HELIX: Accelerating human-in-the-loop machine learning

Doris Xin, Litian Ma, Jialin Liu, Stephen Macke, Shuchen Song, Aditya Parameswaran

Research output: Contribution to journalConference article

Abstract

Data application developers and data scientists spend an inordinate amount of time iterating on machine learning (ML) workflows- by modifying the data pre-processing, model training, and postprocessing steps-via trial-and-error to achieve the desired model performance. Existing work on accelerating machine learning focuses on speeding up one-shot execution of workflows, failing to address the incremental and dynamic nature of typical ML development. We propose HELIX, a declarative machine learning system that accelerates iterative development by optimizing workflow execution end-to-end and across iterations. HELIX minimizes the runtime per iteration via program analysis and intelligent reuse of previous results, which are selectively materialized-trading off the cost of materialization for potential future benefits-to speed up future iterations. Additionally, HELIX offers a graphical interface to visualize workflow DAGs and compare versions to facilitate iterative development. Through two ML applications, in classification and in structured prediction, attendees will experience the succinctness of HELIX's programming interface and the speed and ease of iterative development using HELIX. In our evaluations, HELIX achieved up to an order of magnitude reduction in cumulative run time compared to state-of-the-art machine learning tools.

Original languageEnglish (US)
Pages (from-to)1958-1961
Number of pages4
JournalProceedings of the VLDB Endowment
Volume11
Issue number12
DOIs
StatePublished - Jan 1 2018
Event44th International Conference on Very Large Data Bases, VLDB 2018 - Rio de Janeiro, Brazil
Duration: Aug 27 2017Aug 31 2017

Fingerprint

Learning systems
Processing
Costs

ASJC Scopus subject areas

  • Computer Science (miscellaneous)
  • Computer Science(all)

Cite this

Xin, D., Ma, L., Liu, J., Macke, S., Song, S., & Parameswaran, A. (2018). HELIX: Accelerating human-in-the-loop machine learning. Proceedings of the VLDB Endowment, 11(12), 1958-1961. https://doi.org/10.14778/3229863.3236234

HELIX : Accelerating human-in-the-loop machine learning. / Xin, Doris; Ma, Litian; Liu, Jialin; Macke, Stephen; Song, Shuchen; Parameswaran, Aditya.

In: Proceedings of the VLDB Endowment, Vol. 11, No. 12, 01.01.2018, p. 1958-1961.

Research output: Contribution to journalConference article

Xin, D, Ma, L, Liu, J, Macke, S, Song, S & Parameswaran, A 2018, 'HELIX: Accelerating human-in-the-loop machine learning', Proceedings of the VLDB Endowment, vol. 11, no. 12, pp. 1958-1961. https://doi.org/10.14778/3229863.3236234
Xin D, Ma L, Liu J, Macke S, Song S, Parameswaran A. HELIX: Accelerating human-in-the-loop machine learning. Proceedings of the VLDB Endowment. 2018 Jan 1;11(12):1958-1961. https://doi.org/10.14778/3229863.3236234
Xin, Doris ; Ma, Litian ; Liu, Jialin ; Macke, Stephen ; Song, Shuchen ; Parameswaran, Aditya. / HELIX : Accelerating human-in-the-loop machine learning. In: Proceedings of the VLDB Endowment. 2018 ; Vol. 11, No. 12. pp. 1958-1961.
@article{28b10a5cea374428b274807322d1638d,
title = "HELIX: Accelerating human-in-the-loop machine learning",
abstract = "Data application developers and data scientists spend an inordinate amount of time iterating on machine learning (ML) workflows- by modifying the data pre-processing, model training, and postprocessing steps-via trial-and-error to achieve the desired model performance. Existing work on accelerating machine learning focuses on speeding up one-shot execution of workflows, failing to address the incremental and dynamic nature of typical ML development. We propose HELIX, a declarative machine learning system that accelerates iterative development by optimizing workflow execution end-to-end and across iterations. HELIX minimizes the runtime per iteration via program analysis and intelligent reuse of previous results, which are selectively materialized-trading off the cost of materialization for potential future benefits-to speed up future iterations. Additionally, HELIX offers a graphical interface to visualize workflow DAGs and compare versions to facilitate iterative development. Through two ML applications, in classification and in structured prediction, attendees will experience the succinctness of HELIX's programming interface and the speed and ease of iterative development using HELIX. In our evaluations, HELIX achieved up to an order of magnitude reduction in cumulative run time compared to state-of-the-art machine learning tools.",
author = "Doris Xin and Litian Ma and Jialin Liu and Stephen Macke and Shuchen Song and Aditya Parameswaran",
year = "2018",
month = "1",
day = "1",
doi = "10.14778/3229863.3236234",
language = "English (US)",
volume = "11",
pages = "1958--1961",
journal = "Proceedings of the VLDB Endowment",
issn = "2150-8097",
publisher = "Very Large Data Base Endowment Inc.",
number = "12",

