Data smells in Public DatasetsResearch Paper
The adoption of Artificial Intelligence (AI) in high-stakes domains such as healthcare, wildlife preservation, autonomous driving and criminal justice system calls for a data-centric approach to AI. Data scientists spend the majority of their time studying and wrangling the data, yet tools to aid them with data analysis are lacking. This study identifies the recurrent data quality issues in public datasets. Analogous to code smells, we introduce the notion of data smells that indicate early signs of problems or technical debt in machine learning systems. To the best of our knowledge, such a catalog of smells in the context of data science does not exist. To understand the prevalence of data quality issues in datasets, we analyse 25 public datasets and propose a catalog of 14 data smells.
Tue 17 MayDisplayed time zone: Eastern Time (US & Canada) change
09:30 - 11:00 | AI SmellsCAIN 2022 at CAIN main room Chair(s): Ipek Ozkaya Carnegie Mellon Software Engineering Institute, Thomas Zimmermann Microsoft Research | ||
09:30 30mOther | Activity: Brainwriting CAIN 2022 | ||
10:00 15mResearch paper | Code Smells for Machine Learning ApplicationsResearch Paper CAIN 2022 Haiyin Zhang AI for Fintech Research, ING, Luís Cruz Deflt University of Technology, Arie van Deursen Delft University of Technology, Netherlands Pre-print | ||
10:15 15mResearch paper | Data Smells: Categories, Causes and Consequences, and Detection of Suspicious Data in AI-based SystemsResearch Paper CAIN 2022 Harald Foidl University of Innsbruck, Michael Felderer University of Innsbruck, Rudolf Ramler Software Competence Center Hagenberg Pre-print | ||
10:30 15mResearch paper | Data smells in Public DatasetsResearch Paper CAIN 2022 Arumoy Shome Delft University of Technology, Luís Cruz Deflt University of Technology, Arie van Deursen Delft University of Technology, Netherlands Pre-print | ||
10:45 15mOther | Discussion on Smells in AI CAIN 2022 | ||