A Taxonomy of Property Measures to Unify Active Learning and Human-centered Approaches to Data Labeling
Authors:
Open Access
- 3 September 2021
- journal article
- research article
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Interactive Intelligent Systems
- Vol. 11 (3-4) , 1-42
Abstract
Strategies for selecting the next data instance to label, in service of generating labeled data for machine learning, have been considered separately in the machine learning literature on active learning and in the visual analytics literature on human-centered approaches. We propose a unified design space for instance selection strategies to support detailed and fine-grained analysis covering both of these perspectives. We identify a concise set of 15 properties, namely measureable characteristics of datasets or of machine learning models applied to them, that cover most of the strategies in these literatures. To quantify these properties, we introduce Property Measures (PM) as fine-grained building blocks that can be used to formalize instance selection strategies. In addition, we present a taxonomy of PMs to support the description, evaluation, and generation of PMs across four dimensions: machine learning (ML) Model Output , Instance Relations , Measure Functionality , and Measure Valence . We also create computational infrastructure to support qualitative visual data analysis: a visual analytics explainer for PMs built around an implementation of PMs using cascades of eight atomic functions. It supports eight analysis tasks, covering the analysis of datasets and ML models using visual comparison within and between PMs and groups of PMs, and over time during the interactive labeling process. We iteratively refined the PM taxonomy, the explainer, and the task abstraction in parallel with each other during a two-year formative process, and show evidence of their utility through a summative evaluation with the same infrastructure. This research builds a formal baseline for the better understanding of the commonalities and differences of instance selection strategies, which can serve as the stepping stone for the synthesis of novel strategies in future work.Keywords
Funding Information
- German Academic Exchange Service
- Austrian Research Promotion Agency (866880)
- Lower Austrian Research and Education Company (FTI17-014 and FTI18-005)
- National Sciences and Engineering Research Council of Canada (RGPIN-2014-06309)
- Deutsche Forschungsgemeinschaft (251654672–TRR 161)
This publication has 81 references indexed in Scilit:
- Inactive learning?ACM SIGKDD Explorations Newsletter, 2011
- Active learning in multimedia annotation and retrievalACM Transactions on Intelligent Systems and Technology, 2011
- The Perception of Correlation in ScatterplotsComputer Graphics Forum, 2010
- Selecting good views of high‐dimensional data using class consistencyComputer Graphics Forum, 2009
- Support-vector networksMachine Learning, 1995
- Silhouettes: A graphical aid to the interpretation and validation of cluster analysisJournal of Computational and Applied Mathematics, 1987
- Power and Centrality: A Family of MeasuresAmerican Journal of Sociology, 1987
- Hierarchical Grouping to Optimize an Objective FunctionJournal of the American Statistical Association, 1963
- On Information and SufficiencyThe Annals of Mathematical Statistics, 1951
- Table for Estimating the Goodness of Fit of Empirical DistributionsThe Annals of Mathematical Statistics, 1948