AbstractsViewer: Local and Global Interpretability in Text Corpus Exploration

AbstractsViewer is a system for exploring text corpus. The system supports tasks including identifying a target document, finding related documents and identifying a theme/concept in the text corpus. Currently, we have research papers from Visualization and Robotics conferences and articles from New York Times dating from 1987 to 2006.

The beginnings of the User Guide is available on this site.

We also have a demo available online.

Go to main components for a visual overview of the biggest elements of the AbstractsViewer.

Our user guide provides a basic overview of how to use the AbstractsViewer, such as using it to find similar papers and using it for interpretability.

The walkthrough provides an example of a usage of the AbstractsViewer.

AbstractsViewer is still in experimental process, so please don’t share it in public for now.