference and implementation for a useful medical‐monitoring approach.

Automatic Evaluation of Song Intelligibility Using Singing Adapted STOI and Vocal-Specific Features

An objective machine-driven measure of song intelligibility would be of great utility for various music information retrieval tasks. Song intelligibility mostly depends on two factors, the amount of interference caused by background accompaniment, and the quality of singing vocal. We leverage these two factors to determine the intelligibility of a song. For the first factor, we adapt a well known method for intelligibility prediction of noisy speech, short term objective intelligibility (STOI), to singing. The singing-adapted STOI considers the polyphonic song as a time-frequency weighted noisy version of the extracted singing vocal. We use U-net based audio source separation method to extract singing vocal from a polyphonic song. The singing vocal shares the same underlying physiological mechanism for production as that of speech, with some differences in the pronunciation and prosody of the phonemes. Therefore, for the second factor, we have introduced vocal-specific features to measure the intelligibility of the singing vocal, which are excitation source, spectral, and prosodic singing characteristics. We perform detailed analysis on each of these features to establish their efficacy for quantifying song intelligibility. We train a regression model to derive the intelligibility scores using a combination of the vocal-specific features and singing adapted STOI, obtaining a significant improvement in performance. The correlation between the intelligibility score obtained using proposed framework and human-rated intelligibility score is 0.81, which shows the efficacy of the proposed approach.

A Gated Recurrent Convolutional Neural Network for Robust Spoofing Detection

Automatic speaker verification ASV systems are exposed to spoofing attacks which may compromise their security. While anti-spoofing techniques have been mainly studied for clean scenarios, it has also been shown that they perform poorly in noisy environments. In this work, we aim at improving the performance of spoofing detection for ASV in clean and noisy scenarios. To achieve this, we first propose the use of Gated Recurrent Convolutional Neural Networks GRCNNs as a deep feature extractor to robustly represent speech signals as utterance-level embeddings, which are later used by a back-end recognizer for the final genuine/spoofed classification. Then, to enhance the robustness of the system in noisy conditions, we propose the use of signal-to-noise masks SNMs as new input features to inform the anti-spoofing system about the time-frequency regions of the input spectral features that are mostly affected by noise and, hence, should be neglected when computing the embeddings. To evaluate our proposals, experiments were carried out on the clean and noisy versions of the ASVspoof 2015 corpus for detecting logical access attacks, as well as on the ASVspoof 2017 database to detect replay attacks. Additional results are provided for the ASVspoof 2019 corpus, including both logical and physical scenarios. The experimental results show that our proposal clearly outperforms some well-known methods based on classical features and other similar deep feature based systems for both clean and noisy conditions.

Heterogeneous multireference alignment for images with application to 2-D classification in single particle reconstruction.

Motivated by the task of 2-D classification in single particle reconstruction by cryo-electron microscopy (cryo-EM), we consider the problem of heterogeneous multireference alignment of images. In this problem, the goal is to estimate a (typically small) set of target images from a (typically large) collection of observations. Each observation is a rotated, noisy version of one of the target images. For each individual observation, neither the rotation nor which target image has been rotated are known. As the noise level in cryo-EM data is high, clustering the observations and estimating individual rotations is challenging. We propose a framework to estimate the target images directly from the observations, completely bypassing the need to cluster or register the images. The framework consists of two steps. First, we estimate rotation-invariant features of the images, such as the bispectrum. These features can be estimated to any desired accuracy, at any noise level, provided sufficiently many observations are collected. Then, we estimate the images from the invariant features. Numerical experiments on synthetic cryo-EM datasets demonstrate the effectiveness of the method. Ultimately, we outline future developments required to apply this method to experimental data.

