Corpus ID: 202719337

Deep Reinforcement Learning with Modulated Hebbian plus Q Network Architecture

@article{Ladosz2019DeepRL,
  title={Deep Reinforcement Learning with Modulated Hebbian plus Q Network Architecture},
  author={Pawel Ladosz and Eseoghene Ben-Iwhiwhu and Yang Hu and Nicholas Ketz and Soheil Kolouri and Jeffrey L. Krichmar and Praveen K. Pilly and Andrea Soltoggio},
  journal={ArXiv},
  year={2019},
  volume={abs/1909.09902}
}
  • Pawel Ladosz, Eseoghene Ben-Iwhiwhu, +5 authors Andrea Soltoggio
  • Published 2019
  • Mathematics, Computer Science
  • ArXiv
  • This paper introduces the modulated Hebbian plus Q network architecture (MOHQA) for solving challenging partially observable Markov decision processes (POMDPs) deep reinforcement learning problems with sparse rewards and confounding observations. The proposed architecture combines a deep Q-network (DQN), and a modulated Hebbian network with neural eligibility traces (MOHN). Bio-inspired neural traces are used to bridge temporal delays between actions and rewards. The purpose is to discover… CONTINUE READING

    References

    Publications referenced by this paper.
    SHOWING 1-10 OF 39 REFERENCES

    Deep Recurrent Q-Learning for Partially Observable MDPs

    VIEW 6 EXCERPTS
    HIGHLY INFLUENTIAL

    Episodic Memory Deep Q-Networks

    VIEW 1 EXCERPT

    Human-level control through deep reinforcement learning

    VIEW 6 EXCERPTS
    HIGHLY INFLUENTIAL

    Dueling Network Architectures for Deep Reinforcement Learning

    VIEW 1 EXCERPT

    Asynchronous Methods for Deep Reinforcement Learning

    VIEW 4 EXCERPTS
    HIGHLY INFLUENTIAL