A query engine for probabilistic preferences

Uzi Cohen, Batya Kenig, Haoyue Ping, Benny Kimelfeld, Julia Stoyanovich

    Research output: Chapter in Book/Report/Conference proceedingConference contribution

    Abstract

    Models of uncertain preferences, such as Mallows, have been extensively studied due to their plethora of application domains. In a recent work, a conceptual and theoretical framework has been proposed for supporting uncertain preferences as first-class citizens in a relational database. The resulting database is probabilistic, and, consequently, query evaluation entails inference of marginal probabilities of query answers. In this paper, we embark on the challenge of a practical realization of this framework. We first describe an implementation of a query engine that supports querying probabilistic preferences alongside relational data. Our system accommodates preference distributions in the general form of the Repeated Insertion Model (RIM), which generalizes Mallows and other models. We then devise a novel inference algorithm for conjunctive queries over RIM, and show that it significantly outperforms the state of the art in terms of both asymptotic and empirical execution cost. We also develop performance optimizations that are based on sharing computation among different inference tasks in the workload. Finally, we conduct an extensive experimental evaluation and demonstrate that clear performance benefits can be realized by a query engine with built-in probabilistic inference, as compared to a stand-alone implementation with a black-box inference solver.

    Original languageEnglish (US)
    Title of host publicationSIGMOD 2018 - Proceedings of the 2018 International Conference on Management of Data
    EditorsGautam Das, Christopher Jermaine, Ahmed Eldawy, Philip Bernstein
    PublisherAssociation for Computing Machinery
    Pages1509-1524
    Number of pages16
    ISBN (Electronic)9781450317436
    DOIs
    StatePublished - May 27 2018
    Event44th ACM SIGMOD International Conference on Management of Data, SIGMOD 2018 - Houston, United States
    Duration: Jun 10 2018Jun 15 2018

    Other

    Other44th ACM SIGMOD International Conference on Management of Data, SIGMOD 2018
    CountryUnited States
    CityHouston
    Period6/10/186/15/18

    Fingerprint

    Engines
    Costs

    ASJC Scopus subject areas

    • Software
    • Information Systems

    Cite this

    Cohen, U., Kenig, B., Ping, H., Kimelfeld, B., & Stoyanovich, J. (2018). A query engine for probabilistic preferences. In G. Das, C. Jermaine, A. Eldawy, & P. Bernstein (Eds.), SIGMOD 2018 - Proceedings of the 2018 International Conference on Management of Data (pp. 1509-1524). Association for Computing Machinery. https://doi.org/10.1145/3183713.3196923

    A query engine for probabilistic preferences. / Cohen, Uzi; Kenig, Batya; Ping, Haoyue; Kimelfeld, Benny; Stoyanovich, Julia.

    SIGMOD 2018 - Proceedings of the 2018 International Conference on Management of Data. ed. / Gautam Das; Christopher Jermaine; Ahmed Eldawy; Philip Bernstein. Association for Computing Machinery, 2018. p. 1509-1524.

    Research output: Chapter in Book/Report/Conference proceedingConference contribution

