Search for: All records

Proton beam radiotherapy is a method of cancer treatment that uses proton beams to irradiate cancerous tissue, while minimizing doses to healthy tissue. In order to guarantee that the prescribed radiation dose is delivered to the tumor and ensure that healthy tissue is spared, many researchers have suggested verifying the treatment delivery through the use of real-time imaging using methods which can image prompt gamma rays that are emitted along the beam’s path through the patient such as Compton cameras (CC). However, because of limitations of the CC, their images are noisy and unusable for verifying proton treatment delivery. We provide a detailed description of a deep residual fully connected neural network that is capable of classifying and improving measured CC data with an increase in the fraction of usable data by up to 72% and allows for improved image reconstruction across the full range of clinical treatment delivery conditions. 

Proton beam therapy is a unique form of radiotherapy that utilizes protons to treat cancer by irradiating cancerous tumors, while avoiding unnecessary radiation exposure to surrounding healthy tissues. Real-time imaging of the proton beam can make this form of therapy more precise and safer for the patient during delivery. The use of Compton cameras is one proposed method for the real-time imaging of prompt gamma rays that are emitted by the proton beams as they travel through a patient’s body. Unfortunately, some of the Compton camera data is flawed and the reconstruction algorithm yields noisy and insufficiently detailed images to evaluate the proton delivery for the patient. Previous work used a deep residual fully connected neural network. The use of recurrent neural networks (RNNs) has been proposed, since they use recurrence relationships to make potentially better predictions. In this work, RNN architectures using two different recurrent layers are tested, the LSTM and the GRU. Although the deep residual fully connected neural network achieves over 75% testing accuracy and our models achieve only over 73% testing accuracy, the simplicity of our RNN models containing only 6 hidden layers as opposed to 512 is a significant advantage. Importantly in a clinical setting, the time to load the model from disk is significantly faster, potentially enabling the use of Compton camera image reconstruction in real-time during patient treatment. 

Atmospheric gravity waves are produced when gravity attempts to restore disturbances through stable layers in the atmosphere. They have a visible effect on many atmospheric phenomena such as global circulation and air turbulence. Despite their importance, however, little research has been conducted on how to detect gravity waves using machine learning algorithms. We faced two major challenges in our research: our raw data had a lot of noise and the labeled dataset was extremely small. In this study, we explored various methods of preprocessing and transfer learning in order to address those challenges. We pre-trained an autoencoder on unlabeled data before training it to classify labeled data. We also created a custom CNN by combining certain pre-trained layers from the InceptionV3 Model trained on ImageNet with custom layers and a custom learning rate scheduler. Experiments show that our best model outperformed the best performing baseline model by 6.36% in terms of test accuracy. 

There is a critical nationwide shortage of IT professionals as well as of scientists and engineers with high-performance computing (HPC) and big data related advanced computing skills. Simultaneously, the technology is growing in complexity and sophistication, which has led to the use of multi-disciplinary teams with members from a broad range of home domains everywhere in industry, government, and academia. Moreover, a lot of the vital team collaborations take will place virtually using a variety of software platforms now and in the future. We report here on experiences with preparing undergraduate and graduate students for these career opportunities in several contexts, from regular semester classes, an undergraduate summer research program, to an advanced graduate student CyberTraining program. All these programs are conducted fully online and leveraged concepts of flipped classrooms, recorded lectures, team-based and active learning, regular oral presentations, and more to ensure student engagement and lasting learning. 

During 2018, 2019, and 2020, the UMBC CyberTraining initiative “Big Data + HPC + Atmospheric Sciences” created an online team-based training program for advanced graduate students and junior researchers that trained a total of 58 participants. The year 2020 included 6 undergraduate students. Based on this experience, the authors created the summer undergraduate research program Online Interdisciplinary Big Data Analytics in Science and Engineering that will conduct 8-week online team-based undergraduate research programs (bigdatareu.umbc.edu) in the summers 2021, 2022, and 2023. Given the context of many institutions potentially expanding their online instruction, we share our experiences how the successful lessons from CyberTraining transfer to a high-intensity full-time online summer undergraduate research program. 

Accurately forecasting Arctic sea ice from sub- seasonal to seasonal scales has been a major scientific effort with fundamental challenges at play. In addition to physics-based earth system models, researchers have been applying multiple statistical and machine learning models for sea ice forecast- ing. Looking at the potential of data-driven sea ice forecasting, we propose an attention-based Long Short Term Memory (LSTM) ensemble method to predict monthly sea ice extent up to 1 month ahead. Using daily and monthly satellite retrieved sea ice data from NSIDC and atmospheric and oceanic variables from ERA5 reanalysis product for 39 years, we show that our multi-temporal ensemble method outperforms several baseline and recently proposed deep learning models. This will substantially improve our ability in predicting future Arctic sea ice changes, which is fundamental for forecasting transporting routes, resource development, coastal erosion, threats to Arctic coastal communities and wildlife. 

Important natural resources in the Arctic rely heavily on sea ice, making it important to forecast Arctic sea ice changes. Arctic sea ice forecasting often involves two connected tasks: sea ice concentration at each pixel and overall sea ice extent. Instead of having two separate models for two forecasting tasks, in this report, we study how to use multi-task learning techniques and leverage the connections between ice concentration and ice extent to improve accuracy for both prediction tasks. Because of the spatiotemporal nature of the data, we designed two novel multi-task learning models based on CNNs and ConvLSTMs, respectively. We also developed a custom loss function which trains the models to ignore land pixels when making predictions. Our experiments show our models can have better accuracies than separate models that predict sea ice extent and concentration separately, and that our accuracies are better than or comparable with results in the state-of-the-art studies.