Imbalance Tolerance Research Articles

Joint carrier phase tracking and IQ imbalance tolerance with the dynamic adaption of reference constellation for probabilistic shaped signals

Joint carrier phase tracking and IQ imbalance tolerance with the dynamic adaption of reference constellation for probabilistic shaped signals

Lnclocator-imb: An Imbalance-tolerant Ensemble Deep Learning Framework for Predicting Long Non-coding RNA Subcellular Localization.

Recent studies have highlighted the critical roles of long non-coding RNAs (lncRNAs) in various biological processes, including but not limited to dosage compensation, epigenetic regulation, cell cycle regulation, and cell differentiation regulation. Consequently, lncRNAs have emerged as a central focus in genetic studies. The identification of the subcellular localization of lncRNAs is essential for gaining insights into crucial information about lncRNA interaction partners, post- or co-transcriptional regulatory modifications, and external stimuli that directly impact the function of lncRNA. Computational methods have emerged as a promising avenue for predicting the subcellular localization of lncRNAs. However, there is a need for additional enhancement in the performance of current methods when dealing with unbalanced data sets. To address this challenge, we propose a novel ensemble deep learning framework, termed lncLocator-imb, for predicting the subcellular localization of lncRNAs. To fully exploit lncRNA sequence information, lncLocator-imb integrates two base classifiers, including convolutional neural networks (CNN) and gated recurrent units (GRU). Additionally, it incorporates two distinct types of features, including the physicochemical pattern feature and the distributed representation of nucleic acids feature. To address the problem of poor performance exhibited by models when confronted with unbalanced data sets, we utilize the label-distribution-aware margin (LDAM) loss function during the training process. Compared with traditional machine learning models and currently available predictors, lncLocator-imb demonstrates more robust category imbalance tolerance. Our study proposes an ensemble deep learning framework for predicting the subcellular localization of lncRNAs. Additionally, a novel approach is presented for the management of different features and the resolution of unbalanced data sets. The proposed framework exhibits the potential to serve as a significant resource for various sequence-based prediction tasks, providing a versatile tool that can be utilized by professionals in the fields of bioinformatics and genetics.

Data-Driven Detection of Stealthy False Data Injection Attack Against Power System State Estimation

Power system state estimation (PSSE) is the foundation of energy management system applications. Hence, operators impose stringent requirements on PSSE data integrity. False data injection attacks (FDIAs) can cause risks to PSSE data-driven operations and demand mitigation. In this article, we present a two-step FDIA detector design. In step one, we study a novel stealthy attack policy by simultaneously considering the attacker’s cost reduction and damage production. In step two, with the aid of a deep autoencoding Gaussian mixture model (DAGMM), we design an unsupervised detection scheme to detect the stealthy attack. The DAGMM-based detector can meet the requirement of rapidity, unsupervisedness, and data imbalance tolerance. Eventually, we simulate and validate the stealthy attack policy and the corresponding detector using the benchmark IEEE 39-bus and 118-bus systems.

Combining high-power heterogeneous RF sources for accelerator applications

Facilities with a synchrotron light source at present utilize both vacuum-tube and solid-state RF sources to accelerate an electron beam, thanks to the rapid development of semiconductor technology. In NSRRC, a retired 500-MHz 100-kW klystron-based RF transmitter and a newly developed 80-kW solid-state RF transmitter became available for experimental study. We proposed the usage of a 3-dB waveguide hybrid coupler to combine these two heterogeneous RF sources to provide twice the RF-power of each alone. The intrinsic difference in RF properties of these two RF-power sources seen by the 3-dB hybrid coupler is compensated with an analogue low-level feedback system that locks the klystron side to the solid-state RF source side for a quasi-equal power level and a quadrature phase offset to obtain the greatest efficiency of the combination. In this manner, we demonstrated an efficiency up to 98.3% for combined RF-power 140 kW measured at the target port of a 3-dB waveguide hybrid coupler; the dummy load after the non-target port received negligible power within the full power range. The robustness to a power imbalance between these two RF-power sources was studied theoretically and experimentally; the efficiency of the RF-power combination was maintained greater than 90% for an available power ratio of these two RF sources up to 3 or phase error within ±45°. The results demonstrate an economical power-combination scheme for heterogeneous RF-power sources with high tolerance of imbalance for a symmetric power combiner. The usage of an asymmetric power combiner is foreseen to extend the imbalance tolerance.

