Version Of Dataset Research Articles

Robotic locomotor training for spasticity, pain, and quality of life in individuals with chronic SCI: A pilot randomized controlled trial.

The prevention and treatment of secondary complications is a key priority for people with spinal cord injury and a fundamental goal of rehabilitation. Activity-based Training (ABT) and Robotic Locomotor Training (RLT) demonstrate promising results for reducing secondary complications associated with SCI. However, there is a need for increased evidence through randomized controlled trials. Therefore, we aimed to investigate the effect of RLT and ABT interventions on pain, spasticity, and quality of life in individuals with spinal cord injuries. Participants with chronic motor incomplete tetraplegia (n = 16) were recruited. Each intervention involved 60-minute sessions, 3× per week, over 24-weeks. RLT involved walking in an Ekso GT exoskeleton. ABT involved a combination of resistance, cardiovascular and weight-bearing exercise. Outcomes of interest included the Modified Ashworth Scale, the International SCI Pain Basic Data Set Version 2, and the International SCI Quality of Life Basic Data Set. Neither intervention altered symptoms of spasticity. Pain intensity increased from pre-post intervention for both groups, with a mean increase of 1.55 [-0.82, 3.92] (p = 0.03) and 1.56 [-0.43, 3.55] (p = 0.02) points for the RLT and ABT group, respectively. The ABT group had an increase in pain interference scores of 100%, 50%, and 109% for the daily activity, mood, and sleep domain, respectively. The RLT group had an increase in pain interference scores of 86% and 69% for the daily activity and mood domain respectively, but no change in the sleep domain. The RLT group had increased perceptions of quality of life with changes of 2.37 [0.32, 4.41], 2.00 [0.43, 3.56] and 0.25 [-1.63, 2.13] points, p = 0.03, for the general, physical, and psychological domains, respectively. The ABT group had increased perceptions of general, physical and psychological quality of life with changes of 0.75 [-1.38, 2.88], 0.62 [-1.83, 3.07] and 0.63 [-1.87, 3.13] points, respectively. Despite increased pain ratings and no change in symptoms of spasticity, there was an increase in perceived quality of life for both groups over 24-weeks. This dichotomy warrants additional investigation in future large-scale randomized controlled trials.

VISAL—A novel learning strategy to address class imbalance

In the imbalance data scenarios, Deep Neural Networks (DNNs) fail to generalize well on minority classes. In this letter, we propose a simple and effective learning function i.e, Visually Interpretable Space Adjustment Learning (VISAL) to handle the imbalanced data classification task. VISAL’s objective is to create more room for the generalization of minority class samples by bringing in both the angular and euclidean margins into the cross-entropy learning strategy. When evaluated on the imbalanced versions of CIFAR, Tiny ImageNet, COVIDx and IMDB reviews datasets, our proposed method outperforms the state of the art works by a significant margin.

Predicting Progression from Normal to MCI and from MCI to AD Using Clinical Variables in the National Alzheimer’s Coordinating Center Uniform Data Set Version 3: Application of Machine Learning Models and a Probability Calculator

Clinical trials are increasingly focused on pre-manifest and early Alzheimer’s disease (AD). Accurately predicting clinical progressions from normal to MCI or from MCI to dementia/AD versus non-progression is challenging. Accurate identification of symptomatic progressors is important to avoid unnecessary treatment and improve trial efficiency. Due to large inter-individual variability, biomarker positivity and comorbidity information are often insufficient to identify those destined to have symptomatic progressions. Using only clinical variables, we aimed to predict clinical progressions, estimating probabilities of progressions with a small set of variables selected by machine learning approaches. This work updates our previous work that was applied to the National Alzheimer’s Coordinating Center (NACC) Uniform Data Set Version 2 (V2), by using the most recent version (V3) with additional analyses. We generated a user-friendly conversion probability calculator which can be used for effectively pre-screening trial participants.

