Noisy Reference Research Articles

PROTAX-GPU: a scalable probabilistic taxonomic classification system for DNA barcodes.

DNA-based identification is vital for classifying biological specimens, yet methods to quantify the uncertainty of sequence-based taxonomic assignments are scarce. Challenges arise from noisy reference databases, including mislabelled entries and missing taxa. PROTAX addresses these issues with a probabilistic approach to taxonomic classification, advancing on methods that rely solely on sequence similarity. It provides calibrated probabilistic assignments to a partially populated taxonomic hierarchy, accounting for taxa that lack references and incorrect taxonomic annotation. While effective on smaller scales, global application of PROTAX necessitates substantially larger reference libraries, a goal previously hindered by computational barriers. We introduce PROTAX-GPU, a scalable algorithm capable of leveraging the global Barcode of Life Data System (>14 million specimens) as a reference database. Using graphics processing units (GPU) to accelerate similarity and nearest-neighbour operations and the JAX library for Python integration, we achieve over a 1000 × speedup compared with the central processing unit (CPU)-based implementation without compromising PROTAX's key benefits. PROTAX-GPU marks a significant stride towards real-time DNA barcoding, enabling quicker and more efficient species identification in environmental assessments. This capability opens up new avenues for real-time monitoring and analysis of biodiversity, advancing our ability to understand and respond to ecological dynamics. This article is part of the theme issue 'Towards a toolkit for global insect biodiversity monitoring'.

The "hidden noise" problem in MR image reconstruction.

The performance of modern image reconstruction methods is commonly judged using quantitative error metrics like root mean squared-error and the structural similarity index, which are calculated by comparing reconstructed images against fully sampled reference data. In practice, the reference data will contain noise and is not a true gold standard. In this work, we demonstrate that the "hidden noise" present in reference data can substantially confound standard approaches for ranking different image reconstruction results. Using both experimental and simulated k-space data and several different image reconstruction techniques, we examined whether there was correlation between performance metrics obtained with typical noisy reference data versus those obtained with higher-quality reference data. For conventional performance metrics, the reconstructions that matched best with the higher-quality reference data were substantially different from the reconstructions that matched best with typical noisy reference data. This leads to suboptimal reconstruction results if the performance with respect to noisy reference data is used to select which reconstruction methods/parameters to employ. These issues were reduced when employing alternative error metrics that better account for noise. Reference data containing hidden noise can substantially mislead the ranking of image reconstruction methods when using conventional error metrics, but this issue can be mitigated with alternative error metrics.

HyGAnno: hybrid graph neural network-based cell type annotation for single-cell ATAC sequencing data.

Reliable cell type annotations are crucial for investigating cellular heterogeneity in single-cell omics data. Although various computational approaches have been proposed for single-cell RNA sequencing (scRNA-seq) annotation, high-quality cell labels are still lacking in single-cell sequencing assay for transposase-accessible chromatin (scATAC-seq) data, because of extreme sparsity and inconsistent chromatin accessibility between datasets. Here, we present a novel automated cell annotation method that transfers cell type information from a well-labeled scRNA-seq reference to an unlabeled scATAC-seq target, via a parallel graph neural network, in a semi-supervised manner. Unlike existing methods that utilize only gene expression or gene activity features, HyGAnno leverages genome-wide accessibility peak features to facilitate the training process. In addition, HyGAnno reconstructs a reference-target cell graph to detect cells with low prediction reliability, according to their specific graph connectivity patterns. HyGAnno was assessed across various datasets, showcasing its strengths in precise cell annotation, generating interpretable cell embeddings, robustness to noisy reference data and adaptability to tumor tissues.

Emulating Low-Dose PCCT Image Pairs with Independent Noise for Self-Supervised Spectral Image Denoising.