Architectures and accuracy of artificial neural network for disease classification from omics data

BackgroundDeep learning has made tremendous successes in numerous artificial intelligence applications and is unsurprisingly penetrating into various biomedical domains. High-throughput omics data in the form of molecular profile matrices, such as transcriptomes and metabolomes, have long existed as a valuable resource for facilitating diagnosis of patient statuses/stages. It is timely imperative to compare deep learning neural networks against classical machine learning methods in the setting of matrix-formed omics data in terms of classification accuracy and robustness.ResultsUsing 37 high throughput omics datasets, covering transcriptomes and metabolomes, we evaluated the classification power of deep learning compared to traditional machine learning methods. Representative deep learning methods, Multi-Layer Perceptrons (MLP) and Convolutional Neural Networks (CNN), were deployed and explored in seeking optimal architectures for the best classification performance. Together with five classical supervised classification methods (Linear Discriminant Analysis, Multinomial Logistic Regression, Naïve Bayes, Random Forest, Support Vector Machine), MLP and CNN were comparatively tested on the 37 datasets to predict disease stages or to discriminate diseased samples from normal samples. MLPs achieved the highest overall accuracy among all methods tested. More thorough analyses revealed that single hidden layer MLPs with ample hidden units outperformed deeper MLPs. Furthermore, MLP was one of the most robust methods against imbalanced class composition and inaccurate class labels.ConclusionOur results concluded that shallow MLPs (of one or two hidden layers) with ample hidden neurons are sufficient to achieve superior and robust classification performance in exploiting numerical matrix-formed omics data for diagnosis purpose. Specific observations regarding optimal network width, class imbalance tolerance, and inaccurate labeling tolerance will inform future improvement of neural network applications on functional genomics data.

Simulation Studies of a Trench MOS Device Structure with Small Figures of Merit

We proposed a vertical high permittivity trench power MOS (HKTMOS) device with alternating N&P drift region and high permittivity (HK) trench sandwiched in between. The unique structure guarantees uniform potential distribution for wide voltage range at block state owing to both HK potential modulation effect and superjunction (SJ) charge balance. The specific on-resistance (Rons) of HKTMOS is in orders of magnitude lower than the SJ counterparts at the on state because of the strong accumulation effect brought by HK trench. Although the gate charge also significantly rises because of the accumulation, the figures of merit (FOM) of HKTMOS still reduces considerably than the SJ under same BV. An expression for FOM is derived demonstrating that the FOM of HKTMOS is proportional to the square of HK trench depth, which agrees on with simulation results well. The simulation results indicate that within the BV range of 500~2000V, the Rons and FOM of HKTMOS are in 1~2 orders of magnitude lower and 17.4%~44.1% of SJ, respectively under the same BV condition. Furthermore, HKTMOS also demonstrates better charge imbalance tolerance than SJ.

PMU-Based Detection of Voltage Imbalances with Tolerance Constraints

The problem of voltage imbalance detection in a three-phase power system using phasor measurement unit (PMU) data is considered within a hypothesis testing framework. A general model for the PMU output with a downsampled negative sequence is presented. The new formulation takes into account computation complexity and tolerance of imbalance. A new monitoring tool, referred to as a generalized locally most powerful (GLMP) test, is proposed for detection in the presence of nuisance parameters. A closed-form expression of the GLMP test is developed for the detection of an imbalance problem, based on a single PMU's measurements. Numerical simulations show improved performance over benchmark techniques.

Which design is better? Ehrenfest urn versus biased coin

Two features are desired in designing a sequential clinical trial: randomness and balance. The former makes the ground for valid statistical inferences and the latter strengthens efficiency in inference procedures. Unfortunately randomness and balance can be in conflict with one another, and clinicians may be caught between the need for both of them. This paper raises an interesting question: can one design consistently achieve more balance than another when both designs own the same randomness? The Ehrenfest urn design is presented to allocate two treatments under a sequential clinical trial, and its balance and randomness properties are investigated. The design is compared with the biased coin design with imbalance tolerance.

Which design is better? Ehrenfest urn versus biased coin

Two features are desired in designing a sequential clinical trial: randomness and balance. The former makes the ground for valid statistical inferences and the latter strengthens efficiency in inference procedures. Unfortunately randomness and balance can be in conflict with one another, and clinicians may be caught between the need for both of them. This paper raises an interesting question: can one design consistently achieve more balance than another when both designs own the same randomness? TheEhrenfest urn designis presented to allocate two treatments under a sequential clinical trial, and its balance and randomness properties are investigated. The design is compared with thebiased coin design with imbalance tolerance.

Biased coin design with imbalance tolerance

One of the main aspects of a sampling procedure is to determine how to collect samples. A completely random sampling scheme is free of any bias and provides a basis for valid statistical inferences. A balanced sampling scheme strengthens efficiency in statistical inference procedures. This paper presents a sampling schemebiased coin design with imbalance tolerance, which enforces balance in treatment allocations in sequential clinical trials. The design synthesizes Efron's pioneering work on biased coin design (1971) and Soares and Wu's big stick design(1983). The underlying structure of the design is a Markov chain. An explicit formula of the high-order transition probabilities of the chain is obtained. This paper applies the formula to give exact assessments of randomness, balance, and the trade-off between them under the design