Browsing by Subject "convolutional neural networks"
Now showing 1 - 5 of 5
Results Per Page
Sort Options
Item An Approach to Tuning Hyperparameters in Parallel: An Approach to Tuning Hyperparameters in Parallel(2019-01-01) Barajas, Carlos Alexander; Gobbert, Matthias K; Mathematics and Statistics; Mathematics and StatisticsPredicting violent storms and dangerous weather conditions, for instance predicting tornados, is an important application for public safety. Using numerical weather simulations to classify a weather pattern as tornadic or not tornadic can take a long time due to the immense complexity associated with current models. Machine learning has the potential to classify tornadic weather patterns much more quickly, since a trained model can classify a storm in seconds when provided data. This allows for alerts to be issued more rapidly such that there is more time for the public to respond to the alert. Preprocessing methods must be applied to the natural data prior to training to prevent inaccuracies. Neural networks must have an equal balance of all classifiable cases for accurate prediction of each case. Different data augmentation approaches have been proposed to fix data imbalances. With our specialized HPC enabled framework, we examine the wall time difference of live data augmentation method versus the use of preaugmented data, when used with a convolutional neural network under various hyperparameter configurations. We also compare CPU and GPU based training over varying sizes of augmented data sets. Then we examine the wall time impact associated with varying the number of GPUs used for training a convolutional neural network. Finally, we create a new data augmentation system by implementing a generative adversarial network and qualitatively compare its output to natural data.Item Dust Detection in Satellite Data using Convolutional Neural Networks(HPCF UMBC, 2019) Cai, Changjie; Lee, Jangho; Shi, Yingxi Rona; Zerfas, Camille; Guo, Pei; Zhang, ZhiboAtmospheric dust is known to cause health ailments and impacts earth’s climate and weather patterns. Due to the many issues atmospheric dust contributes to, it is important to study dust patterns and how it enters the atmosphere. In the past, many scientists have used satellite data and physical-based algorithms to detect and track dust, but these algorithms have many shortcomings. Herein, we consider Convolutional Neural Networks to classify dust in satellite images to try to improve the accuracy of dust detection. We describe the satellite data used, discuss the model structures, and provide results for the models built. These models show promising preliminary results.Item Learning Aligned Cross-Modal Representations from Weakly Aligned Data(IEEE, 2016-06-30) Castrejón, Lluís; Aytar, Yusuf; Vondrick, Carl; Pirsiavash, Hamed; Torralba, AntonioPeople can recognize scenes across many different modalities beyond natural images. In this paper, we investigate how to learn cross-modal scene representations that transfer across modalities. To study this problem, we introduce a new cross-modal scene dataset. While convolutional neural networks can categorize cross-modal scenes well, they also learn an intermediate representation not aligned across modalities, which is undesirable for crossmodal transfer applications. We present methods to regularize cross-modal convolutional neural networks so that they have a shared representation that is agnostic of the modality. Our experiments suggest that our scene representation can help transfer representations across modalities for retrieval. Moreover, our visualizations suggest that units emerge in the shared representation that tend to activate on consistent concepts independently of the modality.Item Predicting Mortality of Diabetic ICU Patients(2019-04) Wittler, Ian; Liu, Xinlian; Dong, Aijuan; Computer Science; Hood College Departmental HonorsDiabetes mellitus (DM) is a major public health concern that requires continuing medical care. It is also a leading cause of other serious health complications associated with longer hospital stays and increased mortality rates. Fluctuation of blood glucose levels are easy to monitor. Physicians manage patients' blood glucose to prevent or slow the progress of diabetes. In this paper, the MIMIC-III data set is used to develop and train multiple models that aim to predict the mortality of DM patients. Our deep learning model of convolutional neural network produced a 0.885 AUC score, above all baseline models we constructed, which include decision trees, random forests, and fully connected neural networks. The inputs for each model were comprised of admission type, age, Elixhauser comorbidity score, blood glucose measurements, and blood glucose range. The results obtained from these models are valuable for physicians, patients, and insurance companies. By analyzing the features that drive these models, care management for diabetic patients in an ICU setting can be improved resulting in lowered motality rate.Item Refining Ice Layer Tracking through Wavelet combined Neural Networks (Papers Track)(2021) Varshney, Debvrat; Yari, Masoud; Chowdhury, Tashnim; Rahnemoonfar, MaryamRise in global temperatures is resulting in polar ice caps to melt away, which can lead to drastic sea level rise and coastal floods. Accurate calculation of the ice cap reduction is necessary in order to project its climatic impact. Ice sheets are monitored through Snow Radar sensors which give noisy profiles of subsurface ice layers. The sensors take snapshots of the entire ice sheet regularly, and thus result in large datasets. In this work, we use convolutional neural networks (CNNs) for their property of feature extraction and generalizability on large datasets. We also use wavelet transforms and embed them as a layer in the architecture to help in denoising the radar images and refine ice layer detection. Our results show that incorporating wavelets in CNNs helps in detecting the position of deep subsurface ice layers, which can be used to analyse their change overtime.