Photon counting CT (PCCT) acquires spectral measurements and enables generation of material decomposition (MD) images that provide distinct advantages in various clinical situations. However, noise amplification is observed in MD images, and denoising is typically applied. Clean or high-quality references are rare in clinical scans, often making supervised learning (Noise2Clean) impractical. Noise2Noise is a self-supervised counterpart, using noisy images and corresponding noisy references with zero-mean, independent noise. PCCT counts transmitted photons separately, and raw measurements are assumed to follow a Poisson distribution in each energy bin, providing the possibility to create noise-independent pairs. The approach is to use binomial selection to split the counts into two low-dose scans with independent noise. We prove that the reconstructed spectral images inherit the noise independence from counts domain through noise propagation analysis and also validated it in numerical simulation and experimental phantom scans. The method offers the flexibility to split measurements into desired dose levels while ensuring the reconstructed images share identical underlying features, thereby strengthening the model's robustness for input dose levels and capability of preserving fine details. In both numerical simulation and experimental phantom scans, we demonstrated that Noise2Noise with binomial selection outperforms other common self-supervised learning methods based on different presumptive conditions.

BEDwARS: a robust Bayesian approach to bulk gene expression deconvolution with noisy reference signatures

Differential gene expression in bulk transcriptomics data can reflect change of transcript abundance within a cell type and/or change in the proportions of cell types. Expression deconvolution methods can help differentiate these scenarios. BEDwARS is a Bayesian deconvolution method designed to address differences between reference signatures of cell types and corresponding true signatures underlying bulk transcriptomic profiles. BEDwARS is more robust to noisy reference signatures and outperforms leading in-class methods for estimating cell type proportions and signatures. Application of BEDwARS to dihydropyridine dehydrogenase deficiency identified the possible involvement of ciliopathy and impaired translational control in the etiology of the disorder.

Building insightful, memory-enriched models to capture long-time biochemical processes from short-time simulations

The ability to predict and understand complex molecular motions occurring over diverse timescales ranging from picoseconds to seconds and even hours in biological systems remains one of the largest challenges to chemical theory. Markov state models (MSMs), which provide a memoryless description of the transitions between different states of a biochemical system, have provided numerous important physically transparent insights into biological function. However, constructing these models often necessitates performing extremely long molecular simulations to converge the rates. Here, we show that by incorporating memory via the time-convolutionless generalized master equation (TCL-GME) one can build a theoretically transparent and physically intuitive memory-enriched model of biochemical processes with up to a three order of magnitude reduction in the simulation data required while also providing a higher temporal resolution. We derive the conditions under which the TCL-GME provides a more efficient means to capture slow dynamics than MSMs and rigorously prove when the two provide equally valid and efficient descriptions of the slow configurational dynamics. We further introduce a simple averaging procedure that enables our TCL-GME approach to quickly converge and accurately predict long-time dynamics even when parameterized with noisy reference data arising from short trajectories. We illustrate the advantages of the TCL-GME using alanine dipeptide, the human argonaute complex, and FiP35 WW domain.

ELULC-10, a 10 m European Land Use and Land Cover Map Using Sentinel and Landsat Data in Google Earth Engine

Land Use/Land Cover (LULC) maps can be effectively produced by cost-effective and frequent satellite observations. Powerful cloud computing platforms are emerging as a growing trend in the high utilization of freely accessible remotely sensed data for LULC mapping over large-scale regions using big geodata. This study proposes a workflow to generate a 10 m LULC map of Europe with nine classes, ELULC-10, using European Sentinel-1/-2 and Landsat-8 images, as well as the LUCAS reference samples. More than 200 K and 300 K of in situ surveys and images, respectively, were employed as inputs in the Google Earth Engine (GEE) cloud computing platform to perform classification by an object-based segmentation algorithm and an Artificial Neural Network (ANN). A novel ANN-based data preparation was also presented to remove noisy reference samples from the LUCAS dataset. Additionally, the map was improved using several rule-based post-processing steps. The overall accuracy and kappa coefficient of 2021 ELULC-10 were 95.38% and 0.94, respectively. A detailed report of the classification accuracies was also provided, demonstrating an accurate classification of different classes, such as Woodland and Cropland. Furthermore, rule-based post processing improved LULC class identifications when compared with current studies. The workflow could also supply seasonal, yearly, and change maps considering the proposed integration of complex machine learning algorithms and large satellite and survey data.

