Seeing the Future by Mining the Past: The Exact Average UIST Paper
by Bernstein, M.S., with the largest contributions from eigenauthors Hudson, S., Balakrishnan, R., Myers. B.A., Feiner, S., Hinckley, K., Rekimoto, J., et al.
History repeats itself: computer science research continuously reinvents past work. Thus, to predict the future of UIST, we trained an n-gram language model on twenty years of previous UIST abstracts. We thus present the machine-generated exact average UIST paper, with small edits made for clarity and maximum humor:
Edgewrite is a big problem in current handheld browsing: we describe some common problems experienced by users that are hit. We describe the architecture we have implemented and deduce a general framework that provides on-demand, persistent prototype applications. We discuss how people-tagging is a low-cost off-the-shelf electroencephalograph (EEG) system. We hope to challenge the audience to creatively consider ways that would otherwise result in small thumbnails that are placed away from systematic noise sources which can identify 100% of dishwasher usage. Unlike previous work, we present examples that represent the structure of toolkits. The user glances at the possibilities of olfactory output devices: ubiquitous computing environments need to work together. The result in a user study shows that after the stimulus, the location of the user responds to an automatic graph layout system. Finally, we introduce three new interaction techniques.
I hope that someday I can give the exact average UIST talk.