M.S. AAI Capstone Chronicles 2024

Figure 3 High level architecture of the Stacked Convolutional Transformer model

The SCT is composed of three distinct modules: the Convolutional Module, the Transformer Module and the Classifier Module. Together, the modules perform end-to-end sepsis classification of a patient time-series data stream. Within the Convolutional Module, a total of three one-dimensional convolution operations are performed with two max pooling

273