Quantum Asymmetry and Noisy Multimode Interferometry.

Quantum asymmetry is a physical resource that coincides with the amount of coherence between the eigenspaces of a generator responsible for phase encoding in interferometric experiments. We highlight an apparently counterintuitive behavior that the asymmetry may increase as a result of a decrease of coherence inside a degenerate subspace. We intuitively explain and illustrate the phenomena by performing a three-mode single-photon interferometric experiment, where one arm carries the signal and two noisy reference arms have fluctuating phases. We show that the source of the observed sensitivity improvement is the reduction of correlations between these fluctuations and comment on the impact of the effect when moving from the single-photon quantum level to the classical regime. Finally, we also establish the analogy of the effect in the case of entanglement resource theory.

Improvement of noisy images filtered by bilateral process using a multi-scale context aggregation network

Deep learning has recently received a lot of attention as a feasible solution to a variety of artificial intelligence difficulties. Convolutional neural networks (CNNs) outperform other deep learning architectures in the application of object identification and recognition when compared to other machine learning methods. Speech recognition, pattern analysis, and image identification, all benefit from deep neural networks. When performing image operations on noisy images, such as fog removal or low light enhancement, image processing methods such as filtering or image enhancement are required. The study shows the effect of using Multi-scale deep learning Context Aggregation Network CAN on Bilateral Filtering Approximation (BFA) for de-noising noisy CCTV images. Data-store is used tomanage our dataset, which is an object or collection of data that are huge to enter in memory, it allows to read, manage, and process data located in multiple files as a single entity. The CAN architecture provides integral deep learning layers such as input, convolution, back normalization, and Leaky ReLu layers to construct multi-scale. It is also possible to add custom layers like adaptor normalization (µ) and adaptive normalization (Lambda) to the network. The performance of the developed CAN approximation operator on the bilateral filtering noisy image is proven when improving both the noisy reference image and a CCTV foggy image. The three image evaluation metrics (SSIM, NIQE, and PSNR) evaluate the developed CAN approximation visually and quantitatively when comparing the created de-noised image over the reference image.Compared with the input noisy image, these evaluation metrics for the developed CAN de-noised image were (0.92673/0.76253, 6.18105/12.1865, and 26.786/20.3254) respectively

AutoGeneS: Automatic gene selection using multi-objective optimization for RNA-seq deconvolution.

Knowing cell-type proportions in a tissue is very important to identify which cells or cell types are targeted by a disease or perturbation. Hence, several deconvolution methods have been proposed to infer cell-type proportions from bulk RNA samples. Their performance with noisy reference profiles and closely correlated cell types highly depends on the set of genes undergoing deconvolution. In this work, we introduce AutoGeneS, a platform that automatically extracts discriminative genes and reveals the cellular heterogeneity of bulk RNA samples. AutoGeneS requires no prior knowledge about marker genes and selects genes by simultaneously optimizing multiple criteria: minimizing the correlation and maximizing the distance between cell types. AutoGeneS can be applied to reference profiles from various sources like single-cell experiments or sorted cell populations. Ground truth cell proportions analyzed by flow cytometry confirmed the accuracy of AutoGeneS in identifying cell-type proportions. AutoGeneS is available for use via a standalone Python package (https://github.com/theislab/AutoGeneS).

