Democratizing Data Science through Interactive Curation of ML Pipelines
© 2019 Association for Computing Machinery. Statistical knowledge and domain expertise are key to extract actionable insights out of data, yet such skills rarely coexist together. In Machine Learning, high-quality results are only attainable via mindful data preprocessing, hyperparameter tuning and...
Main Authors: | , , , , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
Association for Computing Machinery (ACM)
2021
|
Online Access: | https://hdl.handle.net/1721.1/132275 |
_version_ | 1811069494062219264 |
---|---|
author | Shang, Zeyuan Zgraggen, Emanuel Buratti, Benedetto Kossmann, Ferdinand Eichmann, Philipp Chung, Yeounoh Binnig, Carsten Upfal, Eli Kraska, Tim |
author_facet | Shang, Zeyuan Zgraggen, Emanuel Buratti, Benedetto Kossmann, Ferdinand Eichmann, Philipp Chung, Yeounoh Binnig, Carsten Upfal, Eli Kraska, Tim |
author_sort | Shang, Zeyuan |
collection | MIT |
description | © 2019 Association for Computing Machinery. Statistical knowledge and domain expertise are key to extract actionable insights out of data, yet such skills rarely coexist together. In Machine Learning, high-quality results are only attainable via mindful data preprocessing, hyperparameter tuning and model selection. Domain experts are often overwhelmed by such complexity, de-facto inhibiting a wider adoption of ML techniques in other elds. Existing libraries that claim to solve this problem, still require well-trained practitioners. Those frameworks involve heavy data preparation steps and are often too slow for interactive feedback from the user, severely limiting the scope of such systems. In this paper we present Alpine Meadow, arst Interactive Automated Machine Learning tool. What makes our system unique is not only the focus on interactivity, but also the combined systemic and algorithmic design approach; on one hand we leverage ideas from query optimization, on the other we devise novel selection and pruning strategies combining cost-based Multi-Armed Bandits and Bayesian Optimization. We evaluate our system on over 300 datasets and compare against other AutoML tools, including the current NIPS winner, as well as expert solutions. Not only is Alpine Meadow able to signicantly outperform the other AutoML systems while - in contrast to the other systems - providing interactive latencies, but also outperforms in 80% of the cases expert solutions over data sets we have never seen before. |
first_indexed | 2024-09-23T08:11:20Z |
format | Article |
id | mit-1721.1/132275 |
institution | Massachusetts Institute of Technology |
language | English |
last_indexed | 2024-09-23T08:11:20Z |
publishDate | 2021 |
publisher | Association for Computing Machinery (ACM) |
record_format | dspace |
spelling | mit-1721.1/1322752021-09-21T04:07:11Z Democratizing Data Science through Interactive Curation of ML Pipelines Shang, Zeyuan Zgraggen, Emanuel Buratti, Benedetto Kossmann, Ferdinand Eichmann, Philipp Chung, Yeounoh Binnig, Carsten Upfal, Eli Kraska, Tim © 2019 Association for Computing Machinery. Statistical knowledge and domain expertise are key to extract actionable insights out of data, yet such skills rarely coexist together. In Machine Learning, high-quality results are only attainable via mindful data preprocessing, hyperparameter tuning and model selection. Domain experts are often overwhelmed by such complexity, de-facto inhibiting a wider adoption of ML techniques in other elds. Existing libraries that claim to solve this problem, still require well-trained practitioners. Those frameworks involve heavy data preparation steps and are often too slow for interactive feedback from the user, severely limiting the scope of such systems. In this paper we present Alpine Meadow, arst Interactive Automated Machine Learning tool. What makes our system unique is not only the focus on interactivity, but also the combined systemic and algorithmic design approach; on one hand we leverage ideas from query optimization, on the other we devise novel selection and pruning strategies combining cost-based Multi-Armed Bandits and Bayesian Optimization. We evaluate our system on over 300 datasets and compare against other AutoML tools, including the current NIPS winner, as well as expert solutions. Not only is Alpine Meadow able to signicantly outperform the other AutoML systems while - in contrast to the other systems - providing interactive latencies, but also outperforms in 80% of the cases expert solutions over data sets we have never seen before. 2021-09-20T18:21:37Z 2021-09-20T18:21:37Z 2021-01-11T15:14:35Z Article http://purl.org/eprint/type/ConferencePaper https://hdl.handle.net/1721.1/132275 en 10.1145/3299869.3319863 Proceedings of the ACM SIGMOD International Conference on Management of Data Creative Commons Attribution-Noncommercial-Share Alike http://creativecommons.org/licenses/by-nc-sa/4.0/ application/pdf Association for Computing Machinery (ACM) Other repository |
spellingShingle | Shang, Zeyuan Zgraggen, Emanuel Buratti, Benedetto Kossmann, Ferdinand Eichmann, Philipp Chung, Yeounoh Binnig, Carsten Upfal, Eli Kraska, Tim Democratizing Data Science through Interactive Curation of ML Pipelines |
title | Democratizing Data Science through Interactive Curation of ML Pipelines |
title_full | Democratizing Data Science through Interactive Curation of ML Pipelines |
title_fullStr | Democratizing Data Science through Interactive Curation of ML Pipelines |
title_full_unstemmed | Democratizing Data Science through Interactive Curation of ML Pipelines |
title_short | Democratizing Data Science through Interactive Curation of ML Pipelines |
title_sort | democratizing data science through interactive curation of ml pipelines |
url | https://hdl.handle.net/1721.1/132275 |
work_keys_str_mv | AT shangzeyuan democratizingdatasciencethroughinteractivecurationofmlpipelines AT zgraggenemanuel democratizingdatasciencethroughinteractivecurationofmlpipelines AT burattibenedetto democratizingdatasciencethroughinteractivecurationofmlpipelines AT kossmannferdinand democratizingdatasciencethroughinteractivecurationofmlpipelines AT eichmannphilipp democratizingdatasciencethroughinteractivecurationofmlpipelines AT chungyeounoh democratizingdatasciencethroughinteractivecurationofmlpipelines AT binnigcarsten democratizingdatasciencethroughinteractivecurationofmlpipelines AT upfaleli democratizingdatasciencethroughinteractivecurationofmlpipelines AT kraskatim democratizingdatasciencethroughinteractivecurationofmlpipelines |