Inferring 3D change detection from bitemporal optical images

In recent years, change detection (CD) using deep learning (DL) algorithms has been a very active research topic in the field of remote sensing (RS). Nevertheless, the CD algorithms developed so far are mainly focused on generating two-dimensional (2D) change maps where the planimetric extent of the areas affected by changes is identified without providing any information on the corresponding elevation variations. The aim of this work is, hence, to establish the basis for the development of DL algorithms able to automatically generate an elevation (3D) CD map along with a standard 2D CD map, using only bitemporal optical images as input, and thus without the need to rely directly on elevation data during the inference phase. Specifically, our work proposes a novel network, capable of solving the 2D and 3D CD tasks simultaneously, and a modified version of the 3DCD dataset, a freely available dataset designed precisely for this twofold task. The proposed architecture consists of a Transformer network based on a semantic tokenizer: the MultiTask Bitemporal Images Transformer (MTBIT). Encouraging results, obtained on the modified version of the 3DCD dataset by comparing the proposed architecture with other networks specifically designed to solve the 2D CD task, are shown. In particular, MTBIT achieves a metric accuracy (represented by the changed root mean squared error) equal to 6.46 m – the best performance among the compared architectures – with a limited number of parameters (13,1 M). The code and the 3DCD dataset are available at https://sites.google.com/uniroma1.it/3dchangedetection/home-page.

Hippocampus Segmentation-Based Alzheimer's Disease Diagnosis and Classification of MRI Images.

Alzheimer's disease represents a neurological condition characterized by steady cognitive decline and eventual memory loss due to the death of brain cells. It is one of the most prominent dementia types observed in patients and which hence underlines the imminent need for potential methods to diagnose the disease early on. This work considers a novel approach by utilizing a reduced version of one of the datasets used in this work to achieve a considerably accurate prediction while also enabling quicker training. It leverages image segmentation to isolate the hippocampus region from brain MRI images and then strikes a comparison between models trained on the segmented portions and models trained on complete images. This research uses two datasets-4 classes of images from Kaggle and a popular OASIS 2 MRI and demographic dataset. A deep learning-based approach was adopted to train the Kaggle dataset to perform severity classification, and the hippocampus region segmented from a reduced version of the OASIS dataset was trained on supervised and ensemble learning algorithms to detect Alzheimer's disease. The metric used for the assessment of model performance is classification accuracy. A comparative analysis between the proposed approach and existing work was also performed, and it was observed that the proposed approach is effective in the early diagnosis of Alzheimer's disease.

Shot Optimization in Quantum Machine Learning Architectures to Accelerate Training

Quantum Machine Learning (QML) has recently emerged as a rapidly growing domain as an intersection of Quantum Computing (QC) and Machine Learning (ML) fields. Hybrid quantum-classical models have demonstrated exponential speedups in various machine learning tasks compared to their classical counterparts. On one hand, training of QML models on real hardware remains a challenge due to long wait queue and the access cost. On the other hand, simulation-based training is not scalable to large QML models due to exponentially growing simulation time. Since the measurement operation converts quantum information to classical binary data, the quantum circuit is executed multiple times (called shots) to obtain the basis state probabilities or qubit expectation values. Higher number of shots worsen the training time of QML models on real hardware and the access cost. Higher number of shots also increase the simulation-based training time. In this paper, we propose shot optimization method for QML models at the expense of minimal impact on model performance. We use classification task as a test case for MNIST and FMNIST datasets using a hybrid quantum-classical QML model. First, we sweep the number of shots for short and full versions of the dataset. We observe that training the full version provides 5-6% higher testing accuracy than short version of dataset with up to 10X higher number of shots for training. Therefore, one can reduce the dataset size to accelerate the training time. Next, we propose adaptive shot allocation on short version dataset to optimize the number of shots over training epochs and evaluate the impact on classification accuracy. We use a (a) linear function where the number of shots reduce linearly with epochs, and (b) step function where the number of shots reduce in step with epochs. We note around 0.01 increase in loss and ~4% (1%) reduction in testing accuracy for reduction in shots by up to 100X (10X) for linear (step) shot function compared to conventional constant shot function for MNIST dataset, and 0.05 increase in loss and ~5-7% (5-7%) reduction in testing accuracy with similar reduction in shots using linear (step) shot function on FMNIST dataset. For comparison, we also use the proposed shot optimization methods to perform ground state energy estimation of different molecules and observe that step function gives the best and most stable ground state energy prediction at 1000X less number of shots.