A Comprehensive Study of the Key Enumeration Problem

In this paper, we will study the key enumeration problem, which is connected to the key recovery problem posed in the cold boot attack setting. In this setting, an attacker with physical access to a computer may obtain noisy data of a cryptographic secret key of a cryptographic scheme from main memory via this data remanence attack. Therefore, the attacker would need a key-recovery algorithm to reconstruct the secret key from its noisy version. We will first describe this attack setting and then pose the problem of key recovery in a general way and establish a connection between the key recovery problem and the key enumeration problem. The latter problem has already been studied in the side-channel attack literature, where, for example, the attacker might procure scoring information for each byte of an Advanced Encryption Standard (AES) key from a side-channel attack and then want to efficiently enumerate and test a large number of complete 16-byte candidates until the correct key is found. After establishing such a connection between the key recovery problem and the key enumeration problem, we will present a comprehensive review of the most outstanding key enumeration algorithms to tackle the latter problem, for example, an optimal key enumeration algorithm (OKEA) and several nonoptimal key enumeration algorithms. Also, we will propose variants to some of them and make a comparison of them, highlighting their strengths and weaknesses.

Prediction law of mixed Gaussian Volterra processes

We study the regular conditional law of mixed Gaussian Volterra processes under the influence of model disturbances. More precisely, we study prediction of Gaussian Volterra processes driven by a Brownian motion in a case where the Brownian motion is not observable, but only a noisy version is observed. As an application, we discuss how our result can be applied to variance reduction in the presence of measurement errors.

Enforcing constraints for interpolation and extrapolation in Generative Adversarial Networks

Generative Adversarial Networks (GANs) are becoming popular machine learning choices for training generators. At the same time there is a concerted effort in the machine learning community to expand the range of tasks in which learning can be applied as well as to utilize methods from other disciplines to accelerate learning. With this in mind, in the current work we suggest ways to enforce given constraints in the output of a GAN generator both for interpolation and extrapolation (prediction). For the case of dynamical systems, given a time series, we wish to train GAN generators that can be used to predict trajectories starting from a given initial condition. In this setting, the constraints can be in algebraic and/or differential form. Even though we are predominantly interested in the case of extrapolation, we will see that the tasks of interpolation and extrapolation are related. However, they need to be treated differently.For the case of interpolation, the incorporation of constraints is built into the training of the GAN. The incorporation of the constraints respects the primary game-theoretic setup of a GAN so it can be combined with existing algorithms. However, it can exacerbate the problem of instability during training that is well-known for GANs. We suggest adding small noise to the constraints as a simple remedy that has performed well in our numerical experiments.The case of extrapolation (prediction) is more involved. During training, the GAN generator learns to interpolate a noisy version of the data and we enforce the constraints. This approach has connections with model reduction that we can utilize to improve the efficiency and accuracy of the training. Depending on the form of the constraints, we may enforce them also during prediction through a projection step. We provide examples of linear and nonlinear systems of differential equations to illustrate the various constructions.

Non‐quantum cryptanalysis of the noisy version of Aaronson–Christiano's quantum money scheme

At STOC 2012, Aaronson and Christiano proposed a noisy and a noiseless version of the first public-key quantum money scheme endowed with a security proof. This paper addresses the so-called noisy hidden subspaces problem, on which the noisy version of their scheme is based. The first contribution of this work is a non-quantum cryptanalysis of the above-mentioned noisy quantum money scheme extended to prime fields F, with |F| ≠ 2, that runs in randomised polynomial time. This finding is supported with experimental results showing that, in practice, the algorithm presented is efficient and succeeds with overwhelming probability. The second contribution is a non-quantum randomised polynomial-time cryptanalysis of the noisy quantum money scheme over F2 succeeding with a certain probability for values of the noise lying within a certain range. This result disproves a conjecture made by Aaronson and Christiano about the non-existence of an algorithm that solves the noisy hidden subspaces problem over F2 and succeeds with such probability.

Constructing Extremal Compatible Quantum Observables by Means of Two Mutually Unbiased Bases

We describe a particular class of pairs of quantum observables which are extremal in the convex set of all pairs of compatible quantum observables. The pairs in this class are constructed as uniformly noisy versions of two mutually unbiased bases (MUB) with possibly different noise intensities affecting each basis. We show that not all pairs of MUB can be used in this construction, and we provide a criterion for determiniing those MUB that actually do yield extremal compatible observables. We apply our criterion to all pairs of Fourier conjugate MUB, and we prove that in this case extremality is achieved if and only if the quantum system Hilbert space is odd-dimensional. Remarkably, this fact is no longer true for general non-Fourier conjugate MUB, as we show in an example. Therefore, the presence or the absence of extremality is a concrete geometric manifestation of MUB inequivalence, that already materializes by comparing sets of no more than two bases at a time.

