Precision medicine in cancer treatment increasingly relies on advanced radiotherapies, such as proton beam radiotherapy, to enhance efficacy of the treatment. When the proton beam in this treatment interacts with pati...
详细信息
ISBN:
(数字)9798350362480
ISBN:
(纸本)9798350362497
Precision medicine in cancer treatment increasingly relies on advanced radiotherapies, such as proton beam radiotherapy, to enhance efficacy of the treatment. When the proton beam in this treatment interacts with patient matter, the excited nuclei may emit prompt gamma ray interactions that can be captured by a Compton camera. The image reconstruction from this captured data faces the issue of mischaracterizing the sequences of incoming scattering events, leading to excessive background noise. To address this problem, several machine learning models such as Feedfoward Neural Networks (FNN) and Recurrent Neural Networks (RNN) were developed in PyTorch to properly characterize the scattering sequences on simulated datasets, including newly-created patient medium data, which were generated by using a pipeline comprised of the GEANT4 and Monte-Carlo Detector Effects (MCDE) softwares. These models were implemented using the novel ‘Big-data REU Integrated Development and Experimentation’ (BRIDE) platform, a modular pipeline that streamlines preprocessing, feature engineering, and model development and evaluation on parallelized GPU processors. Hyperparameter studies were done on the novel patient data as well as on water phantom datasets used during previous research. Patient data was more difficult than water phantom data to classify for both FNN and RNN models. FNN models had higher accuracy on patient medium data but lower accuracy on water phantom data when compared to RNN models. Previous results on several different datasets were reproduced on BRIDE and multiple new models achieved greater performance than in previous research.
Small object detection (SOD) in optical images and videos is a challenging problem that even state-of-the-art generic object detection methods fail to accurately localize and identify such objects. Typically, small ob...
详细信息
The conditional value-at-risk (CVaR) is a useful risk measure in fields such as machine learning, finance, insurance, energy, etc. When measuring very extreme risk, the commonly used CVaR estimation method of sample a...
详细信息
In today's rapidly evolving technological landscape, ensuring the security of systems requires continuous authentication over sessions and comprehensive access management during user interaction with a device. Wit...
In today's rapidly evolving technological landscape, ensuring the security of systems requires continuous authentication over sessions and comprehensive access management during user interaction with a device. With the increasing use of smartphones and Internet of Things (IoT) devices, Split Learning (SL) and Federated Learning (FL) have emerged as promising technologies that can tackle the authentication problem while protecting the user's private data. The SL distributed technology enables users with limited resources to complete neural network model training with server assistance, lessening the computational burden from the client side. In addition, FL aims to combine knowledge between different nodes collaboratively. The privacy and security of the user's data are ensured in both approaches, as only the models' weights are shared with a server. This study employs a cluster-based approach using split learning and federated learning techniques to improve the efficiency and robustness of training Machine Learning (ML) models. We compare the approaches' performance to baseline methods and demonstrate their advantages using the UMDAA-02-FD face detection and MNIST datasets. Our findings show that combining both technologies achieves high accuracy in continuous authentication scenarios while maintaining user privacy. These results highlight the importance of SL and FL in cybersecurity, enabling continuous authentication and demonstrating their potential to revolutionize how we address security.
Proton beam therapy is an advanced form of cancer radiotherapy that uses high-energy proton beams to deliver precise and targeted radiation to tumors. This helps to mit-igate unnecessary radiation exposure in healthy ...
Proton beam therapy is an advanced form of cancer radiotherapy that uses high-energy proton beams to deliver precise and targeted radiation to tumors. This helps to mit-igate unnecessary radiation exposure in healthy tissues. Real-time imaging of prompt gamma rays with Compton cameras has been suggested to improve therapy efficacy. However, the camera's non-zero time resolution leads to incorrect interaction classifications and noisy images that are insufficient for accurately assessing proton delivery in patients. To address the challenges posed by the Compton camera's image quality, machine learning techniques are employed to classify and refine the generated data. These machine-learning techniques include recurrent and feedforward neural networks. A PyTorch model was designed to improve the data captured by the Compton camera. This decision was driven by PyTorch's flexibility, powerful capabilities in handling sequential data, and enhanced G PU usage. This accelerates the model's computations on large-scale radiotherapy data. Through hyperparameter tuning, the validation accuracy of our PyTorch model has been improved from an initial 7 % to over 60 %. Moreover, the PyTorch Distributed Data Parallelism strategy was used to train the RNN models on multiple G PU s, which significantly reduced the training time with a minor impact on model accuracy.
Knowledge Graphs typically suffer from incompleteness. A popular approach to knowledge graph completion is to infer missing knowledge by multi-hop reasoning over the information found along other paths connecting a pa...
详细信息
Predicting the end-of-life or remaining useful life of batteries in electric vehicles is a critical and challenging problem, predominantly approached in recent years using machine learning to predict the evolution of ...
详细信息
XAI refers to the techniques and methods for building AI applications which assist end users to interpret output and predictions of AI models. Black box AI applications in high-stakes decision-making situations, such ...
详细信息
Federated learning extends the centralized machine learning architecture by enabling data privacy for its providers. The distributed structure of the emerged federated architecture imposes a problem of the data being ...
详细信息
ISBN:
(纸本)9781665435413
Federated learning extends the centralized machine learning architecture by enabling data privacy for its providers. The distributed structure of the emerged federated architecture imposes a problem of the data being not independent and identically distributed (non-IID), which drastically affects the performance of the learning process. While the majority of the recent works in the federated learning domain have accepted this limitation, only a few scholars addressed the non-IID problem straightforwardly. Nevertheless, these works lack the fundamental analysis of the data’ IIDness, and/or contradict the privacy feature of the federated learning paradigm. In this paper, we focus on evaluating the harmony of the participants by studying their data distribution and calculating their level of compatibility. The devised tool, in this work, is an assessment technique integrated within the federated learning framework to analyze the data distribution among the trainers. Our proposed method is proven by experimenting with several scenarios, and results show that our utility can fairly assess the selected participants before initiating the learning process.
XAI refers to the techniques and methods for building AI applications which assist end users to interpret output and predictions of AI models. Black box AI applications in high-stakes decision-making situations, such ...
详细信息
暂无评论