Prof. Ke YI's Group Won Triple Awards at ACM SIGMOD/PODS 2022

Prof Ke YI's group have received three awards at this year's ACM SIGMOD/PODS conference, held during June 12 to 17 in Philadelphia, PA, USA:

• SIGMOD Best Paper Award
  R2T: Instance-optimal Truncation for Differentially Private Query Evaluation with Foreign Keys
  Wei Dong (HKUST), Juanru Fang (HKUST), Ke Yi (HKUST), Yuchao Tao (Duke), Ashwin Machanavajjhala (Duke)

  Differential privacy in recent years has attracted significant interests in both academia and industry due to growing privacy concerns and regulatory requirements. While many specific problems have been solved, it remains a challenging problem to handle general SQL queries in a differentially private manner. This paper introduces the first differentially private mechanism for answering general select-project-join-aggregation queries in a database with foreign-key constraints. It achieves a strong notion of optimality, while still simple enough to be implemented on top of any RDBMS and an LP solver.

• SIGMOD Best Paper Honorable Mention
  Conjunctive Queries with Comparisons
  Qichen Wang (HKUST), Ke Yi (HKUST)

  This paper introduces new query processing algorithms for conjunctive queries with predicates in the form of comparisons that span multiple relations. Such queries have regained interest recently, due to their relevance in OLAP queries, spatiotemporal databases, and machine learning over relational data. In addition to achieving theoretical linear time, the algorithms have also been implemented in Spark, yielding order-of-magnitude speedups over SparkSQL and PostgreSQL on a variety of graph pattern and analytical queries.

• PODS Test-of-Time Award
  Mergeable Summaries
  Pankaj K. Agarwal (Duke), Graham Cormode (Warwick), Zengfeng Huang (Fudan; HKUST CSE graduate), Jeff M. Phillips (Utah), Zhewei Wei (Renmin; HKUST CSE graduate), and Ke Yi (HKUST)

  This paper, first published in PODS 2012 with an extended version appeared in ACM Transactions on Database Systems, initiated the study of mergeability of data summaries, including analysing the mergeability of the popular existing sketches, and developing new mergeable summaries for quantiles, heavy hitters, and geometric approximations. Quoting from the award committee, "this paper has generated significant impact since its appearance: mergeability is now an essential property of sketches, and stands as a core principle of the Apache Data Sketches project as well as other products in industry. The paper is also increasingly influential within academic research, in both the database and the algorithms communities."

Congratulations again to Prof. YI and his current/former students: Wei, Juanru, Qichen, Zhewei, and Zengfeng!

ACM SIGMOD/PODS is the premier conference on database systems and theory, organized annually by ACM.

For more details, please refer to the SIGMOD website and PODS website.