Modeling time series (TS) is a research focus in cryospheric sciences because of the complexity and multiscale nature of events of interest. Highly nonuniform sampling of measurements from different sensors with different levels of accuracy, as is typical for measurements of ice sheet elevations, makes the problem even more challenging. In this article, we propose a spline-based approximation framework (ALPS--Approximation by Localized Penalized Splines) for modeling TS of land ice changes. The localized support of the B-spline basis functions enable robustness to nonuniform sampling, a considerable improvement over other global and piecewise local models. With features like discrete-coordinate-difference-based penalization and two-level outlier detection, ALPS further guarantees the stability and quality of approximations. ALPS incorporates rigorous model uncertainty estimates with all approximations. As demonstrated by examples, ALPS performs well for a variety of data sets, including TS of ice sheet thickness, elevation, velocity, and terminus locations. The robust estimation of TS and their derivatives facilitates new applications, such as the reconstruction of high-resolution elevation change records by fusing sparsely sampled TS of ice sheet thickness changes with modeled firn thickness changes, and the analysis of the relationship between different outlet glacier observations to gain new insights into processes and forcing.
This paper describes a hierarchical learning strategy for generating sparse representations of multivariate datasets. The hierarchy arises from approximation spaces considered at successively finer scales. A detailed analysis of stability, convergence and behavior of error functionals associated with the approximations are presented, along with a well chosen set of applications. Results show the performance of the approach as a data reduction mechanism for both synthetic (univariate and multivariate) and a real dataset (geo-spatial). The sparse representation generated is shown to efficiently reconstruct data and minimize error in prediction. The approach is also shown to generalize well to unseen samples, extending its prospective application to statistical learning problems.
As IoT (internet of things) devices grow in prominence, safeguarding them from cyberattacks is becoming a pressing challenge. To bootstrap IoT security, device identification or authentication is crucial for establishing trusted connections among devices without prior trust. In this regard, radio frequency fingerprinting (RFF) is gaining attention because it is more efficient and requires fewer computational resources compared to resource-intensive cryptographic methods, such as digital signatures. RFF works by identifying unique manufacturing defects in the radio circuitry of IoT devices by analyzing over-the-air signals that embed these imperfections, allowing for the identification of the transmitting hardware. Recent studies on RFF often leverage advanced classification models, including classical machine learning techniques such as K-Nearest Neighbor (KNN) and Support Vector Machine (SVM), as well as modern deep learning architectures like Convolutional Neural Network (CNN). In particular, CNNs are well-suited as they use multidimensional mapping to detect and extract reliable fingerprints during the learning process. However, a significant limitation of these approaches is that they require large datasets and necessitate retraining when new devices not included in the initial training set are added. This retraining can cause service interruptions and is costly, especially in large-scale IoT networks. In this paper, we propose a novel solution to this problem: RFF using Siamese networks, which eliminates the need for retraining and allows for seamless authentication in IoT deployments. The proposed Siamese network is trained using in-phase and quadrature (I/Q) samples from 10 different Software-Defined Radios (SDRs). Additionally, we present a new algorithm, the Similarity-Based Embedding Classification (SBEC) for RFF. We present experimental results that demonstrate that the Siamese network effectively distinguishes between malicious and trusted devices with a remarkable 98% identification accuracy.
Metastatic tumors have poor prognoses for progression-free and overall survival for all cancer patients. Rare circulating tumor cells (CTCs) and rarer circulating tumor cell clusters (CTCCs) are potential biomarkers of metastatic growth, with CTCCs representing an increased risk factor for metastasis. Current detection platforms are optimized for ex vivo detection of CTCs only. Microfluidic chips and size exclusion methods have been proposed for CTCC detection; however, they lack in vivo utility and real-time monitoring capability. Confocal backscatter and fluorescence flow cytometry (BSFC) has been used for label-free detection of CTCCs in whole blood based on machine learning (ML) enabled peak classification. Here, we expand to a deep-learning (DL) -based, peak detection and classification model to detect CTCCs in whole blood data. We demonstrate that DL-based BSFC has a low false alarm rate of 0.78 events/min with a high Pearson correlation coefficient of 0.943 between detected events and expected events. DL-based BSFC of whole blood maintains a detection purity of 72% and a sensitivity of 35.3% for both homotypic and heterotypic CTCCs starting at a minimum size of two cells. We also demonstrate through artificial spiking studies that DL-based BSFC is sensitive to changes in the number of CTCCs present in the samples and does not add variability in detection beyond the expected variability from Poisson statistics. The performance established by DL-based BSFC motivates its use for in vivo detection of CTCCs. Further developments of label-free BSFC to enhance throughput could lead to critical applications in the clinical detection of CTCCs and ex vivo isolation of CTCC from whole blood with minimal disruption and processing steps.
Federated Learning (FL) enables collaborative model training while preserving data privacy by keeping raw data locally stored on client devices, preventing access from other clients or the central server. However, recent studies reveal that sharing model gradients creates vulnerability to Model Inversion Attacks, particularly Deep Leakage from Gradients (DLG), which reconstructs private training data from shared gradients. While Homomorphic Encryption has been proposed as a promising defense mechanism to protect gradient privacy, fully encrypting all model gradients incurs high computational overhead. Selective encryption approaches aim to balance privacy protection with computational efficiency by encrypting only specific gradient components. However, the existing literature largely overlooks a theoretical exploration of the spectral behavior of encrypted versus unencrypted parameters, relying instead primarily on empirical evaluations. To address this gap, this paper presents a framework for theoretical analysis of the underlying principles of selective encryption as a defense against model inversion attacks. We then provide a comprehensive empirical study that identifies and quantifies the critical factors, such as model complexity, encryption ratios, and exposed gradients, that influence defense effectiveness. Our theoretical framework clarifies the relationship between gradient selection and privacy preservation, while our experimental evaluation demonstrates how these factors shape the robustness of defenses against model inversion attacks. Collectively, these contributions advance the understanding of selective encryption mechanisms and offer principled guidance for designing efficient, scalable, privacy-preserving federated learning systems.