Summit is an interactive system that scalably and systematically summarizes and visualizes what features a deep learning model has learned and how those features interact to make predictions.
🏔️ Live demo: fredhohman.com/summit
📘 Paper: https://fredhohman.com/papers/19-summit-vast.pdf
🎥 Video: https://youtu.be/J4GMLvoH1ZU
💻 Code: https://github.com/fredhohman/summit
📺 Slides: https://fredhohman.com/slides/19-summit-vast-slides.pdf
🎤 Recording: https://vimeo.com/368704428
Summit: Scaling Deep Learning Interpretability by Visualizing Activation and Attribution Summarizations
Fred Hohman, Haekyu Park, Caleb Robinson, Duen Horng (Polo) Chau
IEEE Transactions on Visualization and Computer Graphics (TVCG, Proc. VAST'19). 2020.
For a live demo, visit: fredhohman.com/summit.
For the Summit notebook code, Visualization: summit-notebooks.
For the Summit data, visit: summit-data.
Download or clone this repository:
git clone https://github.com/fredhohman/summit.gitDownload the data from summit-data:
git clone https://github.com/fredhohman/summit-data.gitPlace summit-data's data folder in the top level of the summit repo.
Then, within summit run:
npm install
npm run build
npm run startSummit requires npm to run.
MIT License. See LICENSE.md.
@article{hohman2020summit,
title={Summit: Scaling Deep Learning Interpretability by Visualizing Activation and Attribution Summarizations},
author={Hohman, Fred and Park, Haekyu and Robinson, Caleb and Chau, Duen Horng},
journal={IEEE Transactions on Visualization and Computer Graphics (TVCG)},
year={2020},
publisher={IEEE},
url={https://fredhohman.com/summit/}
}
For questions or support open an issue or contact Fred Hohman.