Members of the SLIM group have some long papers and workshop papers:
David McNeill’s paper titled Predicting Human Interpretations of Affect and Valence in a Social Robot on identifying how humans interpret affect (i.e., emotion) and valence in the Anki Cozmo robot has been accepted to the RSS Conference in Freiburg, Germany. David painstakingly recorded Cozmo’s over 940 pre-scripted animations, crowd-sourced emotion labels for them, then we looked into approaches to automatically determining the perceived emotion based on the robot’s audio, face, and internal state features. We found that splitting up 16 possible emotions into positive and negative valence pairs yielded reliable results that could be used as informative features for other human-robot interaction tasks.*
*Thanks to NVIDIA for donating the GPU we used to run the experiments!
Dr. Kennington also has papers accepted to the IDC Conference in Boise, Idaho (including a long paper about query formulation for child-directed search tools, a workshop paper on spell checkers for children, and a workshop paper about evaluation of search engines for children).
Dr. Kennington also co-authored a paper entitled Privacy Disclosures Detection in Natural-Language Text Through Linguistically-motivated Artificial Neural Networks. primarily worked on by Nuhil Mehdy and Hoda Merhpouyan which was accepted to the SPNCE Conference in Tianjing, China. The work presented in that paper explored how we can use natural language processing techniques to identify text that has private disclosures in them.