    Cohen, U, Kenig, B, Ping, H, Kimelfeld, B & Stoyanovich, J 2018, A query engine for probabilistic preferences. in G Das, C Jermaine, A Eldawy & P Bernstein (eds), SIGMOD 2018 - Proceedings of the 2018 International Conference on Management of Data. Association for Computing Machinery, pp. 1509-1524, 44th ACM SIGMOD International Conference on Management of Data, SIGMOD 2018, Houston, United States, 6/10/18. https://doi.org/10.1145/3183713.3196923
    Cohen U, Kenig B, Ping H, Kimelfeld B, Stoyanovich J. A query engine for probabilistic preferences. In Das G, Jermaine C, Eldawy A, Bernstein P, editors, SIGMOD 2018 - Proceedings of the 2018 International Conference on Management of Data. Association for Computing Machinery. 2018. p. 1509-1524 https://doi.org/10.1145/3183713.3196923
    Cohen, Uzi ; Kenig, Batya ; Ping, Haoyue ; Kimelfeld, Benny ; Stoyanovich, Julia. / A query engine for probabilistic preferences. SIGMOD 2018 - Proceedings of the 2018 International Conference on Management of Data. editor / Gautam Das ; Christopher Jermaine ; Ahmed Eldawy ; Philip Bernstein. Association for Computing Machinery, 2018. pp. 1509-1524
    @inproceedings{1f5af5323b7041c483422076392398e4,
    title = "A query engine for probabilistic preferences",
    abstract = "Models of uncertain preferences, such as Mallows, have been extensively studied due to their plethora of application domains. In a recent work, a conceptual and theoretical framework has been proposed for supporting uncertain preferences as first-class citizens in a relational database. The resulting database is probabilistic, and, consequently, query evaluation entails inference of marginal probabilities of query answers. In this paper, we embark on the challenge of a practical realization of this framework. We first describe an implementation of a query engine that supports querying probabilistic preferences alongside relational data. Our system accommodates preference distributions in the general form of the Repeated Insertion Model (RIM), which generalizes Mallows and other models. We then devise a novel inference algorithm for conjunctive queries over RIM, and show that it significantly outperforms the state of the art in terms of both asymptotic and empirical execution cost. We also develop performance optimizations that are based on sharing computation among different inference tasks in the workload. Finally, we conduct an extensive experimental evaluation and demonstrate that clear performance benefits can be realized by a query engine with built-in probabilistic inference, as compared to a stand-alone implementation with a black-box inference solver.",
    author = "Uzi Cohen and Batya Kenig and Haoyue Ping and Benny Kimelfeld and Julia Stoyanovich",
    year = "2018",
    month = "5",
    day = "27",
    doi = "10.1145/3183713.3196923",
    language = "English (US)",
    pages = "1509--1524",
    editor = "Gautam Das and Christopher Jermaine and Ahmed Eldawy and Philip Bernstein",
    booktitle = "SIGMOD 2018 - Proceedings of the 2018 International Conference on Management of Data",
    publisher = "Association for Computing Machinery",

    }

    TY - GEN

    T1 - A query engine for probabilistic preferences

    AU - Cohen, Uzi

    AU - Kenig, Batya

    AU - Ping, Haoyue

    AU - Kimelfeld, Benny

    AU - Stoyanovich, Julia

    PY - 2018/5/27

    Y1 - 2018/5/27

    N2 - Models of uncertain preferences, such as Mallows, have been extensively studied due to their plethora of application domains. In a recent work, a conceptual and theoretical framework has been proposed for supporting uncertain preferences as first-class citizens in a relational database. The resulting database is probabilistic, and, consequently, query evaluation entails inference of marginal probabilities of query answers. In this paper, we embark on the challenge of a practical realization of this framework. We first describe an implementation of a query engine that supports querying probabilistic preferences alongside relational data. Our system accommodates preference distributions in the general form of the Repeated Insertion Model (RIM), which generalizes Mallows and other models. We then devise a novel inference algorithm for conjunctive queries over RIM, and show that it significantly outperforms the state of the art in terms of both asymptotic and empirical execution cost. We also develop performance optimizations that are based on sharing computation among different inference tasks in the workload. Finally, we conduct an extensive experimental evaluation and demonstrate that clear performance benefits can be realized by a query engine with built-in probabilistic inference, as compared to a stand-alone implementation with a black-box inference solver.

    AB - Models of uncertain preferences, such as Mallows, have been extensively studied due to their plethora of application domains. In a recent work, a conceptual and theoretical framework has been proposed for supporting uncertain preferences as first-class citizens in a relational database. The resulting database is probabilistic, and, consequently, query evaluation entails inference of marginal probabilities of query answers. In this paper, we embark on the challenge of a practical realization of this framework. We first describe an implementation of a query engine that supports querying probabilistic preferences alongside relational data. Our system accommodates preference distributions in the general form of the Repeated Insertion Model (RIM), which generalizes Mallows and other models. We then devise a novel inference algorithm for conjunctive queries over RIM, and show that it significantly outperforms the state of the art in terms of both asymptotic and empirical execution cost. We also develop performance optimizations that are based on sharing computation among different inference tasks in the workload. Finally, we conduct an extensive experimental evaluation and demonstrate that clear performance benefits can be realized by a query engine with built-in probabilistic inference, as compared to a stand-alone implementation with a black-box inference solver.

    UR - http://www.scopus.com/inward/record.url?scp=85048804055&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=85048804055&partnerID=8YFLogxK

    U2 - 10.1145/3183713.3196923

    DO - 10.1145/3183713.3196923

    M3 - Conference contribution

    SP - 1509

    EP - 1524

    BT - SIGMOD 2018 - Proceedings of the 2018 International Conference on Management of Data

    A2 - Das, Gautam

    A2 - Jermaine, Christopher

    A2 - Eldawy, Ahmed

    A2 - Bernstein, Philip

    PB - Association for Computing Machinery

    ER -