Predictive learning with uncertainty estimation for modeling infants' cognitive development with caregivers: A neurorobotics experiment

Shingo Murata, Saki Tomioka, Ryoichi Nakajo, Tatsuro Yamada, Hiroaki Arie, Tetsuya Ogata, Shigeki Sugano

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

    Abstract

    Dynamic interactions with caregivers are essential for infants to develop cognitive abilities, including aspects of action, perception, and attention. We hypothesized that these abilities can be acquired through the predictive learning of sensory inputs including their uncertainty (inverse precision) in terms of variance. To examine our hypothesis from the perspective of cognitive developmental robotics, we conducted a neurorobotics experiment involving a ball-playing interaction task between a human experimenter representing a caregiver and a small humanoid robot representing an infant. The robot was equipped with a dynamic generative model called a stochastic continuous-time recurrent neural network (S-CTRNN). The S-CTRNN learned to generate predictions about both the visuo-proprioceptive states of the robot and the uncertainty of these states by minimizing a negative log-likelihood consisting of log-uncertainty and precision-weighted prediction error. The experimental results showed that predictive learning with uncertainty estimation enabled the robot to acquire infant-like cognitive abilities through dynamic interactions with the experimenter. We also discuss the effects of infant-directed modifications observed in caregiver-infant interactions on the development of these abilities.

    Original languageEnglish
    Title of host publication5th Joint International Conference on Development and Learning and Epigenetic Robotics, ICDL-EpiRob 2015
    PublisherInstitute of Electrical and Electronics Engineers Inc.
    Pages302-307
    Number of pages6
    ISBN (Print)9781467393201
    DOIs
    Publication statusPublished - 2015 Dec 2
    Event5th Joint International Conference on Development and Learning and Epigenetic Robotics, ICDL-EpiRob 2015 - Providence, United States
    Duration: 2015 Aug 132015 Aug 16

    Other

    Other5th Joint International Conference on Development and Learning and Epigenetic Robotics, ICDL-EpiRob 2015
    CountryUnited States
    CityProvidence
    Period15/8/1315/8/16

    Fingerprint

    Robots
    Recurrent neural networks
    Experiments
    Dynamic models
    Robotics
    Uncertainty

    ASJC Scopus subject areas

    • Artificial Intelligence

    Cite this

    Murata, S., Tomioka, S., Nakajo, R., Yamada, T., Arie, H., Ogata, T., & Sugano, S. (2015). Predictive learning with uncertainty estimation for modeling infants' cognitive development with caregivers: A neurorobotics experiment. In 5th Joint International Conference on Development and Learning and Epigenetic Robotics, ICDL-EpiRob 2015 (pp. 302-307). [7346162] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/DEVLRN.2015.7346162

    Predictive learning with uncertainty estimation for modeling infants' cognitive development with caregivers : A neurorobotics experiment. / Murata, Shingo; Tomioka, Saki; Nakajo, Ryoichi; Yamada, Tatsuro; Arie, Hiroaki; Ogata, Tetsuya; Sugano, Shigeki.

    5th Joint International Conference on Development and Learning and Epigenetic Robotics, ICDL-EpiRob 2015. Institute of Electrical and Electronics Engineers Inc., 2015. p. 302-307 7346162.

