Artificial Intelligence (AI) can now automate the algorithm selection, feature engineering, and hyperparameter tuning steps in a machine learning workflow. Commonly known as AutoML or AutoAI, these technologies aim to relieve data scientists from the tedious manual work. However, today's AutoAI systems often present only limited to no information about the process of how they select and generate model results. Thus, users often do not understand the process, neither do they trust the outputs. In this short paper, we provide a first user evaluation by 10 data scientists of an experimental system, AutoAIViz, that aims to visualize AutoAI's model generation process. We find that the proposed system helps users to complete the data science tasks, and increases their understanding, toward the goal of increasing trust in the AutoAI system.
Description
AutoAIViz | Proceedings of the 25th International Conference on Intelligent User Interfaces
%0 Conference Paper
%1 Weidele_2020
%A Weidele, Daniel Karl I.
%A Weisz, Justin D.
%A Oduor, Erick
%A Muller, Michael
%A Andres, Josh
%A Gray, Alexander
%A Wang, Dakuo
%B Proceedings of the 25th International Conference on Intelligent User Interfaces
%D 2020
%I ACM
%K explanation information-visualization iui2020
%P 308-312
%R 10.1145/3377325.3377538
%T AutoAIViz: opening the blackbox of automated artificial intelligence with conditional parallel coordinates
Share on
%U https://doi.org/10.1145%2F3377325.3377538
%X Artificial Intelligence (AI) can now automate the algorithm selection, feature engineering, and hyperparameter tuning steps in a machine learning workflow. Commonly known as AutoML or AutoAI, these technologies aim to relieve data scientists from the tedious manual work. However, today's AutoAI systems often present only limited to no information about the process of how they select and generate model results. Thus, users often do not understand the process, neither do they trust the outputs. In this short paper, we provide a first user evaluation by 10 data scientists of an experimental system, AutoAIViz, that aims to visualize AutoAI's model generation process. We find that the proposed system helps users to complete the data science tasks, and increases their understanding, toward the goal of increasing trust in the AutoAI system.
@inproceedings{Weidele_2020,
abstract = {Artificial Intelligence (AI) can now automate the algorithm selection, feature engineering, and hyperparameter tuning steps in a machine learning workflow. Commonly known as AutoML or AutoAI, these technologies aim to relieve data scientists from the tedious manual work. However, today's AutoAI systems often present only limited to no information about the process of how they select and generate model results. Thus, users often do not understand the process, neither do they trust the outputs. In this short paper, we provide a first user evaluation by 10 data scientists of an experimental system, AutoAIViz, that aims to visualize AutoAI's model generation process. We find that the proposed system helps users to complete the data science tasks, and increases their understanding, toward the goal of increasing trust in the AutoAI system.},
added-at = {2020-08-03T00:08:10.000+0200},
author = {Weidele, Daniel Karl I. and Weisz, Justin D. and Oduor, Erick and Muller, Michael and Andres, Josh and Gray, Alexander and Wang, Dakuo},
biburl = {https://www.bibsonomy.org/bibtex/24a37b46b2823f5454fe2a4bb6075e20d/brusilovsky},
booktitle = {Proceedings of the 25th International Conference on Intelligent User Interfaces},
description = {AutoAIViz | Proceedings of the 25th International Conference on Intelligent User Interfaces},
doi = {10.1145/3377325.3377538},
interhash = {a3bb1f6287f74db81774ebb7d4ba7966},
intrahash = {4a37b46b2823f5454fe2a4bb6075e20d},
keywords = {explanation information-visualization iui2020},
month = mar,
pages = {308-312},
publisher = {{ACM}},
timestamp = {2020-12-07T21:14:55.000+0100},
title = {{AutoAIViz}: opening the blackbox of automated artificial intelligence with conditional parallel coordinates
Share on},
url = {https://doi.org/10.1145%2F3377325.3377538},
year = 2020
}