Size Of Dataset Research Articles

Estimation of minimal data sets sizes for machine learning predictions in digital mental health interventions

Artificial intelligence promises to revolutionize mental health care, but small dataset sizes and lack of robust methods raise concerns about result generalizability. To provide insights on minimal necessary data set sizes, we explore domain-specific learning curves for digital intervention dropout predictions based on 3654 users from a single study (ISRCTN13716228, 26/02/2016). Prediction performance is analyzed based on dataset size (N = 100–3654), feature groups (F = 2–129), and algorithm choice (from Naive Bayes to Neural Networks). The results substantiate the concern that small datasets (N ≤ 300) overestimate predictive power. For uninformative feature groups, in-sample prediction performance was negatively correlated with dataset size. Sophisticated models overfitted in small datasets but maximized holdout test results in larger datasets. While N = 500 mitigated overfitting, performance did not converge until N = 750–1500. Consequently, we propose minimum dataset sizes of N = 500–1000. As such, this study offers an empirical reference for researchers designing or interpreting AI studies on Digital Mental Health Intervention data.

Global tropical cyclone size and intensity reconstruction dataset for 1959–2022 based on IBTrACS and ERA5 data

Abstract. Tropical cyclones (TCs) are powerful weather systems that can cause extreme disasters. The International Best Track Archive for Climate Stewardship (IBTrACS) dataset provides widely used data to estimate TC climatology. However, it has low data coverage, lacking intensity and outer-size data for more than half of all recorded storms, and is therefore insufficient as a reference for researchers and decision makers. To fill this data gap, we reconstruct a long-term TC dataset by integrating IBTrACS and European Centre for Medium-Range Weather Forecasts Reanalysis 5 (ERA5) data. This reconstructed dataset covers the period 1959–2022, with 3 h temporal resolution. Compared to the IBTrACS dataset, it contains approximately 3–4 times more data points per characteristic. We establish machine learning models to estimate the maximum sustained wind speed (Vmax) and radius of maximum wind (Rmax) in six basins for which TCs are generated, using ERA5-derived 10 m azimuthal mean azimuthal wind profiles as input, with Vmax and Rmax data from the IBTrACS dataset used as learning target data. Furthermore, we employ an empirical wind–pressure relationship and six wind profile models to estimate the minimum central pressure (Pmin) and outer size of the TCs, respectively. Overall, this high-resolution TC reconstruction dataset demonstrates global consistency with observations, exhibiting mean biases of <1 % for Vmax and 3 % for Rmax and Pmin in almost all basins. The dataset is publicly available from https://doi.org/10.5281/zenodo.13919874 (Xu et al., 2024) and substantially advances our understanding of TC climatology, thereby facilitating risk assessments and defenses against TC-related disasters.

Comparative Analysis of Conventional CNN v’s ImageNet Pretrained ResNet in Medical Image Classification

Convolutional Neural Networks (CNNs) are the prevalent technology in computer vision and have become increasingly popular for medical imaging data classification and analysis. In this field, due to the scarcity of medical data, pretrained ResNets on ImageNet can be considered a suitable first approach. This paper examines the medical imaging classification accuracy of conventional basic custom CNNs compared to ImageNet pretrained ResNets on various medical datasets in an effort to give more information about the importance of medical data and its preprocessing techniques for disease studies. Microscope-extracted cytological images were examined along with chest X-rays, MRI brain scans, and melanoma photographs. The medical images were examined in various sets, class combinations, and resolutions. Augmented image datasets and asymmetrical training and validation splits among the classes were also examined. Models were developed after they were tested and fine-tuned with respect to their network size, parameter values and network methods, image resolution, size of dataset, multitude, and genre of class. Overfitting was also examined, and comparative studies regarding the computational cost of different models were performed. The models achieved high accuracy in image classification that varies depending on the dataset and can be easily incorporated in future over-the-internet medical decision-supporting (telemedicine) environments. In addition, it appeared that conventional basic custom CNN overperformed ImageNet pretrained ResNets. The obtained results indicate the importance of utilizing medical image data as a testbed for improvements in CNN classification performance and the possibility of using CNNs and data preprocessing techniques for disease studies.

Automated segmentation of dental restorations using deep learning: exploring data augmentation techniques.

