Confessions of a researchaholic

This LinkedIn post reminded me of the talk I gave in SIGGRAPH Asia 2014 about how to make a SIGGRAPH paper.

That owl image came from here, resonating with my earlier phase of learning how to draw around that time, in addition for being a funny point of analogy for writing research papers.

Retrospect can sometimes feel like a time travel to reconnect with my past self.
My passion for SIGGRAPH peaked around 2011 to 2014, and I have been looking beyond that for interesting or important problems to work on.

When I was a PhD student around the turn of the millennium, SIGGRAPH, despite being the top venue in visual computing, is considered as a conference and thus the published technical papers received a lower count/weight than journal papers in traditional academic rating.
To remedy this problem, starting from 2002 SIGGRAPH technical papers were published as special issues of ToG (ACM Transactions on Graphics), which is a journal.

Now, about 20 years later, people were concerned about the author workload and review speed for ToG, compared to some machine learning and computer vision conferences which tend to have lower requirement for evaluation and faster review/publication cycles.
To remedy this problem, starting from 2022 SIGGRAPH will start accepting a conference-track of papers.

This is one example of the general phenomena where people repeat cycles of identifying problems, devising solutions, causing other (sometimes previous) problems, ad infinitum.

Update: Aaron Hertzmann (the main architect for the conference track) has a recent post about expectation creep that is worth taking a look.

Yesterday I received some automatic notification about an accepted demo to SIGGRAPH Asia Real-Time Live without realizing that I have a submission there (a co-author of a previous UIST paper made the submission without telling others).

https://twitter.com/liyiwei/status/1455674543930953728

I mentioned this in a group meeting today, and a colleague told me that he has a previous intern who submitted their project without including his name and he found out by a review request for that submission.
🙂
So my case is quite benign in comparison.

I remember (20+ years ago while I was still in school) during a faculty seminar, Mark Horowitz couldn’t attend, and Bill Dally presented his research by skipping all results slides by simply saying “good results!” and called it a day.

As someone coming from the SIGGRAPH-side of research, I always wish that all CHI user study sections can contain just one sentence: happy users and good results! (instead of pages of descriptions and statistics which, honestly, I never read).

https://twitter.com/liyiwei/status/1434981016800690180