RNNVis: Understanding Hidden Memories of Recurrent Neural Networks

Yao Ming, Shaozu Cao, Ruixiang Zhang, Zhen Li, Yuanzhe Chen, Yangqiu Song, and Huamin Qu


RNNVis is a visual analytics tool for understanding and comparing recurrent neural networks (RNNs) for text-based applications. The functions of hidden state units are explained using their expected response to the input texts (words). It allows users to gain more comprehensive understandings on the RNN’s hidden mechanism through various visual techniques.


This project has used the following dataset:

Penn Tree Bank: [data source] [paper]

Yelp Data Challenge: [data source]

A collection of Shakespeare’s work (used in supplement case study): [data source]


To appear in Proceedings of VAST 17. [preprint]



[VIS17 Preview]


RNNVis is under active development. If you have any comments or suggestions, feel free to open an issue.