Deep learning has revolutionized image analysis for dentistry. Automated segmentation of dental radiographs is of great importance towards digital dentistry. The performance of deep learning models heavily relies on the quality and diversity of the training data. Data augmentation is a widely used technique implemented in machine learning and deep learning to artificially increase the size and diversity of a training dataset by applying various transformations to the original data. This work aims to automatically segment implants, prostheses, and fillings in panoramic images using 9 different deep learning segmentation models. Later, it explores the effect of data augmentation methods on segmentation performance of the models. Eight different data augmentation techniques are examined. Performance is evaluated by well-accepted metrics such as intersection over union (IoU) and Dice coefficient. While averaging the segmentation results for the three classes, IoU varies between 0.62 and 0.82 while Dice score is between 0.75 and 0.9 among deep learning models used. Augmentation techniques provided performance improvements of up to 3.37%, 5.75% and 8.75% for implant, prosthesis and filling classes, respectively. Findings reveal that choosing optimal augmentation strategies depends on both model architecture and dental structure type.

A Unified Graph Theory Approach: Clustering and Learning in Criminal Data

Crime report clustering plays a critical role in modern law enforcement, enabling the identification of patterns and trends essential for proactive policing. However, traditional clustering approaches face significant challenges with the complex, unstructured nature of crime reports and their inherent sparse relationships. While graph-based clustering shows promise, issues of noise sensitivity and data sparsity persist. This study introduces a unified approach integrating spectral graph-based clustering with Graph Convolutional Networks (GCN) to address these challenges. The proposed approach encompasses data collection, preprocessing, linguistic feature extraction, vectorization, graph construction, graph learning, and clustering to effectively capture the intricate similarities between crime reports. The proposed approach achieved significant improvements over existing methods: a Silhouette Score of 0.77, a Davies–Bouldin Index of 0.51, and consistent performance across varying dataset sizes (100–1000 nodes). These results demonstrate the potential for enhanced crime pattern detection in law enforcement operations.

Separation and Dynamic Binding Mechanism of 3D Model Animation by Integrating Improved Genetic Algorithm and WebAR

With the development of the economy, the demand for 3D model animation in fields such as games, film and television, and education is constantly increasing. At present, many 3D model animations are built based on genetic algorithm. However, due to the slow convergence speed and poor local search ability of genetic algorithm, this model still has many shortcomings. Therefore, a cellular automaton is used to optimize the genetic algorithm, and a cellular genetic algorithm is proposed. It is integrated with WebAR technology, which has the characteristics of lightweight and low cost, aiming to improve the separation and dynamic binding mechanism of 3D model animation. Experimental analysis showed that when the number of chromosome genes was 90, the vertex gene value of the research algorithm was the highest, ranging from 281 to 501, and the convergence was the highest. The file size of the model dataset based on the research method was reduced by 50% compared with before. The FPDT based on the animation separation and dynamic binding mechanism was decreased by 15.2 s on average. To sum up, it can be seen that the designed 3D model animation is more detailed and the animation dynamic is also the best. This research method can significantly improve the separation and dynamic binding mechanism of 3D model animation. This research will improve the integration of genetic algorithm and WebAR technology, which can not only effectively deal with complex 3D models and animations, but also has broad application prospects.

An Optimised CNN Hardware Accelerator Applicable to IoT End Nodes for Disruptive Healthcare

In the evolving landscape of computer vision, the integration of machine learning algorithms with cutting-edge hardware platforms is increasingly pivotal, especially in the context of disruptive healthcare systems. This study introduces an optimized implementation of a Convolutional Neural Network (CNN) on the Basys3 FPGA, designed specifically for accelerating the classification of cytotoxicity in human kidney cells. Addressing the challenges posed by constrained dataset sizes, compute-intensive AI algorithms, and hardware limitations, the approach presented in this paper leverages efficient image augmentation and pre-processing techniques to enhance both prediction accuracy and the training efficiency. The CNN, quantized to 8-bit precision and tailored for the FPGA’s resource constraints, significantly accelerates training by a factor of three while consuming only 1.33% of the power compared to a traditional software-based CNN running on an NVIDIA K80 GPU. The network architecture, composed of seven layers with excessive hyperparameters, processes downscale grayscale images, achieving notable gains in speed and energy efficiency. A cornerstone of our methodology is the emphasis on parallel processing, data type optimization, and reduced logic space usage through 8-bit integer operations. We conducted extensive image pre-processing, including histogram equalization and artefact removal, to maximize feature extraction from the augmented dataset. Achieving an accuracy of approximately 91% on unseen images, this FPGA-implemented CNN demonstrates the potential for rapid, low-power medical diagnostics within a broader IoT ecosystem where data could be assessed online. This work underscores the feasibility of deploying resource-efficient AI models in environments where traditional high-performance computing resources are unavailable, typically in healthcare settings, paving the way for and contributing to advanced computer vision techniques in embedded systems.