Data-Driven Caching With Users’ Content Preference Privacy in Information-Centric Networks

Information-centric networking (ICN) as an emerging networking paradigm has recently gained significant attention, due to the improvement of content delivery efficiency. The built-in network storage for caching is a key component in ICN to provide low latency service and reduce high backhaul traffic by caching popular content. However, users' content preference contains individual sensitive characteristics which is distinguishable from others. Therefore, in this work, we propose a data-driven caching revenue maximization problem with the considerations of users' local differential privacy. Specifically, we employ dBitFlip, a local differential privacy (LDP) mechanism, to locally add differential private noise to the users' preference content information. We leverage data-driven approach to predict the content popularity based on the reference distribution constructed by the reported noisy preference content data from users, mathematically present the distance between the noisy reference distribution and the true distribution by the tolerance level, and prove the relationship among the tolerance level, differential privacy budget and the confidence level. We provide feasible solutions to the proposed revenue maximization problem, and conduct simulations to show the effectiveness of the proposed scheme.

Performance Analysis of the Extended Binaural MVDR Beamformer With Partial Noise Estimation

Besides reducing undesired noise sources and limiting speech distortion, another important objective of a binaural noise reduction algorithm is the preservation of the binaural cues of all sound sources in the acoustic scene. In this paper, we consider the binaural minimum variance distortionless response beamformer with partial noise estimation (BMVDR-N), which allows to trade off between noise reduction performance and binaural cue preservation of the noise component by mixing the output signals of the BMVDR beamformer with the noisy reference microphone signals. For a directional noise source, it has been shown that incorporating an external microphone in addition to the head-mounted microphones enables both the noise reduction performance as well as the interaural time and level difference cues of the noise component to be improved in the output signals. In this paper, we consider an arbitrary noise field and analytically show that incorporating an external microphone in the BMVDR-N beamformer enables 1) a larger output signal-to-noise ratio (SNR) for the same mixing parameter, 2) the same output SNR for a larger mixing parameter, and 3) the same desired output magnitude squared coherence (MSC) of the noise component for a smaller mixing parameter to be obtained. The derived analytical expressions are firstly validated using simulated anechoic acoustic transfer functions, where the listener’s head is modelled as a rigid sphere. Experimental results using recorded signals for a binaural hearing device setup in a reverberant environment also show that in a realistic scenario incorporating an external microphone in the BMVDR-N beamformer significantly improves the output SNR and reduces the mixing parameter that is required to obtain a desired output MSC of the noise component compared to using only the head-mounted microphones.

Half2Half: deep neural network based CT image denoising without independent reference data

Reducing radiation dose of x-ray computed tomography (CT) and thereby decreasing the potential risk to patients are desirable in CT imaging. Deep neural network (DNN) has been proposed to reduce noise in low-dose CT (LdCT) images and showed promising results. However, most existing DNN-based methods require training a neural network using high-quality CT images as the reference. Lack of high-quality reference data has therefore been the bottleneck in the current DNN-based methods. Recently, a noise-to-noise (Noise2Noise) training method was proposed to train a denoising neural network with only noisy images. It has also been applied to LdCT data in both the count domain and image domain. However, the method still requires a separately acquired independent noisy reference image for supervising the training procedure. To address this limitation, we propose a novel method to generate both training inputs and training labels from the existing CT scans, which does not require any additional high-dose CT images or repeated scans. Therefore, existing large noisy dataset can be fully exploited for training a denoising neural network. Our experimental results show that the trained networks can reduce noise in existing CT image and hence improve the image quality for clinical diagnosis.

