Hybrid computing using a neural network with dynamic external memory

@article{Graves2016HybridCU,
  title={Hybrid computing using a neural network with dynamic external memory},
  author={A. Graves and Greg Wayne and M. Reynolds and T. Harley and Ivo Danihelka and Agnieszka Grabska-Barwinska and Sergio Gomez Colmenarejo and Edward Grefenstette and Tiago Ramalho and J. Agapiou and Adri{\`a} Puigdom{\`e}nech Badia and K. Hermann and Yori Zwols and Georg Ostrovski and A. Cain and H. King and C. Summerfield and P. Blunsom and K. Kavukcuoglu and Demis Hassabis},
  journal={Nature},
  year={2016},
  volume={538},
  pages={471-476}
}
  • A. Graves, Greg Wayne, +17 authors Demis Hassabis
  • Published 2016
  • Computer Science, Medicine
  • Nature
  • Artificial neural networks are remarkably adept at sensory processing, sequence learning and reinforcement learning, but are limited in their ability to represent variables and data structures and to store data over long timescales, owing to the lack of an external memory. [...] Key Result Taken together, our results demonstrate that DNCs have the capacity to solve complex, structured tasks that are inaccessible to neural networks without external read–write memory.Expand Abstract
    Artificial intelligence: Deep neural reasoning
    • 27
    • PDF
    Neural Map: Structured Memory for Deep Reinforcement Learning
    • 130
    • PDF
    Relational recurrent neural networks
    • 85
    • Highly Influenced
    • PDF
    Neural Stored-program Memory
    • 7
    • Highly Influenced
    • PDF
    Memory Augmented Neural Networks with Wormhole Connections
    • 44
    • Highly Influenced
    • PDF
    Scaling Memory-Augmented Neural Networks with Sparse Reads and Writes
    • 96
    • PDF
    Learning to reinforcement learn
    • 397
    • PDF
    A Theory of Sequence Indexing and Working Memory in Recurrent Neural Networks
    • 30
    • PDF

    References

    Publications referenced by this paper.
    SHOWING 1-10 OF 42 REFERENCES
    Sparse Distributed Memory
    • 929
    Human-level control through deep reinforcement learning
    • 9,842
    • PDF
    Long Short-Term Memory
    • 31,084
    • PDF
    Pointer Networks
    • 1,011
    • PDF
    Memory Networks
    • 912
    • PDF
    End-To-End Memory Networks
    • 1,497
    • PDF