Wednesday, January 10, 2024

Guest Post by Mohammad Hajiaghayi on SODA Business Meeting 2024

This is a guest post Mohammad Hajiaghayi on the SODA business Meeting from 2024. The meeting was held in Alexandria Virginia, Jan 7-10, 2024. The following meetings were held jointly:

SODA: ACM-SIAM Symposium on Discrete Algorithms

ALENEX: SIAM Symposium on Algorithmic Engineering and Experiments

SOSA: Symposium on Simplicity in  Algorithms

The website for SODA 2024 is here.

And now the guest post:


SODA Business Meeting Report:

After a few years of being unable to attend the SODA business meeting due to factors like Covid, I was able to participate this time, especially since it took place nearby in Alexandria, VA. So I thought it's a good idea to provide a brief report on the business meeting.

DISCLAIMER: Please be aware that this report is solely based on MY PERSONAL recollection of the meeting and is intended solely as an informal update.

The session started with Piotr Indyk, the chair of the SODA Steering Committee, presenting a slide outlining the meeting agenda. Following that, SODA24 PC Chair David Woodruff gave a brief overview of SODA24's numbers, including the record-high number of submissions, exceeding 650. Following that, Seth Pettie, SOSA PC Chair, shared a concise overview, highlighting the enjoyable experience of being a PC in SOSA, where papers are typically short, often less than 15 pages, emphasizing simplicity in algorithms.Best paper awards were then presented for both SODA and SOSA, detailed  here  and Greg Bodwin's An Alternate Proof of Near-Optimal Light Spanners for SOSA. Pettie bragged about a distinctive trophy for SOSA's best paper authors, in contrast to SODA, which only offers paper acknowledgments for awards. :)

Looking ahead, SODA is scheduled for New Orleans next year, followed by Vancouver in the subsequent year.

The majority of the discussion focused on the acceptance rate for PC submissions, particularly as David mentioned since the process is double-blind, and the PC is sizable, allowing PC members to submit can enhance the overall paper quality. Initially  at around 42%, significantly higher than the average acceptance rate of 29%, the Steering Committee of SODA requested a reduction to 35% for PC submissions this year.

In the past year, the fixed rate was 30%, prompting a debate on determining the appropriate acceptance rate. Nikhil Bansal shared insights into the acceptance rate of FAMOUS AUTHORS (authors with more than 20 papers in SODA) over the past few years, revealing no significant changes. While some argued against controlling the PC acceptance rate, I think the summary was to examine the average acceptance rate of PC members who were not part of the PC in the past few years. This evaluation could potentially guide the establishment of a threshold rate. This aligns with a truthful mechanism that I discussed with David Woodruff before the session.

During the discussion of PC submissions, Mikkel Thorup raised a subtle point about double-blind submissions,  where the authors might subtly boast about their results compared to their own previous work, potentially resulting in incremental advancements, as improving upon others is highly regarded in our community. I believe PC members, rather than reviewers, should diligently verify such claims for all papers, particularly regarding previous work, without the need for author names.  Additionally, if necessary, the PC could inquire with the PC chair who has access to the authors' names for cross-referencing with previous works.

Last but not least, in the concluding business segment of the meeting, I expressed my satisfaction with the adoption of double-blind submissions as the default for SODA/STOC/FOCS, consistently advocating for it in theory conference business meetings and behind the scenes (e.g., refer to my earlier conversation with SODA founder, the late David Johnson, here.) Despite experiencing increased difficulty in getting my own papers accepted in a double-blind SODA, being ranked second among SODA Famous Authors by DBLP (as defined by Nikhil), I firmly believe that double-blind submissions are the correct and fair approach. This is based on the principle that an author's name should not influence the fate of a scientific claim. I also voiced my happiness at the extensive discussion on fair policies for PC authors and even advocated for transparent criteria in selecting PC Chairs for SODA/FOCS/STOC. I stressed in particular the significance of valuing contributions to the conference, commonly measured by the number of papers, in selecting PC Chairs as those who dedicate effort to a conference should have a say in the process. This stands in contrast to relying solely on potentially biased personal opinions when such a quantitative criterion does not exist. Regardless, I highlighted the need for transparency in any criteria used to select a PC Chair, who can contribute further to diversity by forming the PC. I suggested discussing these criteria in business meetings, akin to our approach to other topics, to foster fairness, a sense of belonging, and prevent the conference from becoming an Insider Club (the term used by Martin Farach-Colton during the meeting), ensuring inclusivity.

No comments:

Post a Comment