In this paper, we present a visual analytics tool for enabling hypothesis-based evaluation of machine learning (ML) models. We describe a novel ML-testing framework that combines the traditional statistical hypothesis testing (commonly used in empirical research) with logical reasoning about the conclusions of multiple hypotheses. The framework defines a controlled configuration for testing a number of hypotheses as to whether and how some extra information about a "concept" or "feature" may benefit or hinder a ML model. Because reasoning multiple hypotheses is not always straightforward, we provide HypoML as a visual analysis tool, with which, the multi-thread testing data is transformed to a visual representation for rapid observation of the conclusions and the logical flow between the testing data and hypotheses.We have applied HypoML to a number of hypothesized concepts, demonstrating the intuitive and explainable nature of the visual analysis.
@article{arxiv.2002.05271,
title = {HypoML: Visual Analysis for Hypothesis-based Evaluation of Machine Learning Models},
author = {Qianwen Wang and William Alexander and Jack Pegg and Huamin Qu and Min Chen},
journal= {arXiv preprint arXiv:2002.05271},
year = {2020}
}
Comments
This article was submitted to EuroVis 2020 on 5 December 2020. It was not accepted. Because the reviews have not identified any technical problems that would undermine the novelty and validity of this work, we think that the article is ready to be released as an arXiv report. The EuroVis 2020 reviews and authors' short feedback can be found in the anc folder