}

TY - JOUR

T1 - HELIX

T2 - Accelerating human-in-the-loop machine learning

AU - Xin, Doris

AU - Ma, Litian

AU - Liu, Jialin

AU - Macke, Stephen

AU - Song, Shuchen

AU - Parameswaran, Aditya

PY - 2018/1/1

Y1 - 2018/1/1

N2 - Data application developers and data scientists spend an inordinate amount of time iterating on machine learning (ML) workflows- by modifying the data pre-processing, model training, and postprocessing steps-via trial-and-error to achieve the desired model performance. Existing work on accelerating machine learning focuses on speeding up one-shot execution of workflows, failing to address the incremental and dynamic nature of typical ML development. We propose HELIX, a declarative machine learning system that accelerates iterative development by optimizing workflow execution end-to-end and across iterations. HELIX minimizes the runtime per iteration via program analysis and intelligent reuse of previous results, which are selectively materialized-trading off the cost of materialization for potential future benefits-to speed up future iterations. Additionally, HELIX offers a graphical interface to visualize workflow DAGs and compare versions to facilitate iterative development. Through two ML applications, in classification and in structured prediction, attendees will experience the succinctness of HELIX's programming interface and the speed and ease of iterative development using HELIX. In our evaluations, HELIX achieved up to an order of magnitude reduction in cumulative run time compared to state-of-the-art machine learning tools.

AB - Data application developers and data scientists spend an inordinate amount of time iterating on machine learning (ML) workflows- by modifying the data pre-processing, model training, and postprocessing steps-via trial-and-error to achieve the desired model performance. Existing work on accelerating machine learning focuses on speeding up one-shot execution of workflows, failing to address the incremental and dynamic nature of typical ML development. We propose HELIX, a declarative machine learning system that accelerates iterative development by optimizing workflow execution end-to-end and across iterations. HELIX minimizes the runtime per iteration via program analysis and intelligent reuse of previous results, which are selectively materialized-trading off the cost of materialization for potential future benefits-to speed up future iterations. Additionally, HELIX offers a graphical interface to visualize workflow DAGs and compare versions to facilitate iterative development. Through two ML applications, in classification and in structured prediction, attendees will experience the succinctness of HELIX's programming interface and the speed and ease of iterative development using HELIX. In our evaluations, HELIX achieved up to an order of magnitude reduction in cumulative run time compared to state-of-the-art machine learning tools.

UR - http://www.scopus.com/inward/record.url?scp=85058887157&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85058887157&partnerID=8YFLogxK

U2 - 10.14778/3229863.3236234

DO - 10.14778/3229863.3236234

M3 - Conference article

AN - SCOPUS:85058887157

VL - 11

SP - 1958

EP - 1961

JO - Proceedings of the VLDB Endowment

JF - Proceedings of the VLDB Endowment

SN - 2150-8097

IS - 12

ER -