The following paper, featuring joint work between TU/e, Athena R.C. and EPFL, was presented last week at the EDBT’20 Conference:
Hamid Shahrivari, Matthaios Olma, Odysseas Papapetrou, Dimitrios Skoutas, Anastasia Ailamaki: A Parallel and Distributed Approach for Diversified Top-k Best Region Search. EDBT 2020: 265-276 [pdf]
The paper addresses the Best Region Search problem: given a set of points in a two-dimensional space, this problem aims at finding the best placement of an axis-aligned rectangle of user-defined width and height such that the value of a user-defined monotone objective function over its contents is maximized. The paper presents a multi-round, a single-round, and a hybrid algorithm for computing top-k non-overlapping regions in a parallel and distributed manner.
Due to the current COVID-19 pandemic, EDBT’20, originally planned to be held in Copenhagen, took place as a fully online event, becoming the first database conference to be held online. Adapting to this format, the paper was presented through a 10 min video, followed by a live Q&A session.