Today marked the opening of Ubicomp 2008. There are approximately 341 attendees, of whom 154 are students. 76 people signed up for the workshops. There were 226 paper submissions (160 full papers, 66 notes), of which 42 were accepted, making the acceptance rate 19%.
Attendees were urged to make their presentations available on SlideShare, and it was recommended that ubicomp2008 be used as a standardized tag for the conference across all sharing sites (SlideShare, Flickr), etc) (hmm, folksonomy breaks down 🙂 )
The first session today was on activity recognition. While the papers were mostly about human motion (in buildings or moving around), PET’s transportation data analysis is clearly a type of activity recognition and presumably the same sets of techniques (Hidden Markov Model, Conditional Random Field) will be relevant, as will the methods of evaluation (such as a confusion table). Good to know there’s a literature I can look towards there.
During a paper on activity recognition in the home, the presenter mentioned providing software for annotating activities to provide ground truth for the activity data set. This might be helpful in my annotation of transportation activities.
Another interesting talk was a fitness motivation system using a glanceable display on a cell phone. The idea is that as the users work towards their fitness goals (recorded either manually through a journal or automatically via a fitness sensor), a garden of flowers is placed as the background on their cell phone interface. This provides a constant reminder of how they are doing in pursuit of their goals. This got me thinking that maybe PET could benefit from a similar glanceable display of GHG emission estimates.
I’m looking forward to the talk tomorrow on indoor inertial tracking.