CogSci 2018 Tools

CogSci 2018 bespoke paper recommendations

Brought to you by the LUCID graduate training program

 So many talks, so little time! Which ones should you prioritize? Try the CogSci 2018 bespoke paper recommendation system. Just check the topics that interest you, indicate your interest in 25  papers based on title and/or abstract, and receive your own personalized top-10 list. Provide additional judgements to improve prediction accuracy.

NOTE this is a system prototype. We are interested to know (1) whether the system works smoothely without substantial delay and (2) whether the recommendations seem appropriate. Please email your feedback to with the subject line: Paper Recommender

Graphing relations among CogSci 2018 talks

To help organize sessions this year, we computed latent topics for the discipline from publications submitted by reviewers and editors, then represented each accepted talk as its posterior probability over topics. To visualize the conceptual relations among talks, the resulting similarities were used to generate a 5-nearest-neighbor graph, which you can view by clicking one of the links below.  Hover over nodes to see both the paper title and nearest neighbors on thew graph. Click a node to highlight it and its nearest neighbors. Drag nodes to move them. Colors indicate graph communities identified by the spinglass algorithm. The static graph shows an arrangement generated by the Fruchterman-Reingold algorithm. The dynamic graph runs the settling process for a force-directed layout from an initial random assignment. Pull the nodes around to see what happens to the layout!

Click here to see the graph!