Human Pose Estimation Using Thermal Images

This study addresses the human pose estimation problem on thermal images using Convolutional Neural Networks and Vision Transformer architectures. To do this, eight human pose estimation methods designed for visible images were extended to be applied in the thermal domain. Due to the lack of large, representative datasets containing labeled thermal images, this extension requires transfer learning between the visible and the thermal domain, and a database for fine-tuning the networks in the thermal domain. Thus, it is proposed to train the networks using a grayscale version of the COCO dataset, and then fine-tune them in the thermal domain. Fine-tuning is carried out using the new UCH-Thermal-Pose database presented in this work. This database includes 600 thermal images for training, 200 for validation, and 104 for testing, all of them fully labeled. Moreover, in the paper, a comparative study of the eight extended deep-based methods for human pose detection is carried out. The UCH-Thermal-Pose database and the source code of all the methods compared are made available for future studies.

Malware Detection Using Deep Learning and Correlation-Based Feature Selection

Malware is one of the most frequent cyberattacks, with its prevalence growing daily across the network. Malware traffic is always asymmetrical compared to benign traffic, which is always symmetrical. Fortunately, there are many artificial intelligence techniques that can be used to detect malware and distinguish it from normal activities. However, the problem of dealing with large and high-dimensional data has not been addressed enough. In this paper, a high-performance malware detection system using deep learning and feature selection methodologies is introduced. Two different malware datasets are used to detect malware and differentiate it from benign activities. The datasets are preprocessed, and then correlation-based feature selection is applied to produce different feature-selected datasets. The dense and LSTM-based deep learning models are then trained using these different versions of feature-selected datasets. The trained models are then evaluated using many performance metrics (accuracy, precision, recall, and F1-score). The results indicate that some feature-selected scenarios preserve almost the same original dataset performance. The different nature of the used datasets shows different levels of performance changes. For the first dataset, the feature reduction ratios range from 18.18% to 42.42%, with performance degradation of 0.07% to 5.84%, respectively. The second dataset reduction rate is between 81.77% and 93.5%, with performance degradation of 3.79% and 9.44%, respectively.

Examining the "Potentially Preventable Emergency Department Transfer" Indicator Among Nursing Home Residents.

To determine if nursing home (NH) resident characteristics associated with potentially preventable emergency department transfers (PPEDs) are similarly associated with non-potentially preventable emergency department transfers (non-PPEDs). We conducted a population-level retrospective cohort study using linked administrative data reported using the Resident Assessment Instrument-Minimum Data Set Version 2.0 and the National Ambulatory Care Reporting System for emergency department transfers. We assessed all NH residents transferred to the emergency department within 92days after admission. The cohort included 56,433 NH resident admissions assessment of which 3498 NH residents experienced PPEDs, and 9331 residents experienced non-PPEDs. We assessed Ontario NH residents admission assessments collected between January 1, 2017, and December 31, 2018. We used cumulative incidence functions and Cox regression to compare resident characteristics between residents experiencing PPEDs and non-PPEDs. PPEDs were defined based on the International Classification of Diseases, 10th Revision. Approximately 23% of residents experienced an emergency department transfer within 92days of NH admission. The cumulative incidence of PPEDs was 6.3% and non-PPEDs was 16.8%. After adjusting for clinically relevant features, 14 of 18 resident admission characteristics were associated with both types of transfers. Resident admission characteristics associated with a greater risk of PPEDs solely were pneumonia [hazard ratio (HR) 1.48; CI 1.25-1.70] and oxygen therapy (HR 1.88; CI 1.69-2.10). Resident admission characteristics associated with a greater risk of non-PPEDs solely are experiencing a change in mood (HR 1.09; CI 1.01-1.18) and delirium (HR 1.08; CI 1.04-1.13). PPEDs were associated with a similar cluster of NH resident characteristics as those transferred for non-ambulatory reasons, suggesting that the clinical distinction between PPEDs vs non-PPEDs within the NH might be unclear. These findings highlight that the PPED indicator could be revised to improve specificity.

Prevalence and Patterns of Cognitive Impairment in a Sample of Community Dwelling Older People in Nigeria.