Neural operators for hydrodynamic modeling of underground gas storage facilities

Objectives. Much of the research in deep learning has focused on studying mappings between finite-dimensional spaces. While hydrodynamic processes of gas filtration in underground storage facilities can be described by partial differential equations (PDE), the requirement to study the mappings between functional spaces of infinite dimension distinguishes this problem from those solved using traditional mapping approaches. One of the most promising approaches involves the construction of neural operators, i.e., a generalization of neural networks to approximate mappings between functional spaces. The purpose of the work is to develop a neural operator to speed up calculations involved in hydrodynamic modeling of underground gas storages (UGS) to an acceptable degree of accuracy.Methods. In this work, a modified Fourier neural operator was built and trained for hydrodynamic modeling of gas filtration processes in underground gas storages.Results. The described method is shown to be capable of successful application to problems of three-dimensional gas filtration in a Cartesian coordinate system at objects with many wells. Despite the use of the fast Fourier transform algorithm in the architecture, the developed model is also effective for modeling objects having a nonuniform grid and complex geometry. As demonstrated not only on the test set, but also on artificially generated scenarios with significant changes made to the structure of the modeled object, the neural operator does not require a large training dataset size to achieve high accuracy of approximation of PDE solutions. A trained neural operator can simulate a given scenario in a fraction of a second, which is at least 106 times faster than a traditional numerical simulator.Conclusions. The constructed and trained neural operator demonstrated efficient hydrodynamic modeling of underground gas storages. The resulting algorithm reproduces adequate solutions even in the case of significant changes in the modeled object that had not occurred during the training process. The model can be recommended for use in planning and decision-making purposes regarding various aspects of UGS operation, such as optimal control of gas wells, pressure control, and management of gas reserves.

Model Parameter Prediction Method for Accelerating Distributed DNN Training

As the size of deep neural network (DNN) models and datasets increases, distributed training becomes popular to reduce the training time. However, a severe communication bottleneck in distributed training limits its scalability. Many methods aim to solve this communication bottleneck by reducing communication traffic, such as gradient sparsification and quantification. However, these methods either are at the expense of losing model accuracy or introducing lots of computing overhead. We observe that the data distribution between layers of neural network models is similar. Thus, we propose a model parameter prediction method (MP2) to accelerate distributed DNN training under parameter server (PS) framework, where workers push only a subset of model parameters to PS, and residual model parameters are locally predicted by an already-trained deep neural network model on PS. Some key challenges are addressed. First, we build a hierarchical parameters dataset by randomly sampling a subset of model from normal distributed trainings. Second, we build a neural network model with the structure of “convolution + channel attention + Max pooling” for predicting model parameters by using a prediction result-based evaluation method. For VGGNet, ResNet, and AlexNet models on CIFAR10 and CIFAR100 datasets, compared with Baseline, Top-k, deep gradient compression (DGC), and weight nowcaster network (WNN), MP2 can reduce traffic by up to 88.98%; and accelerates the training by up to 47.32% while not losing the model accuracy. MP2 has shown good generalization.

Generating implicit object fragment datasets for machine learning

One of the primary challenges inherent in utilizing deep learning models is the scarcity and accessibility hurdles associated with acquiring datasets of sufficient size to facilitate effective training of these networks. This is particularly significant in object detection, shape completion, and fracture assembly. Instead of scanning a large number of real-world fragments, it is possible to generate massive datasets with synthetic pieces. However, realistic fragmentation is computationally intensive in the preparation (e.g., pre-factured models) and generation. Otherwise, simpler algorithms such as Voronoi diagrams provide faster processing speeds at the expense of compromising realism. In this context, it is required to balance computational efficiency and realism. This paper introduces a GPU-based framework for the massive generation of voxelized fragments derived from high-resolution 3D models, specifically prepared for their utilization as training sets for machine learning models. This rapid pipeline enables controlling how many pieces are produced, their dispersion and the appearance of subtle effects such as erosion. We have tested our pipeline with an archaeological dataset, producing more than 1M fragmented pieces from 1,052 Iberian vessels (). Although this work primarily intends to provide pieces as implicit data represented by voxels, triangle meshes and point clouds can also be inferred from the initial implicit representation. To underscore the unparalleled benefits of CPU and GPU acceleration in generating vast datasets, we compared against a realistic fragment generator that highlights the potential of our approach, both in terms of applicability and processing time. We also demonstrate the synergies between our pipeline and realistic simulators, which frequently cannot select the number and size of resulting pieces. To this end, a deep learning model was trained over realistic fragments and our dataset, showcasing similar results.