Implementing positive maps with multiple copies of an input state

Valid transformations between quantum states are necessarily described by completely positive maps, instead of just positive maps. Positive but not completely positive maps such as the transposition map cannot be implemented due to the existence of entanglement in composite quantum systems, but there are classes of states for which the positivity is guaranteed, e.g., states not correlated to other systems. In this paper, we introduce the concept of N-copy extension of maps to quantitatively analyze the difference between positive maps and completely positive maps. We consider implementations of the action of positive but not completely positive maps on uncorrelated states by allowing an extra resource of consuming multiple copies of the input state and characterize the positive maps in terms of implementability with multiple copies. We show that by consuming multiple copies, the set of implementable positive maps becomes larger, and almost all positive maps are implementable with finite copies of an input state. The number of copies of the input state required to implement a positive map quantifies the degree by which a positive map violates complete positivity. We then analyze the optimal N-copy implementability of a noisy version of the transposition map.

Inhibition increases response variability and reduces stimulus discrimination in random networks of cortical neurons

Much of what is known about the contribution of inhibition to stimulus discrimination is due to extensively studied sensory systems, which are highly structured neural circuits. The effect of inhibition on stimulus representation in less structured networks is not as clear. Here we exercise a biosynthetic approach in order to study the impacts of inhibition on stimulus representation in non-specialized network anatomy. Combining pharmacological manipulation, multisite electrical stimulation and recording from ex-vivo randomly rewired networks of cortical neurons, we quantified the effects of inhibition on response variability and stimulus discrimination at the population and single unit levels. We find that blocking inhibition quenches variability of responses evoked by repeated stimuli and enhances discrimination between stimuli that invade the network from different spatial loci. Enhanced stimulus discrimination is reserved for representation schemes that are based on temporal relation between spikes emitted in groups of neurons. Our data indicate that – under intact inhibition – the response to a given stimulus is a noisy version of the response evoked in the absence of inhibition. Spatial analysis suggests that the dispersion effect of inhibition is due to disruption of an otherwise coherent, wave-like propagation of activity.

Quantitative robustness analysis of quantum programs

Quantum computation is a topic of significant recent interest, with practical advances coming from both research and industry. A major challenge in quantum programming is dealing with errors (quantum noise) during execution. Because quantum resources (e.g., qubits) are scarce, classical error correction techniques applied at the level of the architecture are currently cost-prohibitive. But while this reality means that quantum programs are almost certain to have errors, there as yet exists no principled means to reason about erroneous behavior. This paper attempts to fill this gap by developing a semantics for erroneous quantum while-programs, as well as a logic for reasoning about them. This logic permits proving a property we have identified, called є-robustness, which characterizes possible “distance” between an ideal program and an erroneous one. We have proved the logic sound, and showed its utility on several case studies, notably: (1) analyzing the robustness of noisy versions of the quantum Bernoulli factory (QBF) and quantum walk (QW); (2) demonstrating the (in)effectiveness of different error correction schemes on single-qubit errors; and (3) analyzing the robustness of a fault-tolerant version of QBF.

Multivariate Data Retrieval Modified by Random Noise using Lattice Autoassociative Memories with Eroded or Dilated Input Residuals

Lattice associative memories were proposed as an alternative approach to work with a set of associated vector pairs for which the storage and retrieval stages are based in the theory of algebraic lattices. Several techniques have been established to deal with the problem of binary or real valued vector recall from corrupted inputs. This paper presents a thresholding technique coupled with statistical correlation pattern index search to enhance the recall performance of lattice auto-associative memories for multivariate data inputs degraded by random noise. By thresholding a given noisy input, a lower bound is generated to produce an eroded noisy version used to boost the min-lattice auto-associative memory inherent retrieval capability. Similarly, an upper bound is generated to obtain a dilated noisy version used to enhance the max-lattice auto-associave memory response. A self contained theoretical foundation is provided including a visual example of a multivariate data set composed of grayscale images that show the increased retrieval capability of this type of associative memories.

