APLAS 2023 Student Research Competition and Poster Session: Chair’s report
October and November were hectic for me: I was on the programme committees of CPP, ESOP, and TFP; there were AIM to attend and APLAS to help organise (but thankfully both took place at IIS, and I could sneak back to my office to take a breath when needed); and I bought tickets to too many concerts/recitals (the number of which surged after the border was reopened). I made it through somehow and can resume writing now, starting with this report before I forget everything.
The primary aim of SRC & posters is to promote unfinished works, in particular those by students (as well as give more exposure to works already published elsewhere if there is space, which there usually is). Another aim is for students to get familiar with writing (though restricted to extended abstracts), hallway chats (though in a much more structured form with posters in place of scribbles on napkins), and formal presentations. I regard the competition format more like a necessary evil, adopted just for arousing interests/excitement and justifying looking at such works in more detail. The negative elements (e.g., stress) should be mitigated as much as possible: for example, given that there are only four entrants this year, I decided to give everyone a prize, and let the entrants know this in advance to make the competition less stressful. On the other hand, the positive elements (e.g., excitement) should be amplified: for example, in addition to letting a small group of judges decide the prizes, I set up audience awards to motivate all the conference participants to engage in the event.
Audience awards
Inspired by music competitions, the audience awards were decided based on voting by conference participants. Since this was somewhat experimental for PL student research competitions, I started with just three award categories: the best SRC and non-SRC posters, and the best finalist presentation, aiming to encourage people to come to the poster session and the SRC finalist presentation session.
I think the key organising principle is to minimise the overhead of voting and make the audience willing to vote. For example, it’s important to keep the instructions within one bite: In an announcement reminding people to vote, I said ‘just take a look at the online voting form, which is very simple’ and ‘choose two posters and one presentation you like’. And a voter could vote for any two posters they liked without having to distinguish between SRC and non-SRC posters — this may be dangerous if one group is clearly inferior to the other and gets significantly fewer votes, but luckily this didn’t happen this time. Also (like any other new stuff) extra care must be taken to make sure that everyone knew about the awards: The conference check-in procedure should be thoroughly checked to ensure that ballots are distributed to registrants in one of the steps — I told the organising team about the ballots but didn’t check, and the ballots were forgotten until I checked-in myself. 😓 And voters need to be constantly reminded to vote from early on (or they may simply leave their ballot in their hotel and forget about it). In the end, we got about half of the registrants to vote.
The online voting form was implemented using Google Forms. The pros was convenience; the cons was that I couldn’t disallow voting multiple times in the same category, and had to come up with more complex rules that specify what to do in such cases, resulting in more complex instructions on the ballot (shown below). I wrote scripts to generate unique ballot IDs and QR codes that link to the voting form pre-filled with a ballot ID, and count the votes according to the more complex rules. But in the end everyone voted just once in each category, so the rules weren’t actually needed, and the summary produced by Google Forms gave the accurate results directly. Next time I’d just put the most important messages such as ‘choose two posters and one presentation you like’ prominently on the ballot, and leave the more complex instructions in small print.
If someone’s interested in setting up audience awards at a bigger conference, it might be a good idea to try and give a larger number of special awards designed to allow a more diverse range of participants to get one — see, for example, the awards given at the Chopin competition.
Selection committee
The selection committee was responsible for reviewing the extended abstracts and providing the first set of feedback to help the authors prepare their presentations. (We intended to accept all reasonable submissions, so the committee didn’t actually do much selection despite the name.) I wanted to have worldwide experts with extensive experience on the committee to make the feedback high-quality and attractive to participants — ideally, this committee should even be slightly more senior than a usual programme committee to be able to give good advice to students. On the other hand, APLAS was a more regional conference, and I surmised (correctly) that people would be less willing to accept my invitation to serve. Therefore I planned to assemble a committee that was double the usual size to reduce the workload, so that the task would look like something easy enough to do. Half of the committee members were based in East Asia, to make it more likely for more committee members to come to the conference and continue as a judge. Alas, more people declined my invitation than I anticipated, and eventually I still had to turn to more familiar people for help.
This year’s APLAS was in Taiwan, which created a problem for Chinese passport holders, who may need an unusually long time to apply for an entry permit. I was forced to bring the notification date forwards by first making acceptance decisions based on quick skimming, and later finishing the reviews. I’d recommend avoiding this unless it’s absolutely necessary, because first impressions can indeed be deceiving, and some problems become clear only after proper reviewing.
Judges
The judges were selected from the conference registrants after the early registration deadline. As mentioned above, selection committee members who came to the conference were automatically invited. For the other judges, I tried to match their expertise with the entrants’ works so that for each work there’s at least one judge who was more likely to understand it well and was at least potentially sympathetic. However, I ended up with a group of judges that were predominantly Japanese, and couldn’t find, for example, a Korean judge since all the Korean registrants seemed fixated on program analysis and automated program verification, which none of the entrants worked on.
Usually the judges’ task at the poster session is to select the finalists, but since there were only four entrants, it turned into intensive discussion, and we spent like 20–30 minutes with each entrant. I sent (possibly revised) extended abstracts and posters to the judges two weeks before the conference and asked them to familiarise themselves with the works and prepare some questions in advance; I also told them that we’d be evaluating the overall performance rather than just the research. The entrants were also informed beforehand that when the judges arrived, they’d be asked to give a 5-minute introduction, followed by questions from the judges. It went well in general, but we should probably have allowed judges to serve as interpreters for those entrants who weren’t a native English speaker, especially for critical questions and answers.
At the judges’ meeting, I asked for concrete comments about each entrant (as opposed to just vague expressions such as ‘I feel this one should get the first prize’); these were potentially valuable feedback for the entrants too, and maybe it would’ve been a good idea to summarise these and send them to the entrants. One particular difficulty that doesn’t appear in usual PC discussions was that we were supposed to focus on the entrants’ own work, excluding, for example, choice of research direction, which may be set by the supervisor, or lack of theoretical development, which may be assigned to a more senior colleague. Taiwanese white tea was served during the meeting — I couldn’t possibly miss this rare opportunity chairing an event at IIS. And after the meeting there was a judges’ dinner at a nearby Taiwanese restaurant, which is a favourite of mine.
Award announcement
The prizes and awards were presented at noon after the judges’ meeting in the previous evening. This posed a somewhat severe time restriction on the preparation of certificates. In particular, it wasn’t easy to typeset the certificates because the titles of the works can be quite long; APLAS hired a firm to do some visual design (in particular the impressive logo), and they were absolutely helpful by fine-tuning the certificates for every title while leaving the prize part blank, which was easier to handle by ourselves. Right after the judges’ meeting, my administrative assistant Po-Fen finalised all the certificates and got them printed at a nearby shop, ready for signing by Shin and me next morning.
We put the prize money in red envelopes with a Taiwanese theme and clever design found by Po-Fen: the front shows either a Chinese character or an English word depending on the orientation.
At the award announcement, I used just one slide each for the audience awards and the SRC prizes to save time and avoid unnecessary ranking of the entrants — everyone was a winner in the end, and should celebrate together.
The two triumphant climaxes of the fourth movement of Mahler’s first symphony were used as background music for the two slides respectively.