Ecogeographic body size clines in turtles: analysis and synthesis of pattern and process

Abstract The mechanisms underpinning the evolution of ectotherm body size across temperature and latitude remain an enduring mystery. Oxygen limitation in warmer aquatic environments is often invoked to explain steeper temperature–size relationships of aquatic species across latitude, but the role of aquatic environments on body size evolution of aquatic species that respire primarily in air is unclear. We tested the generality of intraspecific temperature–size responses in Testudines, assessing whether turtles generally follow Bergmann’s Rule, and whether aquatic versus terrestrial turtles exhibit different temperature–size responses. Using a global dataset of turtle body sizes, we found that turtles generally show decreasing body size with increasing temperature. However, we also found that aquatic, semi-aquatic, and terrestrial turtles respond similarly to temperature. Our study affirms the possibility that temperature–size relationships are driven partly by oxygen limitation in other ectotherms, but casts doubt on hypotheses related to oxygen limitation for groups that respire primarily in air. We formalize and explore a hypothesis that temperature–size responses in turtles are a result of heat balance, arguing small size in warm environments reduces overheating risk, and large size in cool environments minimizes heat loss. Ultimately, we provide new insight and direction in the study of ecogeographic clines.

Ageing drop by drop: Disturbance of the membrane-less organelle biogenesis as an aging hallmark

Despite extensive research, the features associated with the aging phenotype are not all-inclusive and need to be updated on a regular basis to incorporate new findings. We propose to include the dysfunction of membrane-less organelle (MLO) as a new aging hallmark. Special scaffold proteins with a high degree of intrinsic disorder drive the formation of MLOs via the liquid-liquid phase separation (LLPS) process. Aberrant behavior of MLOs was shown to be associated with the pathogenesis of many neurodegenerative diseases. In this work, we challenge the aging through bidirectional bioinformatics analysis of human proteins found in Granulome consisting of 7264 protein and Ageome containing 1624 aging-related proteins. The analysis indicates the interconnectivity of MLOs and aging. Approximately 67 % of the Ageome are presented in Granulome thereby constituting the Intersectome that include 1084 proteins showing an enrichment significantly higher than for the random datasets of the same size. Furthermore, for proteins in 10 representative MLOs, we analyzed in detail molecular functions, association with the already known aging hallmarks, and the roles in MLO formation (scaffold, client, or regulator). Cumulatively, our results strengthen the hypothesis that the dysfunction of MLOs can serve as a potent new aging hallmark.

GPU-Enabled Volume Renderer for Use with MATLAB

Traditional tools, such as 3D Slicer, Fiji, and MATLAB®, often encounter limitations in rendering performance and data management as the dataset sizes increase. This work presents a GPU-enabled volume renderer with a MATLAB® interface that addresses these issues. The proposed renderer uses flexible memory management and leverages the GPU texture-mapping features of NVIDIA devices. It transfers data between the CPU and the GPU only in the case of a data change between renderings, and uses texture memory to make use of specific hardware benefits of the GPU and improve the quality. A case study using the ViBE-Z zebrafish larval dataset demonstrated the renderer’s ability to produce visualizations while managing extensive data effectively within the MATLAB® environment. The renderer is available as open-source software.

영상처리 기법 기반 겹친 넙치 객체 탐지를 위한 자동 라벨링 방법

Detecting objects in underwater environments is challenging due to factors such as object overlap, noise, and lighting variations. In particular, labeling underwater organisms like Paralichthys olivaceus requires significant time and effort when done manually. To address these challenges, this paper proposes a method to automate the labeling of Paralichthys olivaceus using image processing techniques and to increase the size of the dataset for training. The proposed method applies morphological operations and contour detection to remove noise and accurately extract object boundaries. Additionally, the YOLOv8 model is used to analyze performance variations based on dataset size and training methods. The results demonstrate that expanding the dataset through automated image preprocessing improves object detection performance, even in cases of object overlap, and proves to be more time-efficient compared to manual labeling.

