Learning what to ignore

Memetic climbing in topology and weight space

Julian Togelius, Faustino Gomez, Jürgen Schmidhuber

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

    Abstract

    We present the memetic climber, a simple search algorithm that learns topology and weights of neural networks on different time scales. When applied to the problem of learning control for a simulated racing task with carefully selected inputs to the neural network, the memetie climber outperforms a standard hill-climber. When inputs to the network are less carefully selected, the difference is drastic. We also present two variations of the memetie climber and discuss the generalization of the underlying principle to population-based neuroevolution algorithms.

    Original languageEnglish (US)
    Title of host publication2008 IEEE Congress on Evolutionary Computation, CEC 2008
    Pages3274-3281
    Number of pages8
    DOIs
    StatePublished - 2008
    Event2008 IEEE Congress on Evolutionary Computation, CEC 2008 - Hong Kong, China
    Duration: Jun 1 2008Jun 6 2008

    Other

    Other2008 IEEE Congress on Evolutionary Computation, CEC 2008
    CountryChina
    CityHong Kong
    Period6/1/086/6/08

    Fingerprint

    Topology
    Neuroevolution
    Neural Networks
    Neural networks
    Learning Control
    Search Algorithm
    Time Scales
    Learning
    Standards
    Generalization

    ASJC Scopus subject areas

    • Computational Theory and Mathematics
    • Theoretical Computer Science

    Cite this

    Togelius, J., Gomez, F., & Schmidhuber, J. (2008). Learning what to ignore: Memetic climbing in topology and weight space. In 2008 IEEE Congress on Evolutionary Computation, CEC 2008 (pp. 3274-3281). [4631241] https://doi.org/10.1109/CEC.2008.4631241

    Learning what to ignore : Memetic climbing in topology and weight space. / Togelius, Julian; Gomez, Faustino; Schmidhuber, Jürgen.

    2008 IEEE Congress on Evolutionary Computation, CEC 2008. 2008. p. 3274-3281 4631241.

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

    Togelius, J, Gomez, F & Schmidhuber, J 2008, Learning what to ignore: Memetic climbing in topology and weight space. in 2008 IEEE Congress on Evolutionary Computation, CEC 2008., 4631241, pp. 3274-3281, 2008 IEEE Congress on Evolutionary Computation, CEC 2008, Hong Kong, China, 6/1/08. https://doi.org/10.1109/CEC.2008.4631241
    Togelius J, Gomez F, Schmidhuber J. Learning what to ignore: Memetic climbing in topology and weight space. In 2008 IEEE Congress on Evolutionary Computation, CEC 2008. 2008. p. 3274-3281. 4631241 https://doi.org/10.1109/CEC.2008.4631241
    Togelius, Julian ; Gomez, Faustino ; Schmidhuber, Jürgen. / Learning what to ignore : Memetic climbing in topology and weight space. 2008 IEEE Congress on Evolutionary Computation, CEC 2008. 2008. pp. 3274-3281
    @inproceedings{028020bfe0e0415f9e912d96fcf72f87,
    title = "Learning what to ignore: Memetic climbing in topology and weight space",
    abstract = "We present the memetic climber, a simple search algorithm that learns topology and weights of neural networks on different time scales. When applied to the problem of learning control for a simulated racing task with carefully selected inputs to the neural network, the memetie climber outperforms a standard hill-climber. When inputs to the network are less carefully selected, the difference is drastic. We also present two variations of the memetie climber and discuss the generalization of the underlying principle to population-based neuroevolution algorithms.",
    author = "Julian Togelius and Faustino Gomez and J{\"u}rgen Schmidhuber",
    year = "2008",
    doi = "10.1109/CEC.2008.4631241",
    language = "English (US)",
    isbn = "9781424418237",
    pages = "3274--3281",
    booktitle = "2008 IEEE Congress on Evolutionary Computation, CEC 2008",

    }

    TY - GEN

    T1 - Learning what to ignore

    T2 - Memetic climbing in topology and weight space

    AU - Togelius, Julian

    AU - Gomez, Faustino

    AU - Schmidhuber, Jürgen

    PY - 2008

    Y1 - 2008

    N2 - We present the memetic climber, a simple search algorithm that learns topology and weights of neural networks on different time scales. When applied to the problem of learning control for a simulated racing task with carefully selected inputs to the neural network, the memetie climber outperforms a standard hill-climber. When inputs to the network are less carefully selected, the difference is drastic. We also present two variations of the memetie climber and discuss the generalization of the underlying principle to population-based neuroevolution algorithms.

    AB - We present the memetic climber, a simple search algorithm that learns topology and weights of neural networks on different time scales. When applied to the problem of learning control for a simulated racing task with carefully selected inputs to the neural network, the memetie climber outperforms a standard hill-climber. When inputs to the network are less carefully selected, the difference is drastic. We also present two variations of the memetie climber and discuss the generalization of the underlying principle to population-based neuroevolution algorithms.

    UR - http://www.scopus.com/inward/record.url?scp=55749093699&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=55749093699&partnerID=8YFLogxK

    U2 - 10.1109/CEC.2008.4631241

    DO - 10.1109/CEC.2008.4631241

    M3 - Conference contribution

    SN - 9781424418237

    SP - 3274

    EP - 3281

    BT - 2008 IEEE Congress on Evolutionary Computation, CEC 2008

    ER -