今年在只有15%的录用率下，祝贺孙菱同学的研究成果被正式录用，迎来了开门红。信息的传播预测一直是领域中的关键性问题，该研究采用了一个具有存储能力的序列超图注意力网络，实现了对节点细粒度传播能力进行预测，再次祝贺。

Dear Yuan Rao,

Congratulations! We are delighted to inform you that your paper 4362 (MS-HGAT: Memory-enhanced Sequential Hypergraph Attention Network for Information Diffusion Prediction) has been accepted for presentation at the Thirty-Sixth AAAI Conference on Artificial Intelligence (AAAI-22). This year we received a record 9,251 submissions, of which 9,020 were reviewed. Based on a thorough and rigorous review process we have accepted 1,349 papers. This yields an overall acceptance rate of 15%.  

To view your final reviews please visit the CMT website:  https://cmt3.research.microsoft.com/AAAI2022/.

The AAAI-22 review process was systematic and thorough, involving approximately 7,266 PC members, 1,120 SPC members, and 288 Area Chairs. We used a two-phase review system that led to much more in-depth consideration of papers that were in the top 50%. In Phase 1, we solicited two reviews per submission and rejected those that did not receive sufficiently positive reviews. We also offered a “Fast Track” for NeurIPS papers that were rejected after the AAAI Main Track submission deadline that received above-average scores.

Almost all Main and AISI Track papers that made it to Phase 2 received at least three reviews across both phases; over 85% received four or more reviews. Virtually all Fast Tracks papers received at least two reviews in Phase 2; over 80% received four or more new reviews. Additional reviews were solicited as needed, with many papers receiving five, six, or even seven reviews. Each paper also received an additional meta-review written by a Senior Program Committee member. For over 90% of submissions--even in most cases where papers received unanimous review recommendations--reviewers, SPC and AC members discussed the merits of the paper and the authors’ response to the reviews. Discussions often led reviewers to change their minds; in such cases they often but not always changed their scores. Area chairs oversaw the review of each paper, providing expertise and guidance where needed, handling issues that arose, consulting on meta-reviews, and ultimately making recommendations. Finally, the AAAI-22 Associate Program Chairs and Program Chairs individually considered each paper, flagging hundreds for further discussion and where warranted departing from AC recommendations to ensure that each paper was held to a uniform standard.

We note that decisions did not solely depend on average reviewer scores; we did not decide acceptance and rejection using a numerical cutoff. Shallow reviews were discounted in favor of substantive reviews.  Senior members of the committee (SPC members, area chairs and program chairs) also weighed in at different stages of the review process as needed. Our process paid particular attention to papers receiving borderline or weak accept ratings or SPC or AC recommendations that diverged from the reviewer ratings.

We understand that each paper is the product of careful thought and hard work on the part of its authors. AAAI’s high bar means that we have to inevitably reject many papers that make meaningful contributions to their respective fields. We made every effort to avoid conflicts of interest, minimize implicit and explicit bias, and ensure the integrity of the review process. The program committee devoted tremendous effort to identify the small fraction of papers that were accepted, we recognize that no process is perfect. Despite our best efforts, the outcomes can sometimes be influenced by random and subjective factors, and some reviewer comments may be inaccurate and some decisions could have justifiably have been different than they ended up being. While we tried our best to ensure that the reviews and meta-reviews reflected an accurate and complete picture of the program committee’s deliberations, we did not always succeed in doing so. 

Nevertheless, we urge you to take very seriously the feedback provided by the reviews and meta-reviews in preparing a preprint that will be included in all electronic conference materials. This preprint is due December 9th 2021, which is a firm deadline. Please prepare this preprint in the submission format (but removing line numbers if you used them for submission). As before, you are allowed 7 pages of technical content plus two pages for references and acknowledgements; if you choose, you may additionally purchase up to two pages of technical content. Please watch your email for detailed instructions for preparing your preprint in a separate message in the next few days.

The final camera ready version of your paper is not due until March 22nd, 2022, providing you with further opportunity to improve your manuscript based on the comments you receive at the conference.

AAAI-22 will be held from Feb 22nd to Mar 1st. AAAI-22 will be held in person in Vancouver. Plans are currently underway to accommodate authors who cannot travel to Vancouver because of COVID travel restrictions, visa issues, or other reasons. These plans will be finalized in mid-December.

At least one author of every accepted paper is required to register for and attend the conference to present the paper; we encourage all authors to attend if possible. Registration information and the online form will be available soon, and you will receive a separate message with details at that time. In addition to the outstanding technical program, AAAI-22 will include thought-provoking panels and invited talks (this year, by Francesca Rossi; Gil Alterovitz; Marta Kwiatkowska; and Patrick Schnable), and exciting co-located events including IAAI-21 and EAAI-21. Information on the conference will be continually updated at https://aaai.org/Conferences/AAAI-22/.

Once again, congratulations!

Sincerely,

Vasant Honavar, Pennsylvania State University, AAAI-22 Program Co-Chair
Matthijs Spaan, Delft University of Technology, AAAI-22 Program Co-Chair

Jennifer Dy, Northeastern University, AAAI-22 Associate Program Co-Chair
Timothy Hospedales, University of Edinburgh, AAAI-22 Associate Program Co-Chair
Nihar Shah, Carnegie Mellon University, AAAI-22 Associate Program Co-Chair
Vivek Srikumar, University of Utah, AAAI-22 Associate Program Co-Chair

Sriraam Natarajan, University of Texas at Dallas, AI for Social Impact Special Track Chair
Fei Fang, Carnegie Mellon University, AI for Social Impact Special Track Chairs