LEARNING FROM NOISY SAMPLES FOR MAN-MADE IMPERVIOUS SURFACE MAPPING

Abstract. Man-made impervious surfaces, indicating the human footprint on Earth, are an environmental concern because it leads to a chain of events that modifies urban air and water resources. To better map man-made impervious surfaces in any region of interest (ROI), we propose a framework for learning to map impervious areas in any ROIs from Sentinel-2 images with noisy reference data, using a pre-trained fully convolutional network (FCN). The FCN is first trained with reference data only available in Europe, which is able to provide reasonable mapping results even in areas outside of Europe. The proposed framework, aiming to achieve an improvement over the preliminary predictions for a specific ROI, consists of two steps: noisy training data pre-processing and model fine-tuning with robust loss functions. The framework is validated over four test areas located in different continents with a measurable improvement over several baseline results. It has been shown that a better impervious mapping result can be achieved through a simple fine-tuning with noisy training data, and label updating through robust loss functions allows to further enhance the performances. In addition, by analyzing and comparing the mapping results to baselines, it can be highlighted that the improvement is mainly coming from a decreased omission error. This study can also provide insights for similar tasks, such as large-scale land cover/land use classification when accurate reference data is not available for training.

Target Detection in Passive Radar Under Noisy Reference Channel: A New Threshold-Setting Strategy

In the detection theory framework, it is customary to assign a bound to the false alarm probability and to attempt to maximize the detection probability subject to this constraint. In the problem of moving target detection in passive radar with a noisy reference channel, we formulate a detection problem as a composite hypothesis-testing problem and solve it with the likelihood ratio test (LRT) principle, which is known as generalized LRT in the electrical engineering works of literature. In such a problem, we show that any uncertainty in the value of the direct signal-to-noise ratio of the reference channel, abbreviated as $\text{DNR}_r$ , can result in excessive false alarm probability of the proposed noisy-reference-channel-based detector in the low- $\text{DNR}_r$ regime. To facilitate efficient operation under uncertainty in $\text{DNR}_r$ , we propose a new threshold-setting strategy to adjust the level of the proposed detector. Through extensive Monte–Carlo simulations, we examine the above problem and investigate the efficiency of the proposed threshold-setting strategy. Besides, we apply the framework of the kernel theory to the target detection problem of a noisy and ideal reference channel passive radar to propose two new detectors. As such, we replace the inner products of the proposed tests with appropriate polynomial kernel functions allowing for richer feature space to be deployed in the detection, achieving better detection performance. In this case, our detection performance results show that the kernelized detectors offer more that 1-dB signal-to-noise ratio gain as compared to their conventional counterparts.

Non-stationary Gaussian process regression applied in validation of vehicle dynamics models

This work compares methods to compute confidence bands in a validation task of a vehicle single-track model. The confidence bands are computed from time series by naïve method, Gaussian process regression and heteroscedastic and non-stationary Gaussian process regression. The simulation model considers the epistemic uncertainty of the vehicle mass parameter by Latin hypercube sampling. The validation procedure compares all stochastically simulated time series of the vehicle yaw rate with the confidence band of the reference data. The model is marked as valid if the yaw rate for each time step is within the confidence band of the reference data. The data was challenging due to noise and time-varying variance and smoothness. Due to required data pre-processing and the high sensitivity to noise in the reference data, the naïve method has generated unusable confidence bands and cannot be recommended for similar validation tasks. Gaussian process regression solved the problem of noise sensitivity, but was not able to model the time-varying length scale of the reference data. Therefore, heteroscedastic and non-stationary Gaussian process regression is proposed to calculate accurate confidence bands of time-varying and noisy reference data for the validation of dynamic models by a confidence band approach.

Performance bounds for two-channel delay-doppler estimation using unknown waveforms

In this paper, we derive Cramer-Rao Lower Bound (CRLB) expressions for two-channel radar target delay-Doppler estimation with unknown transmit waveforms. We have access to these unknown waveforms only via the noisy reference measurements. Further, we assume that the waveform source transmits periodic pulses as is common for most pulsed radar illuminators. Recent CRLB literature for radar has focussed on specific types of illuminators and derived the CRLB assuming these specific transmitted waveforms are known and incorporating them directly into the measurement model. Further, some papers while assuming the waveform itself to be unknown, make simplifying statistical assumptions on the waveforms or the waveform structure. Contrary to this, we compute a more general result that is not restricted to any specific waveform and we also include the noisy reference channel measurements in our derivation of the CRLB to obtain a more accurate performance analysis. The only assumption we make is on the periodicity of the transmit signal, which is true in practice. We demonstrate the CRLB using extensive numerical simulations both in the presence and absence of phase synchronization between the transmitter and receiver.

