Publications by authors named "Adam Spannaus"
Article Synopsis
- Machine learning models, specifically deep neural networks (DNNs), are increasingly used in decision-making alongside humans, emphasizing the need for reliable classifications.
- This paper highlights the use of DNNs to automate the extraction of cancer-related data from electronic pathology reports, while introducing new selective classification methods to improve accuracy and reduce the number of unreliable predictions.
- The proposed methods outperform existing models by achieving high accuracy with lower rejection rates, demonstrating their effectiveness in processing complex medical data.
The role of epidemiological models is crucial for informing public health officials during a public health emergency, such as the COVID-19 pandemic. However, traditional epidemiological models fail to capture the time-varying effects of mitigation strategies and do not account for under-reporting of active cases, thus introducing bias in the estimation of model parameters. To infer more accurate parameter estimates and to reduce the uncertainty of these estimates, we extend the SIR and SEIR epidemiological models with two time-varying parameters that capture the transmission rate and the rate at which active cases are reported to health officials.
View Article and Find Full Text PDF