Noise Reduction in ECG Signals Using Fully Convolutional Denoising Autoencoders

The electrocardiogram (ECG) is an efficient and noninvasive indicator for arrhythmia detection and prevention. In real-world scenarios, ECG signals are prone to be contaminated with various noises, which may lead to wrong interpretation. Therefore, significant attention has been paid on denoising of ECG for accurate diagnosis and analysis. A denoising autoencoder (DAE) can be applied to reconstruct the clean data from its noisy version. In this paper, a DAE using the fully convolutional network (FCN) is proposed for ECG signal denoising. Meanwhile, the proposed FCN-based DAE can perform compression with regard to the DAE architecture. The proposed approach is applied to ECG signals from the MIT-BIH Arrhythmia database and the added noise signals are obtained from the MIT-BIH Noise Stress Test database. The denoising performance is evaluated using the root-mean-square error (RMSE), percentage-root-mean-square difference (PRD), and improvement in signal-to-noise ratio (SNR imp ). The results of the experiments conducted on noisy ECG signals of different levels of input SNR show that the FCN acquires better performance as compared to the deep fully connected neural network- and convolutional neural network-based denoising models. Moreover, the proposed FCN-based DAE reduces the size of the input ECG signals, where the compressed data is 32 times smaller than the original. The results of the study demonstrate the superiority of FCN in denoising, with lower RMSE and PRD, as well as higher SNR imp . According to the results, we believe that the proposed FCN-based DAE has a good application prospect in clinical practice.

Coding for Racetrack Memories

Racetrack memory is a new technology, which utilizes magnetic domains along a nanoscopic wire in order to obtain extremely high storage density. In racetrack memory, each magnetic domain can store a single bit of information, which can be sensed by a reading port (head). The memory is structured like a tape, which supports a shift operation that moves the domains to be read sequentially by the head. In order to increase the memory’s speed, prior work studied how to minimize the latency of the shift operation, while the no less important reliability of this operation has received only a little attention. In this paper, we design codes, which combat shift errors in racetrack memory, called position errors, namely, shifting the domains is not an error-free operation and the domains may be over shifted or are not shifted, which can be modeled as deletions and sticky insertions. While it is possible to use conventional deletion and insertion-correcting codes, we tackle this problem with the special structure of racetrack memory, where the domains can be read by multiple heads. Each head outputs a noisy version of the stored data and the multiple outputs are combined in order to reconstruct the data. This setup is a special case of the reconstruction problem studied by Levenshtein, however, in our case, the position errors from different heads are correlated. We will show how to take advantage of this special feature of racetrack memories in order to construct codes correcting deletions and sticky insertions. In particular, under this paradigm, we will show that it is possible to correct, with at most a single bit of redundancy, $d$ deletions with $d+1$ heads if the heads are well separated. Similar results are provided for burst of deletions, sticky insertions, and combinations of both deletions and sticky insertions.

Cold Boot Attacks on Ring and Module LWE Keys Under the NTT

In this work, we consider the ring- and module- variants of the LWE problem and investigate cold boot attacks on cryptographic schemes based on these problems, wherein an attacker is faced with the problem of recovering a scheme’s secret key from a noisy version of that key. The leakage resilience of cryptography based on the learning with errors (LWE) problem has been studied before, but there are only limited results considering the parameters observed in cold boot attack scenarios. There are two main encodings for storing ring- and module-LWE keys, and, as we show, the performance of cold boot attacks can be highly sensitive to the exact encoding used. The first encoding stores polynomial coefficients directly in memory. The second encoding performs a number theoretic transform (NTT) before storing the key, a commonly used method leading to more efficient implementations. We first give estimates for a cold boot attack complexity on the first encoding method based on standard algorithms; this analysis confirms that this encoding method is vulnerable to cold boot attacks only at very low bit-flip rates. We then show that, for the second encoding method, the structure introduced by using an NTT is exploitable in the cold boot setting: we develop a bespoke attack strategy that is much cheaper than our estimates for the first encoding when considering module-LWE keys. For example, at a 1% bit-flip rate (which corresponds roughly to what can be achieved in practice for cold boot attacks when applying cooling), a cold boot attack on Kyber KEM parameters has a cost of 243 operations when the second, NTT-based encoding is used for key storage, compared to 270 operations with the first encoding. On the other hand, in the case of the ring-LWE-based KEM, New Hope, the cold boot attack complexities are similar for both encoding methods.