Surrogate decision tree visualization interpreting and visualizing black-box classification models with surrogate decision tree

Federica Di Castro, Enrico Bertini

    Research output: Contribution to journalConference article

    Abstract

    With the growing interest towards the application of Machine Learning techniques to many application domains, the need for transparent and interpretable ML is getting stronger. Visualizations methods can help model developers understand and refine ML models by making the logic of a given model visible and interactive. In this paper we describe a visual analytics tool we developed to support developers and domain experts (with little to no expertise in ML) in understanding the logic of a ML model without having access to the internal structure of the model (i.e., a model-agnostic method). The method is based on the creation of a “surrogate” decision tree which simulates the behavior of the black-box model of interest and presents readable rules to the end-users. We evaluate the effectiveness of the method with a preliminary user study and analysis of the level of fidelity the surrogate decision tree can reach with respect to the original model.

    Original languageEnglish (US)
    JournalCEUR Workshop Proceedings
    Volume2327
    StatePublished - Jan 1 2019
    Event2019 Joint ACM IUI Workshops, ACMIUI-WS 2019 - Los Angeles, United States
    Duration: Mar 20 2019 → …

    Fingerprint

    Decision trees
    Visualization
    Learning systems

    Keywords

    • Classification
    • Decision tree
    • Dendrograms
    • Explanation
    • Interpretability
    • Machine learning
    • User interface
    • Visual analytic

    ASJC Scopus subject areas

    • Computer Science(all)

    Cite this

    Surrogate decision tree visualization interpreting and visualizing black-box classification models with surrogate decision tree. / Di Castro, Federica; Bertini, Enrico.

    In: CEUR Workshop Proceedings, Vol. 2327, 01.01.2019.

    Research output: Contribution to journalConference article

    @article{e1be14668b3f41d5a59ba1c7633d4802,
    title = "Surrogate decision tree visualization interpreting and visualizing black-box classification models with surrogate decision tree",
    abstract = "With the growing interest towards the application of Machine Learning techniques to many application domains, the need for transparent and interpretable ML is getting stronger. Visualizations methods can help model developers understand and refine ML models by making the logic of a given model visible and interactive. In this paper we describe a visual analytics tool we developed to support developers and domain experts (with little to no expertise in ML) in understanding the logic of a ML model without having access to the internal structure of the model (i.e., a model-agnostic method). The method is based on the creation of a “surrogate” decision tree which simulates the behavior of the black-box model of interest and presents readable rules to the end-users. We evaluate the effectiveness of the method with a preliminary user study and analysis of the level of fidelity the surrogate decision tree can reach with respect to the original model.",
    keywords = "Classification, Decision tree, Dendrograms, Explanation, Interpretability, Machine learning, User interface, Visual analytic",
    author = "{Di Castro}, Federica and Enrico Bertini",
    year = "2019",
    month = "1",
    day = "1",
    language = "English (US)",
    volume = "2327",
    journal = "CEUR Workshop Proceedings",
    issn = "1613-0073",
    publisher = "CEUR-WS",

    }

    TY - JOUR

    T1 - Surrogate decision tree visualization interpreting and visualizing black-box classification models with surrogate decision tree

    AU - Di Castro, Federica

    AU - Bertini, Enrico

    PY - 2019/1/1

    Y1 - 2019/1/1

    N2 - With the growing interest towards the application of Machine Learning techniques to many application domains, the need for transparent and interpretable ML is getting stronger. Visualizations methods can help model developers understand and refine ML models by making the logic of a given model visible and interactive. In this paper we describe a visual analytics tool we developed to support developers and domain experts (with little to no expertise in ML) in understanding the logic of a ML model without having access to the internal structure of the model (i.e., a model-agnostic method). The method is based on the creation of a “surrogate” decision tree which simulates the behavior of the black-box model of interest and presents readable rules to the end-users. We evaluate the effectiveness of the method with a preliminary user study and analysis of the level of fidelity the surrogate decision tree can reach with respect to the original model.

    AB - With the growing interest towards the application of Machine Learning techniques to many application domains, the need for transparent and interpretable ML is getting stronger. Visualizations methods can help model developers understand and refine ML models by making the logic of a given model visible and interactive. In this paper we describe a visual analytics tool we developed to support developers and domain experts (with little to no expertise in ML) in understanding the logic of a ML model without having access to the internal structure of the model (i.e., a model-agnostic method). The method is based on the creation of a “surrogate” decision tree which simulates the behavior of the black-box model of interest and presents readable rules to the end-users. We evaluate the effectiveness of the method with a preliminary user study and analysis of the level of fidelity the surrogate decision tree can reach with respect to the original model.

    KW - Classification

    KW - Decision tree

    KW - Dendrograms

    KW - Explanation

    KW - Interpretability

    KW - Machine learning

    KW - User interface

    KW - Visual analytic

    UR - http://www.scopus.com/inward/record.url?scp=85063203196&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=85063203196&partnerID=8YFLogxK

    M3 - Conference article

    AN - SCOPUS:85063203196

    VL - 2327

    JO - CEUR Workshop Proceedings

    JF - CEUR Workshop Proceedings

    SN - 1613-0073

    ER -