Binaural LCMV Beamforming With Partial Noise Estimation

Besides reducing undesired sources (interfering sources and background noise), another important objective of a binaural beamforming algorithm is to preserve the spatial impression of the acoustic scene, which can be achieved by preserving the binaural cues of all sound sources. While the binaural minimum variance distortionless response (BMVDR) beamformer provides a good noise reduction performance and preserves the binaural cues of the desired source, it does not allow to control the reduction of the interfering sources and distorts the binaural cues of the interfering sources and the background noise. Hence, several extensions have been proposed. First, the binaural linearly constrained minimum variance (BLCMV) beamformer uses additional constraints, enabling to control the reduction of the interfering sources while preserving their binaural cues. Second, the BMVDR with partial noise estimation (BMVDR-N) mixes the output signals of the BMVDR with the noisy reference microphone signals, enabling to control the binaural cues of the background noise. Merging the advantages of both extensions, in this paper we propose the BLCMV with partial noise estimation (BLCMV-N). We show that the output signals of the BLCMV-N can be interpreted as a mixture of the noisy reference microphone signals and the output signals of a BLCMV using an adjusted interference scaling parameter. We provide a theoretical comparison between the BMVDR, the BLCMV, the BMVDR-N and the proposed BLCMV-N in terms of noise and interference reduction performance and binaural cue preservation. Experimental results using recorded signals as well as the results of a perceptual listening test show that the BLCMV-N is able to preserve the binaural cues of an interfering source (like the BLCMV), while enabling to trade off between noise reduction performance and binaural cue preservation of the background noise (like the BMVDR-N).

Average BER and Noisy Reference Loss of Partially Coherent PSK Demodulation Over Shadowed Multipath Fading Channel

In this correspondence paper, we present novel results related to the error performance of coherent systems in the presence of the imperfect carrier phase estimation over Gamma shadowed Nakagami- $m$ fading environment. Taking the Tikhonov-distributed phase error process into account, binary and quaternary phase-shift keying signal detection is analyzed. Based on the Maclaurin series expansion of the complementary error function, approximated closed-form expressions for the bit error rate (BER) are derived. These expressions are simplified to those already reported in literature, which are valid in the case without shadowing. The asymptotic closed-form expressions for noisy reference loss are also presented. We examine the effect of phase-locked loop parameters on both BER and noisy reference loss. Analytical results are validated by independent Monte Carlo simulations.

Passive Radar Detection With Noisy Reference Channel Using Principal Subspace Similarity

Traditional passive radar detectors compute cross correlation of the raw data in the reference and surveillance channels. However, there is no optimality guarantee for this detector in the presence of a noisy reference. Here, we develop a new detector that utilizes a test statistic based on the cross correlation of the principal left singular vectors of the reference and surveillance signal-plus-noise matrices. This detector offers better performance by exploiting the inherent low-rank structure when the transmitted signals are a weighted periodic summation of several identical waveforms (amplitude and phase modulation), as is the case with commercial digital illuminators as well as noncooperative radar. We consider a scintillating target. We provide analytical detection performance guarantees establishing signal-to-noise ratio thresholds above which the proposed detection statistic reliably discriminates, in an asymptotic sense, the signal versus no-signal hypothesis. We validate these results using extensive numerical simulations. We demonstrate the “near constant false alarm rate (CFAR)” behavior of the proposed detector with respect to a fixed, SNR-independent threshold and contrast that with the need to adjust the detection threshold in an SNR-dependent manner to maintain CFAR for other detectors found in the literature. Extensions of the proposed detector for settings applicable to orthogonal frequency division multiplexing (OFDM), adaptive radar are discussed.