Integrating clustering and sequential analysis for improving the spectral density estimation and dependency structure of time series

Time series models (TSMs) are used to forecast events based on verified historical data that are widely applied to describe natural and social phenomena. The spectral density function (SDF) of the TSM has the benefit to determine trends, periodic patterns and behavior of observable events and is used in degree of memory (DOM) which is used to identify independent and dependent structures. Therefore, paying attention to the estimator of SDF (SDE) is the cornerstone and desired goal in the statistical properties of TSMs. This paper gives a novel perspective for improving the SDE then dependency structure of TSMs by integrating clustering and sequential analysis. The accuracy of the lag window estimator (LWE) and periodogram estimator (PE), which are commonly used as SDEs in practice, is examined both with and without integrating the new proposed technique in order to figure out the efficacy of the new technique. Several circumstances, including independent and dependent data structures, TSMs with long and short DOMs, different numbers of clusters, and different dataset sizes, are used to carry out the investigation. The main findings indicate that: (i) The SDEs (PE and LWE) perform better when combined with the new proposed technique than when used alone. (ii) Using the new technique in conjunction with the SDEs results in absolutely perfect estimators for independent processes. (iii) High efficiency estimators for short memory TSMs arise when the new proposed technique is combined with the SDEs in the dependent processes; however, high efficiency in long memory TSMs (LMTSMs) is not guaranteed. (iv) Using the new technique, the correlation estimator behavior between SDEs and frequencies improves significantly with increasing DOM values of LMTSMs. (v) By applying the new technique for LMTSMs, the DOM estimator provides results indicating significant dependence. This study demonstrates that there are other efficient ways to improve the estimator than estimate techniques.

Bilateral symmetry-based augmentation method for improved tooth segmentation in panoramic X-rays

Panoramic X-rays are crucial in dental radiology, providing detailed images that are essential for diagnosing and planning treatment for various oral conditions. The advent of automated methods that learn from annotated data promises to significantly aid clinical experts in making accurate diagnoses. However, these methods often require large amounts of annotated data, making the generation of high-quality annotations for panoramic X-rays both challenging and time-consuming. This paper introduces a novel bilateral symmetry-based augmentation method specifically designed to enhance tooth segmentation in panoramic X-rays. By exploiting the inherent bilateral symmetry of these images, our proposed method systematically generates augmented data, leading to substantial improvements in the performance of tooth segmentation models. By increasing the training data size fourfold, our approach proportionately reduces the effort required to manually annotate extensive datasets. These findings highlight the potential of leveraging the symmetrical properties of medical images to enhance model performance and accuracy in dental radiology. The effectiveness of the proposed method is evaluated on three widely adopted deep learning models: U-Net, SE U-Net, and TransUNet. Significant improvements in segmentation accuracy are observed with the proposed augmentation method across all models. For example, the average Dice Similarity Coefficient (DSC) increases by over 8%, reaching 76.7% for TransUNet. Further, comparisons with existing augmentation methods, including rigid transform-based and elastic grid-based techniques, show that the proposed method consistently outperforms them with additional improvements up to 5% in terms of average DSC, with the exact improvement varying depending on the model and training dataset size. We have made the data augmentation codes and tools developed based on our method available at https://github.com/wathoresanket/bilateralsymmetrybasedaugmentation.

FL-YOLOv8: Lightweight Object Detector Based on Feature Fusion

In recent years, anchor-free object detectors have become predominant in deep learning, the YOLOv8 model as a real-time object detector based on anchor-free frames is universal and influential, it efficiently detects objects across multiple scales. However, the generalization performance of the model is lacking, and the feature fusion within the neck module overly relies on its structural design and dataset size, and it is particularly difficult to localize and detect small objects. To address these issues, we propose the FL-YOLOv8 object detector, which is improved based on YOLOv8s. Firstly, we introduce the FSDI module in the neck, enhancing semantic information across all layers and incorporating rich detailed features through straightforward layer-hopping connections. This module integrates both high-level and low-level information to enhance the accuracy and efficiency of image detection. Meanwhile, the structure of the model was optimized and designed, and the LSCD module is constructed in the detection head; adopting a lightweight shared convolutional detection head reduces the number of parameters and computation of the model by 19% and 10%, respectively. Our model achieves a comprehensive performance of 45.5% on the COCO generalized dataset, surpassing the benchmark by 0.8 percentage points. To further validate the effectiveness of the method, experiments were also performed on specific domain urine sediment data (FCUS22), and the results on category detection also better justify the FL-YOLOv8 object detection algorithm.

