Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Meta-learning in reinforcement learning.

Nicolas Schweighofer1, Kenji Doya

  • 1CREST, Japan Science and Technology Corporation, ATR, Human Information Science Laboratories, 2-2-2 Hikaridai, Seika-cho, Soraku-gun, 619-0288, Kyoto, Japan. nicolas@atr.co.jp

Neural Networks : the Official Journal of the International Neural Network Society
|February 11, 2003
PubMed
Summary
This summary is machine-generated.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

A composite measure of cerebral small vessel disease predicts cognitive change after stroke.

medRxiv : the preprint server for health sciences·2026
Same author

CALM-VLM: CALIBRATION AND SELECTIVE PREDICTION IN VISION-LANGUAGE MODELS FOR RELIABLE BRAIN MRI CLASSIFICATION.

bioRxiv : the preprint server for biology·2026
Same author

Self-Organizing Recruitment of Compensatory Cortical Areas Post-Stroke Can Maximize Residual Motor Performance.

IEEE transactions on neural systems and rehabilitation engineering : a publication of the IEEE Engineering in Medicine and Biology Society·2026
Same author

Associations between contralesional neuroplasticity and motor impairment through deep learning-derived MRI regional brain age in chronic stroke (ENIGMA): a multicohort, retrospective, observational study.

The Lancet. Digital health·2026
Same author

A Streamlined 4-item Wolf Motor Function Test for Efficient Assessment of Upper Extremity Motor Function in Chronic Stroke Survivors.

Neurorehabilitation and neural repair·2026
Same author

Towards AI-based precision rehabilitation via contextual model-based reinforcement learning.

Journal of neuroengineering and rehabilitation·2025
Same journal

Exploiting audio-visual modalities in videos: Object detection via multi-stage bilateral coupling network.

Neural networks : the official journal of the International Neural Network Society·2026
Same journal

Reliability-aware modality completion with cross-modal distillation for federated learning with missing modalities.

Neural networks : the official journal of the International Neural Network Society·2026
Same journal

IGFD-Net: Illumination-guided frequency decoupling for polarization image fusion.

Neural networks : the official journal of the International Neural Network Society·2026
Same journal

Multiple-Strategies dung beetle optimizer and its applications in engineering optimization and bankruptcy prediction.

Neural networks : the official journal of the International Neural Network Society·2026
Same journal

Aggregating global-scale pixel-wise forgery cues within a graph.

Neural networks : the official journal of the International Neural Network Society·2026
Same journal

Finite-Time intermittent control for secure synchronization of Neutral-Type stochastic delayed neural networks under aperiodic DoS attacks.

Neural networks : the official journal of the International Neural Network Society·2026
See all related articles

This study introduces a biologically plausible meta-reinforcement learning algorithm to adaptively tune reinforcement learning meta-parameters. The algorithm successfully optimizes parameters in dynamic environments, suggesting dopamine neuron firing encodes meta-learning signals.

Area of Science:

  • Computational neuroscience
  • Machine learning
  • Behavioral science

Background:

  • Reinforcement learning (RL) meta-parameters require tuning to match environmental dynamics and agent performance.
  • Adaptive tuning is crucial for optimizing RL agents in complex, changing environments.

Purpose of the Study:

  • To propose and evaluate a biologically plausible meta-reinforcement learning algorithm for adaptive meta-parameter tuning.
  • To investigate the algorithm's robustness in both simulated and real-world control tasks.
  • To explore the role of dopamine neuron firing in encoding meta-learning signals.

Main Methods:

  • Development of a novel meta-reinforcement learning algorithm inspired by biological plausibility.
  • Testing the algorithm in a simulated Markov decision task.

Related Experiment Videos

  • Validation of the algorithm in a non-linear control task.

    • Main Results:

    • The algorithm robustly identified appropriate meta-parameter values across different environments.
    • The algorithm effectively controlled the meta-parameter time course in both static and dynamic settings.
    • Demonstrated successful adaptation to environmental changes and task demands.

    Conclusions:

    • The proposed meta-reinforcement learning algorithm offers a robust and adaptive approach to tuning RL parameters.
    • The findings suggest that dopamine neuron activity, specifically phasic and tonic firing, may serve as a neural substrate for meta-learning in reinforcement learning.
    • This work bridges computational approaches to reinforcement learning with neurobiological mechanisms.