Fall has traditionally been my toughest term. It’s the term I traditionally teach an overload (of sorts), so I don’t usually have any students working for me, and research is done in bits and snatches whenever I get 20-30 minutes of free time (such as, er, during faculty meetings). That said, research is Still Happening, even if some weeks it’s only happening in my head or as a wistful item on my to-do list.
The big lab news is that I am presenting 2 papers next month at a conference and co-located workshop (QShine and AAA-IDEA, respectively). This is exciting because both papers represent major steps forward in the research agenda of the lab.
The QShine paper is something that came from the work of 2 students in the summer of 2008. (The students are co-authors on the paper, too.) It is a major proof-of-concept of something that our lab has been speculating about and alluding to for, oh, maybe 4-5 years now: is real-time prediction of streaming video quality possible? The students’ work demonstrated that it may in fact be possible, because we can get a pretty accurate idea of video quality from pretty small portions of the video stream (10 seconds). So these results, while preliminary, are pretty significant for our lab.
The AAA-IDEA paper is interesting because it is my first invited paper (yay!), and because it represents a full-circle point for me. Back in 2002-2003, while I was still at HP Labs, I wrote a paper discussing a proposed architecture for a streaming video quality assessment system. This paper became the basis for my subsequent research work. In the position paper, I revisit my design decisions and comment on what I got right and what I had to amend. It’s been a while since I’ve had the chance to think deeply about architectural issues, and the act of writing the paper rekindled my interest in some long-abandoned research questions. I’m thinking I’m going to get back to some of these big-picture, system questions as a regular part of my research agenda.
Looking ahead to Winter term, I’m planning on a new round of user experiments for another conference paper, possibly a journal article too. I may also hire one or more students work specifically on the real-time stuff. And I’m thinking ahead to summer and possible projects for our new CBL fellows, as well as trying to figure out how many students I can and want to support in my lab next summer.