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

    Murata, S, Tomioka, S, Nakajo, R, Yamada, T, Arie, H, Ogata, T & Sugano, S 2015, Predictive learning with uncertainty estimation for modeling infants' cognitive development with caregivers: A neurorobotics experiment. in 5th Joint International Conference on Development and Learning and Epigenetic Robotics, ICDL-EpiRob 2015., 7346162, Institute of Electrical and Electronics Engineers Inc., pp. 302-307, 5th Joint International Conference on Development and Learning and Epigenetic Robotics, ICDL-EpiRob 2015, Providence, United States, 15/8/13. https://doi.org/10.1109/DEVLRN.2015.7346162
    Murata S, Tomioka S, Nakajo R, Yamada T, Arie H, Ogata T et al. Predictive learning with uncertainty estimation for modeling infants' cognitive development with caregivers: A neurorobotics experiment. In 5th Joint International Conference on Development and Learning and Epigenetic Robotics, ICDL-EpiRob 2015. Institute of Electrical and Electronics Engineers Inc. 2015. p. 302-307. 7346162 https://doi.org/10.1109/DEVLRN.2015.7346162
    Murata, Shingo ; Tomioka, Saki ; Nakajo, Ryoichi ; Yamada, Tatsuro ; Arie, Hiroaki ; Ogata, Tetsuya ; Sugano, Shigeki. / Predictive learning with uncertainty estimation for modeling infants' cognitive development with caregivers : A neurorobotics experiment. 5th Joint International Conference on Development and Learning and Epigenetic Robotics, ICDL-EpiRob 2015. Institute of Electrical and Electronics Engineers Inc., 2015. pp. 302-307
    @inproceedings{6b1301ee92454205a84ae9ae64ddd6da,
    title = "Predictive learning with uncertainty estimation for modeling infants' cognitive development with caregivers: A neurorobotics experiment",
    abstract = "Dynamic interactions with caregivers are essential for infants to develop cognitive abilities, including aspects of action, perception, and attention. We hypothesized that these abilities can be acquired through the predictive learning of sensory inputs including their uncertainty (inverse precision) in terms of variance. To examine our hypothesis from the perspective of cognitive developmental robotics, we conducted a neurorobotics experiment involving a ball-playing interaction task between a human experimenter representing a caregiver and a small humanoid robot representing an infant. The robot was equipped with a dynamic generative model called a stochastic continuous-time recurrent neural network (S-CTRNN). The S-CTRNN learned to generate predictions about both the visuo-proprioceptive states of the robot and the uncertainty of these states by minimizing a negative log-likelihood consisting of log-uncertainty and precision-weighted prediction error. The experimental results showed that predictive learning with uncertainty estimation enabled the robot to acquire infant-like cognitive abilities through dynamic interactions with the experimenter. We also discuss the effects of infant-directed modifications observed in caregiver-infant interactions on the development of these abilities.",
    author = "Shingo Murata and Saki Tomioka and Ryoichi Nakajo and Tatsuro Yamada and Hiroaki Arie and Tetsuya Ogata and Shigeki Sugano",
    year = "2015",
    month = "12",
    day = "2",
    doi = "10.1109/DEVLRN.2015.7346162",
    language = "English",
    isbn = "9781467393201",
    pages = "302--307",
    booktitle = "5th Joint International Conference on Development and Learning and Epigenetic Robotics, ICDL-EpiRob 2015",
    publisher = "Institute of Electrical and Electronics Engineers Inc.",

    }

    TY - GEN

    T1 - Predictive learning with uncertainty estimation for modeling infants' cognitive development with caregivers

    T2 - A neurorobotics experiment

    AU - Murata, Shingo

    AU - Tomioka, Saki

    AU - Nakajo, Ryoichi

    AU - Yamada, Tatsuro

    AU - Arie, Hiroaki

    AU - Ogata, Tetsuya

    AU - Sugano, Shigeki

    PY - 2015/12/2

    Y1 - 2015/12/2

    N2 - Dynamic interactions with caregivers are essential for infants to develop cognitive abilities, including aspects of action, perception, and attention. We hypothesized that these abilities can be acquired through the predictive learning of sensory inputs including their uncertainty (inverse precision) in terms of variance. To examine our hypothesis from the perspective of cognitive developmental robotics, we conducted a neurorobotics experiment involving a ball-playing interaction task between a human experimenter representing a caregiver and a small humanoid robot representing an infant. The robot was equipped with a dynamic generative model called a stochastic continuous-time recurrent neural network (S-CTRNN). The S-CTRNN learned to generate predictions about both the visuo-proprioceptive states of the robot and the uncertainty of these states by minimizing a negative log-likelihood consisting of log-uncertainty and precision-weighted prediction error. The experimental results showed that predictive learning with uncertainty estimation enabled the robot to acquire infant-like cognitive abilities through dynamic interactions with the experimenter. We also discuss the effects of infant-directed modifications observed in caregiver-infant interactions on the development of these abilities.

    AB - Dynamic interactions with caregivers are essential for infants to develop cognitive abilities, including aspects of action, perception, and attention. We hypothesized that these abilities can be acquired through the predictive learning of sensory inputs including their uncertainty (inverse precision) in terms of variance. To examine our hypothesis from the perspective of cognitive developmental robotics, we conducted a neurorobotics experiment involving a ball-playing interaction task between a human experimenter representing a caregiver and a small humanoid robot representing an infant. The robot was equipped with a dynamic generative model called a stochastic continuous-time recurrent neural network (S-CTRNN). The S-CTRNN learned to generate predictions about both the visuo-proprioceptive states of the robot and the uncertainty of these states by minimizing a negative log-likelihood consisting of log-uncertainty and precision-weighted prediction error. The experimental results showed that predictive learning with uncertainty estimation enabled the robot to acquire infant-like cognitive abilities through dynamic interactions with the experimenter. We also discuss the effects of infant-directed modifications observed in caregiver-infant interactions on the development of these abilities.

    UR - http://www.scopus.com/inward/record.url?scp=84962142831&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=84962142831&partnerID=8YFLogxK

    U2 - 10.1109/DEVLRN.2015.7346162

    DO - 10.1109/DEVLRN.2015.7346162

    M3 - Conference contribution

    AN - SCOPUS:84962142831

    SN - 9781467393201

    SP - 302

    EP - 307

    BT - 5th Joint International Conference on Development and Learning and Epigenetic Robotics, ICDL-EpiRob 2015

    PB - Institute of Electrical and Electronics Engineers Inc.

    ER -