+ Note: This app is in pre-release mode for ICML 2017. Please send feedback and bug reports to icml17app@gmail.com or file an issue on our github project page. +
- This app helps navigate the rich program of ICML 2017. The app enables users to quickly find interesting papers, publicly comment them, register their interest using 'like' and 'bookmark', and browse what others are interested in. In addition, users' preferences feed a personalized paper recommendation engine.
-
+ This app helps navigate the rich program of ICML 2017. The app provides an easily
+ navigable schedule, enables users to quickly find interesting papers, publicly
+ comment them, register their interest using 'like' and 'bookmark', and browse what
+ others are interested in. In addition, users' preferences feed a personalized paper
+ recommendation engine.
+
+
Papers you choose to bookmark or like are made available through your "Library" page.
+The difference between bookmarks and likes is that liked papers are fed into the recommendation engine.
Recommendations are available through the "Feed" page and are updated every two minutes. On that page recommended papers are preceded by trending papers which shows some of the most popular papers.
+The recommendation engine is based on the collaborative topic regression (CTR) model of C. Wang and D. Blei.
+There are a few reasons one of which is that it can be challenging to keep up with all of the great work published at conferences. We believe that an easy to use app and a good recommendation engine can help surface interesting papers.
+Moreover, we built this app as a service to the community, we will open source the app. We would like others to use and even build on it. We believe that the app can also be used as a test bed for recommender engines. As such the app keeps detailed (and anonymous) logs for future offline analysis and model comparisons.
+The app was designed and built by a team of machine learning students and researchers spread across England, France, and Canada.