Prevalence and patterns of cognitive impairment were studied in older people from Nigeria. Four hundred and forty one participants (263 females; age: 60-87) were recruited from community dwelling adults in Anambra state Nigeria. Five domains of cognition were tested using the Uniform Data Set Version 3 (UDS-3). Prevalence: 49.7% were classified as normal cognition, 34% as borderline, 12.9% as MCI (2.72% with amnesic MCI) and 3.4% as dementia. We showed in descending order in that 13% of the participants were impaired on visual-spatial index; 6.8% on memory index; 5.2% on attention/concentration index; 2.7% were impaired on executive function index and 34.80% (based on mean) of the participants were impaired on processing speed index. There were significant interaction effects for gender and education on visual spatial and attention domains respectively. Significant effects of education were seen on executive function and processing speed while interaction effect was found on executive function alone. 8% scored 1.5 SD below the mean on MoCA. There was a significant effect of education on MoCA with the pairwise comparison showing a significant difference between tertiary education and other two levels of education. The groups did differ significantly for hypertension on MoCA. This study showed a high prevalence of cognitive impairment among older adult population from Nigeria. A significant proportion of the sample were impaired on the visual spatial domain and at least half of the participants were impaired on one cognitive domain. Hypertensive participants performed significantly poor on MoCA compared to non-hypertensive group.

EnsembleDet: ensembling against adversarial attack on deepfake detection

Deepfakes and manipulation techniques such as adversarial attacks pose a significant threat to society today. It is essential for the detection methods used against them to be more robust than ever. Generation techniques such as neural texture, along with adversarial attacks, result in visually realistic manipulated media and make it difficult for the current state-of-the-art (SoTA) to detect the manipulation. We propose an ensemble-based approach and a genetic algorithm-based approach for fake imagery detection that has been perturbed using fast gradient sign method and basic iterative method techniques. We collectively put such methods under the umbrella term, “EnsembleDet.” We evaluate the proposed approaches using a perturbed version of the benchmark dataset of FaceForensics++, which consists of data generated from manipulation techniques such as Face2Face, neural textures, among others. We present a comparative analysis of quantitative results with SoTA, XceptionNet, and MesoNet, which depict the superiority of the proposed methods over the current SoTA. We also highlight the aspects where the proposed methods have a scope for improvement.

The curse and blessing of abundance-the evolution of drug interaction databases and their impact on drug network analysis.

Widespread bioinformatics applications such as drug repositioning or drug-drug interaction prediction rely on the recent advances in machine learning, complex network science, and comprehensive drug datasets comprising the latest research results in molecular biology, biochemistry, or pharmacology. The problem is that there is much uncertainty in these drug datasets-we know the drug-drug or drug-target interactions reported in the research papers, but we cannot know if the not reported interactions are absent or yet to be discovered. This uncertainty hampers the accuracy of such bioinformatics applications. We use complex network statistics tools and simulations of randomly inserted previously unaccounted interactions in drug-drug and drug-target interaction networks-built with data from DrugBank versions released over the plast decade-to investigate whether the abundance of new research data (included in the latest dataset versions) mitigates the uncertainty issue. Our results show that the drug-drug interaction networks built with the latest dataset versions become very dense and, therefore, almost impossible to analyze with conventional complex network methods. On the other hand, for the latest drug database versions, drug-target networks still include much uncertainty; however, the robustness of complex network analysis methods slightly improves. Our big data analysis results pinpoint future research directions to improve the quality and practicality of drug databases for bioinformatics applications: benchmarking for drug-target interaction prediction and drug-drug interaction severity standardization.

A too-good-to-be-true prior to reduce shortcut reliance

Despite their impressive performance in object recognition and other tasks under standard testing conditions, deep networks often fail to generalize to out-of-distribution (o.o.d.) samples. One cause for this shortcoming is that modern architectures tend to rely on ǣshortcutsǥ superficial features that correlate with categories without capturing deeper invariants that hold across contexts. Real-world concepts often possess a complex structure that can vary superficially across contexts, which can make the most intuitive and promising solutions in one context not generalize to others. One potential way to improve o.o.d. generalization is to assume simple solutions are unlikely to be valid across contexts and avoid them, which we refer to as the too-good-to-be-true prior. A low-capacity network (LCN) with a shallow architecture should only be able to learn surface relationships, including shortcuts. We find that LCNs can serve as shortcut detectors. Furthermore, an LCN’s predictions can be used in a two-stage approach to encourage a high-capacity network (HCN) to rely on deeper invariant features that should generalize broadly. In particular, items that the LCN can master are downweighted when training the HCN. Using a modified version of the CIFAR-10 dataset in which we introduced shortcuts, we found that the two-stage LCN-HCN approach reduced reliance on shortcuts and facilitated o.o.d. generalization.

