May 21, 2020

Aditya Parameswaran Awarded Best Paper at SIGMOD 2020

Assistant Professor Aditya Parameswaran has been awarded the Best Paper Award at the 2020 ACM SIGMOD/PODS Conference for his joint paper: “ShapeSearch: A Flexible and Efficient System for Shape-based Exploration of Trendlines.” This paper was one of two that received the top award out of over 144 accepted research papers and 450 submissions.

The annual ACM SIGMOD/PODS Conference is a leading international forum for database researchers, developers, and users to explore innovative ideas and results, as well as share techniques and experiences in all forms of data management. This year, the conference will be held in Portland, Oregon, from June 14 to June 19, 2020. The Special Interest Group on Management of Data (SIGMOD) is one of three Special Interest Groups from the Association for Computing Machinery (ACM), the organization that sponsors the annual conference. 

Parameswaran completed this research paper with several researchers from the University of Illinois, Urbana Champaign: Tarique Siddiqui, Paul Luh, Zesheng Wang and Karrie Karahalios. The paper proposes the implementation of ShapeSearch, a tool that mitigates issues with existing visual analytics tools, which they identify as limited flexibility, expressiveness, and scalability.

From the paper’s abstract, the authors pitch their team’s new tool: “We propose ShapeSearch, an efficient and flexible pattern-searching tool, that enables the search for desired patterns via multiple mechanisms: sketch, natural-language, and visual regular expressions.”

“ShapeSearch is a promising step towards accelerating the search for insights in data, while catering to the needs of expert and novice programmers alike.”

The paper first sets out to take a comprehensive look at the development of ShapeSearch, breaking down the team’s processes and the tool’s various functionalities. It shares an in-depth discussion of the creation of ShapeQuery, which was designed to help ShapeSearch express a wide variety of patterns with a minimal set of primitives and operators, as well as the team’s selected execution engine, which enables interactive pattern matching on a large collection of visualizations.

Then, Parameswaran and his team present an evaluation of the system. This includes a user study, a case study, and performance experiments that compare ShapeSearch against existing high-end trendline shape matching approaches. By doing so, Parameswaran and his co-authors hope to demonstrate the usability and scalability of ShapeSearch.

In the paper’s conclusion, the authors share a final thought: “ShapeSearch is a promising step towards accelerating the search for insights in data, while catering to the needs of expert and novice programmers alike.”

Last updated:

May 21, 2020