Summarize with Wikum
By harnessing the power of people backed by state-of-the-art human-computer interaction and machine learning techniques, users get complete control over the collaborative summarization process.
By harnessing the power of people backed by state-of-the-art human-computer interaction and machine learning techniques, users get complete control over the collaborative summarization process.
Through Wikum's easy-to-navigate tree visualization, you can get a snapshot of the various comments, replies, and existing summaries of the entire discussion thread.
You can click, drag-to-select, Command/Ctrl-click, and expand or collapse replies using the visualization in order to select the discussion level at which you want read or summarize.
After selecting a comment or a subtree or group of comments, you can write a summary of these comments, with the help of automatically highlighted important sentences, and features to add citations. You can also tag, hide, or automatically cluster comments to organize the discussion.
Wikum uses a recursive summarization approach, allowing you to build summaries on top of other summaries, your own or others. That is, once you've written a summary of a subset of comments, another summary can summarize a larger subset of comments, taking into account your summary.
With the ability to flag and edit summaries as well as see the entire edit history, Wikum empowers the community to keep track of who has participated and the overall quality of summaries.
By adding auto-suggested or user-generated tags as well as automatic clustering, similar comments can be grouped together for further action.
Flags empower users to highlight when summaries need further work due to incompleteness, non-neutral language, or writing quality.
From a wiki philosophy, Wikum permits users edit any existing summaries to make them better. User permissions are also controllable for those wishing a private workspace.
Wikum keeps a log of all edits so that the community can see who has edited each summary and keep a record of all activities.
Wikum has both public and private workspaces for summarizing and creating discussions. Explore and contribute to public summaries of discussions or log in to create a private workspace for yourself and your collaborators.
Currently, Wikum supports pulling discussion from Wikipedia Talk Pages, Disqus forums on articles, Reddit, Decide Madrid, JOIN Taiwan, and mbox email files.
We believe in open source software and invite you to collaborate with us on our Github repository.
We have written research papers about Wikum and are actively conducting additional studies.
Wikum researchers at Computer Science and Artificial Intelligence Laboratory, MIT
Creator
Creator
Current and former graduate and undergraduate students in the Haystack Group at MIT CSAIL.
Contributor
Contributor
Contributor
Wikum Collaborators at Collective Intelligence for Democracy 2017
Medialab Prado, Madrid.
Contributor
Contributor
Contributor
Contributor
Contributor
Contributor
Contributor