THE ASSOCIATION BETWEEN ACCESS TO MEDICAL CARE AND RESIDENT OUTCOMES IN NURSING HOMES

Abstract The nursing home population is vulnerable and medically complex, yet little is known about models of medical service provision and associated quality outcomes. The goal of this study is to examine the association between physician (MD) and nurse practitioner (NP) accessibility and practice sensitive outcomes. This project used data from the Translating Research in Elder Care (TREC) longitudinal study and the routinely collected Resident Assessment Instrument – Minimum Data Set version 2.0 (RAI-MDS 2.0) to test the association between the availability of MDs and NPs in nursing homes (NH) and clinically-relevant resident outcomes of antipsychotic medication (APM) use without indication of psychosis, physical restraint use, hospitalization and emergency department (ED) transfers, and polypharmacy. Eight models were created using logistic regression to test the association between the access measures of daily presence of MD or NP on unit and MDs being involved in care planning and each of the four resident outcomes. The sample consisted of 10,888 residents across 320 units in 92 facilities. Staff from 277 (86%) units reported an MD or NP visited daily and 318 (99%) units reported that the MD or NP could be reached when needed. Following adjustment for multiple confounding variables, there were no associations between either measure of access and any of the resident outcomes. Although we did not find any associations between our measures of access and resident outcomes, additional research which more directly measures physician and NP activities in the NH is required.

A New Range-based Breast Cancer Prediction Model Using the Bayes' Theorem and Ensemble Learning

Breast cancer prediction is essential for preventing and treating cancer. In this research, a novel breast cancer prediction model is introduced. In addition, this research aims to provide a range-based cancer score instead of binary classification results (yes or no). The Breast Cancer Surveillance Consortium dataset (BCSC) dataset is used and modified by applying a proposed probabilistic model to achieve the range-based cancer score. The suggested model analyses a sub dataset of the whole BCSC dataset, including 67632 records and 13 risk factors. Three types of statistics are acquired (general cancer and non-cancer probabilities, previous medical knowledge, and the likelihood of each risk factor given all prediction classes). The model also uses the weighting methodology to achieve the best fusion of the BCSC's risk factors. The computation of the final prediction score is done using the post probability of the weighted combination of risk factors and the three statistics acquired from the probabilistic model. This final prediction is added to the BCSC dataset, and the new version of the BCSC dataset is used to train an ensemble model consisting of 30 learners. The experiments are applied using the sub and the whole datasets (including 317880 medical records). The results indicate that the new range-based model is accurate and robust with an accuracy of 91.33%, a false rejection rate of 1.12%, and an AUC of 0.9795. The new version of the BCSC dataset can be used for further research and analysis.

Bayesian data synthesis and the utility-risk trade-off for mixed epidemiological data

Much of the microdata used for epidemiological studies contain sensitive measurements on real individuals. As a result, such microdata cannot be published out of privacy concerns, and without public access to these data, any statistical analyses originally published on them are nearly impossible to reproduce. To promote the dissemination of key datasets for analysis without jeopardizing the privacy of individuals, we introduce a cohesive Bayesian framework for the generation of fully synthetic high-dimensional microdatasets of mixed categorical, binary, count, and continuous variables. This process centers around a joint Bayesian model that is simultaneously compatible with all of these data types, enabling the creation of mixed synthetic datasets through posterior predictive sampling. Furthermore, a focal point of epidemiological data analysis is the study of conditional relationships between various exposures and key outcome variables through regression analysis. We design a modified data synthesis strategy to target and preserve these conditional relationships, including both nonlinearities and interactions. The proposed techniques are deployed to create a synthetic version of a confidential dataset containing dozens of health, cognitive, and social measurements on nearly 20,000 North Carolina children.

