Search for: All records

In recent years, neural networks (NNs) have been embraced by several scientific and engineering disciplines for diverse modeling and inferencing applications. The importance of quantifying the confidence in NN predictions has escalated due to the increasing adoption of these decision models. Nevertheless, conventional NN do not furnish uncertainty estimates associated with their predictions and are therefore ill-calibrated. Uncertainty quantification techniques offer probability distributions or CIs to represent the uncertainty associated with NN predictions, instead of solely presenting the point predictions/estimates. Once the uncertainty in NN is quantified, it is crucial to leverage this information to modify training objectives and improve the accuracy and reliability of the corresponding decision models. This work presents a novel framework to utilize the knowledge of input and output uncertainties in NN to guide querying process in the context of Active Learning. We also derive the lower and upper bounds for label complexity. The efficacy of the proposed framework is established by conducting experiments across classification and regression tasks. 

Time series forecasting is an important application in various domains such as energy management, traffic planning, financial markets, meteorology, and medicine. However, real-time series data often present intricate temporal variability and sharp fluctuations, which pose significant challenges for time series forecasting. Previous models that rely on 1D time series representations usually struggle with complex temporal variations. To address the limitations of 1D time series, this study introduces the Times2D method that transforms the 1D time series into 2D space. Times2D consists of three main parts: first, a Periodic Decomposition Block (PDB) that captures temporal variations within a period and between the same periods by converting the time series into a 2D tensor in the frequency domain. Second, the First and Second Derivative Heatmaps (FSDH) capture sharp changes and turning points, respectively. Finally, an Aggregation Forecasting Block (AFB) integrates the output tensors from PDB and FSDH for accurate forecasting. This 2D transformation enables the utilization of 2D convolutional operations to effectively capture long and short characteristics of the time series. Comprehensive experimental results across large-scale data in the literature demonstrate that the proposed Times2D model achieves state-of-the-art performance in both short-term and long-term forecasting. 

Ovarian cancer survival depends strongly on the time of diagnosis. Detection at stage 1 must be the goal of liquid biopsies for ovarian cancer detection. We report the development and validation of graphene-based optical nanobiosensors (G-NBSs) that quantify the activities of a panel of proteases, which were selected to provide a crowd response that is specific for ovarian cancer. These G-NBSs consist of few-layer explosion graphene featuring a hydrophilic coating, which is linked to fluorescently labeled highly selective consensus sequences for the proteases of interest, as well as a fluorescent dye. The panel of G-NBSs showed statistically significant differences in protease activities when comparing localized (early-stage) ovarian cancer with both metastatic (late-stage) and healthy control groups. A hierarchical framework integrated with active learning (AL) as a prediction and analysis tool for early-stage detection of ovarian cancer was implemented, which obtained an overall accuracy score of 94.5%, with both a sensitivity and specificity of 0.94. 

A composite detection technique against stealthy data manipulations is developed in this paper for distribution networks that are low observable. Attack detection strategies typically rely on state estimation which becomes challenging when limited measurements are available. In this paper, a modified matrix completion approach provides estimates of the system state and its error variances for the locations in the network where measurements are unavailable. Using the error statistics and their corresponding state estimates, bad data detection can be carried out using the chi-squared test. The proposed approach employs a moving target defence strategy (MTD) where the network parameters are perturbed through distributed flexible AC transmission system (D-FACTS) devices such that stealthy data manipulation attacks can be exposed in the form of bad data. Thus, the bad data detection approach developed in this paper can detect stealthy attacks using the MTD strategy. This technique is implemented on 37-bus and 123-bus three-phase unbalanced distribution networks to demonstrate the attack detection accuracy even for a low observable system.