Loughborough University
1-s2.0-S0893608022001368-mmc1.pdf (1.61 MB)
Download file

Supplementary information files for Context meta-reinforcement learning via neuromodulation

Download (1.61 MB)
posted on 2023-06-28, 11:31 authored by Eseoghene Ben-IwhiwhuEseoghene Ben-Iwhiwhu, Jeffery Dick, Nicholas A Ketz, Praveen K Pilly, Andrea SoltoggioAndrea Soltoggio

Supplementary files for article Context meta-reinforcement learning via neuromodulation

Meta-reinforcement learning (meta-RL) algorithms enable agents to adapt quickly to tasks from few samples in dynamic environments. Such a feat is achieved through dynamic representations in an agent’s policy network (obtained via reasoning about task context, model parameter updates, or both). However, obtaining rich dynamic representations for fast adaptation beyond simple benchmark problems is challenging due to the burden placed on the policy network to accommodate different policies. This paper addresses the challenge by introducing neuromodulation as a modular component to augment a standard policy network that regulates neuronal activities in order to produce efficient dynamic representations for task adaptation. The proposed extension to the policy network is evaluated across multiple discrete and continuous control environments of increasing complexity. To prove the generality and benefits of the extension in meta-RL, the neuromodulated network was applied to two state-of-the-art meta-RL algorithms (CAVIA and PEARL). The result demonstrates that meta-RL augmented with neuromodulation produces significantly better result and richer dynamic representations in comparison to the baselines. 


United States Air Force Research Laboratory (AFRL) and Defense Advanced Research Projects Agency (DARPA) under Contract No. FA8750-18-C0103



  • Science


  • Computer Science

Usage metrics

    Computer Science


    No categories selected