Radiation Dose Prediction for Cervical Cancer Patients Using IMRT Technique with a Machine Learning Model Based on Support Vector Regression (SVR)

Cervical cancer poses significant global health challenges, necessitating the need for innovative treatment approaches. This study addresses the gap in current radiotherapy methods by integrating Support Vector Regression (SVR) to predict radiation doses for cervical cancer treatment, thereby enhancing the precision of Intensity Modulated Radiation Therapy (IMRT). Using datasets from 102 and 173 cervical cancer cases, we developed and validated an SVR model to predict dose distributions based on radiomic and dosiomic features. The model demonstrated strong performance, achieving a Mean Absolute Error (MAE) of 0.069 for the testing data, with specific performance metrics as follows: bladder mean dose MAE of 0.0693, bowel mean dose MAE of 0.0926, and rectum mean dose MAE of 0.0779. These findings highlight the potential of machine learning to refine radiotherapy planning, reduce the workload on medical physicists, and improve patient outcomes. Future research should focus on expanding dataset sizes and enhancing model precision, particularly for anatomically challenging regions.

Generative modeling of the Circle of Willis using 3D-StyleGAN

The circle of Willis (CoW) is a network of cerebral arteries with significant inter-individual anatomical variations. Deep learning has been used to characterize and quantify the status of the CoW in various applications for the diagnosis and treatment of cerebrovascular disease. In medical imaging, the performance of deep learning models is limited by the diversity and size of training datasets. To address medical data scarcity, generative AI models have been applied to generate synthetic vessel neuroimaging data. However, the proposed methods produce synthetic data with limited anatomical fidelity or downstream utility in tasks concerning vessel characteristics.We adapted the StyleGANv2 architecture to 3D to synthesize Time-of-Flight Magnetic Resonance Angiography (TOF MRA) volumes of the CoW. For generative modeling, we used 1782 individual TOF MRA scans from 6 open source datasets. To train the adapted 3D StyleGAN model with limited data we employed differentiable data augmentations, used mixed precision and a cropped region of interest of size 32 × 128 × 128 to tackle computational constraints. The performance was evaluated quantitatively using the Fréchet Inception Distance (FID), MedicalNet distance (MD) and Area Under the Curve of the Precision and Recall Curve for Distributions (AUC-PRD). Qualitative analysis was performed via a visual Turing test. We demonstrated the utility of generated data in a downstream task of multiclass semantic segmentation of CoW arteries. Vessel segmentation performance was assessed quantitatively using the Dice coefficient and the Hausdorff distance.The best-performing 3D StyleGANv2 architecture generated high-quality and diverse synthetic TOF MRA volumes (FID: 12.17, MD: 0.00078, AUC-PRD: 0.9610). Multiclass vessel segmentation models trained on synthetic data alone achieved comparable performance to models trained using real data in most arteries. The addition of synthetic data to a baseline training set improved segmentation performance in underrepresented artery segments, similar to the addition of real data.In conclusion, generative modeling of the Circle of Willis via synthesis of 3D TOF MRA data paves the way for generalizable deep learning applications in cerebrovascular disease. In the future, the extensions of the provided methodology to other medical imaging problems or modalities with the inclusion of pathological datasets has the potential to advance the development of more robust AI models for clinical applications.

Instability results for cosine-dissimilarity-based nearest neighbor search on high dimensional Gaussian data

Because many dissimilarity functions behave differently in low versus high-dimensional spaces, the behavior of high-dimensional nearest neighbor search has been studied extensively. One line of research involves the characterization of nearest neighbor queries as unstable if their query points have nearly identical dissimilarity with most points in the dataset. This research has shown that, for various data distributions and dissimilarity functions, the probability of query instability approaches one. Previous work in Information Processing Letters by C. Giannella in 2021 explicated this phenomenon for centered Gaussian data and Euclidean distance. This paper addresses the problem of characterizing query instability behavior over centered Gaussian data and a fundamentally different dissimilarity function, cosine dissimilarity. Conditions are provided on the covariance matrices and dataset size function guaranteeing that the probability of query instability goes to one. Furthermore, conditions are provided under which the instability probability is bounded away from one.