Dynamic Multidimensional Knapsack Problem benchmark datasets

With increasing research on solving Dynamic Optimization Problems (DOPs), many metaheuristic algorithms and their adaptations have been proposed to solve them. However, from currently existing research results, it is hard to evaluate the algorithm performance in a repeatable way for combinatorial DOPs due to the fact that each research work has created its own version of a dynamic problem dataset using stochastic methods. Up to date, there are no combinatorial DOP benchmarks with replicable qualities. This work introduces a non-stochastic consistent Dynamic Multidimensional Knapsack Problem (Dynamic MKP) dataset generation method that is also extensible to solve the research replicability problem. Using this method, generated and published 1405 Dynamic MKP benchmark datasets using existing famous static MKP benchmark instances as the initial state. Then the datasets are quantitatively and qualitatively analyzed. Furthermore, 445 datasets have the optimal result found of each state using a linear solver. The optimal results and result scores are included with published datasets.

Predicting Conversions from Normal to MCI and from MCI to AD with Clinical Variables in the National Alzheimer’s Coordinating Center Uniform Data Set Version 3: Application of Big Data Analytics and Transition Probability Calculators

AbstractBackgroundClinical trials are increasingly focused on pre‐manifest and early Alzheimer’s disease (AD). Accurately predicting those who transit from normal to MCI or from MCI to dementia/AD (true positives) versus non‐progressors (false positives) is challenging. Biomarker positivity and comorbidity information are often not sufficient to identify those who are destined to have symptom progressions. Accurate identification of progressors is important to avoid un‐necessary treatment as well as to improve trial efficiency.MethodUsing only clinical variables and big data analytics, we aimed to predict those who convert from normal to MCI and from MCI to AD, estimating probabilities of these conversions with a small set of selected clinical variables. This work updates our previous work using National Alzheimer’s Coordinating Center (NACC) Uniform Data Set Version 2. (1) We also generated a user‐friendly conversion probability calculator to estimate individual subject conversion. We used data from the NACC Uniform Data Set Version 3 and machine learning feature selection methods to identify clinical variables that predicted MCI and dementia conversion. We built a pipeline with machine learning techniques to test model performance.ResultUsing NACC data (frozen 3/2020), 577 subjects were used to predict conversions from normal to MCI within 4 years, and 538 and 882 subjects were used to predict conversion from MCI to AD within 2 years and 3 years, respectively. An exploratory analysis was also conducted for amnestic MCI (aMCI) conversion. For each prediction model, we selected 20 clinical variables optimizing predictions using Receiver Operating Characteristics Area‐under‐Curve (AUC). Our model achieved 82.1% prediction accuracy, 75.9% specificity, 85.1% sensitivity, and 80.5% AUC for MCI to AD conversion within 3 years. Similar results were obtained for normal to MCI conversion. The variables which ranked highest in predicting transitions include clinician’s judgement of subject’s meaningful cognitive impairment (conversion to MCI) and CDR community affairs item (conversion to AD in 3 years).ConclusionAlthough NACC UDS is extensive, our algorithm is based on a large sample size of potential trial participants who visited or are referred to Alzheimer’s Disease Centers. The probability calculator can be used to effectively pre‐screen and enrich trial participants.

Estimating Chinese bilateral aid for health: an analysis of AidData’s Global Chinese Official Finance Dataset Version 2.0

BackgroundAlthough it is difficult to quantify, previous estimates suggested that China’s global health aid has increased sharply since the early 2000s. Unlike many donors, China has no official aid reporting...

Visual question answering from another perspective: CLEVR mental rotation tests

Different types of mental rotation tests have been used extensively in psychology to understand human visual reasoning and perception. Understanding what an object or visual scene would look like from another viewpoint is a challenging problem that is made even harder if it must be performed from a single image. We explore a controlled setting whereby questions are posed about the properties of a scene if that scene was observed from another viewpoint. To do this we have created a new version of the CLEVR dataset that we call CLEVR Mental Rotation Tests (CLEVR-MRT). Using CLEVR-MRT we examine standard methods, show how they fall short, then explore novel neural architectures that involve inferring volumetric representations of a scene. These volumes can be manipulated via camera-conditioned transformations to answer the question. We examine the efficacy of different model variants through rigorous ablations and demonstrate the efficacy of volumetric representations.