From a podcast I bumped into this paper about the economics job market rumors (EJMR) forum, which is a place for anonymous discussions about the job market for economists.
I was initially interested in the methodology (in particular reverse hashing) and results in the paper, but then I found the EJMR site itself to be even more fascinating as it sheds more light on human nature than economics (which is already a relatively “well to do” branch of social/political science).
2023-07-23
EJMR
2023-06-15
Journal AE versus paper committee
After serving as an associate editor (AE) for a few journals (e.g., ToG, CGF, TVCG) and paper committee members for a few conferences/journals (e.g., SIGGRAPH, SIGGRAPH Asia, Eurographics), here is my personal take on the similarities and differences between the two roles.
Similarities: a main (and probably the most important) task for both is to recruit and manage reviewers.
Differences: an AE does not have to read the papers (even though I read every single one so far just to make sure my final recommendation is sound) and can decide the final recommendation without consulting other AEs, while a paper committee member has to read all the papers and present the recommendations in a committee meeting before the final decisions can be made.
2023-06-12
How to invite reviewers
Are you a paper committee paper or journal associate editor (AE) who want to increase the likelihood of reviewers responding positively to your invitations?
Here are some tips based on my own experience:
Instead of a machine-generated generic email, craft a custom message telling the candidate reviewer why you think they are a good fit for the paper, such as their background respect to the topic or specific components of the paper.
You should have this information already when deciding to invite that particular reviewer.
Add some personal greetings such as if the reviewer has a recent job change or life event, a new paper, or someone you both know, which could be found on their social media or personal website.
Do so particularly if you know the reviewer (or someone close to them like their collaborator/advisor) personally.
If you already read the paper, provide some estimation of the time commitment required for reviewing the paper, e.g., “I think it will take you about 2 hours to review this paper”.
If you have accepted review requests from that reviewer in the past, mention this as well as a hint of reciprocity.
This is another reason to accept review requests if you have the bandwidth.
2023-04-26
Sketch creativity
It was told that I would be too run down to have meetings after the surgery, but I managed to come back 10 minutes prior to the start of the seminar and gave a live talk without any issues.
But just as a back-up plan, I also pre-recorded a rehearsal.
2023-04-23
PhD student recruiting philosophy
Throughout my research career I have been very conservative in recruiting PhD students, especially for those whom I would be the (de facto) advisor.
(I am a bit more relaxed for hiring interns as the collaborations are shorter term and thus the risks are lower.)
I prefer to have deep involvement for each student and project, and the cost of having a student not suitable for independent research is higher than the risk of occasionally passing on a top candidate.
However, there are other professors/researchers out there who have been very successful in managing large groups.
So definitely go for that if your style is like a VC incubating startups, you have enough funding, and your projects require teamwork (e.g., one student probably is not going to build a new operating system or programming language).
2023-04-18
Single review paper committee member
It turns out that I only need to review one paper as a committee member for a top venue thanks to “the large number of IPC members”. (Cannot go any lower unless we want to list someone who does not review any paper as a PC member.)
Expanding the committee to include more members can lighten the individual workloads, but also reduce the visibility of each PC member for calibration purposes.
I wonder what the sweet spot would be as a function of the number of submissions and the number of PC members.
1 is definitely slow while 20 is high (based on my recent SIGGRAPH/SIGGRAPH-Asia PC experiences), while 4 to 8 feels about right to me.
2023-04-16
Outside the mandates
Recently, during the preparation for an invited talk on a topic I have been working on for a while, I realized that even though I am familiar with the topic, I have yet to organize my thoughts into a coherent story, and I am glad that this invited talk motivated me to do just that.
Giving academic talks is outside my current main job duties and thus I wouldn’t be as incentivized to do that as teaching courses as an university professor or shipping products as an industry researcher.
But this experience made me realize that there are activities outside our official job mandates that can still benefit our careers in a less direct way or in a longer run.
2023-04-09
The potential of being on both SIGGRAPH and SIGGRAPH Asia technical papers PC in the same year
I have already served on the SIGGRAPH 2023 technical papers program committee and just got asked to do that for SIGGRAPH Asia 2023 as well.
I have not yet decided (at this moment of writing) whether to turn down the latter or do both at the same year, but either will be the first time in my career, and I will feel bad about potentially disappointing the paper chairs or spending another month’s worth of evenings of reviewing papers.
2023-04-06
Benefits of reviewing papers
Reviewing papers can take a lot of time and effort, but also has the following benefits:
It is a good way to build reputation and relationship with the research community by writing an informative opinion on time and participating in the discussions to reach the final recommendation.
If you submit to a venue, it is only fair to review for that venue, especially if you have complained that your submissions have not received good/fair reviews due to lack of expertise or efforts from the reviewers.
It is a good way to learn about the latest research, as you might not feel pressured to keep up to date with the latest research in your day job otherwise.
Writing a good review (see above) requires deep enough understanding of the paper beyond the usually cursory glance at the published papers.
It is a good way to learn how to get your own submissions accepted in the future by looking at how decisions are made and what the reviewers are looking for.
Accept only reviews that you think you have the time and expertise to do a decent job, and do not feel bad about declining reviews, preferably within a few days of receiving the request with suggestions for alternative reviewers so that the organizers can find replacements in time.
Other things being equal, prioritize review requests from higher-quality venues for which your reviews would tend to make a relatively bigger impact.