understandability

This project demonstrates the use of generic bi-directional LSTM models for predicting importance of words in a spoken dialgoue for understanding its meaning. The model operates on human-annotated corpus of word importance for its training and evaluation. The corpus can be downloaded from: http://latlab.ist.rit.edu/lrec2018

In this research project, we aim to create an environment to gather structured data about machine learning experiments in order to analyze data and algorithmich dependencies.