Multiple Timescale Gated Recurrent Unit (MTGRU)

This is the original implementation of MTGRU developed at Artificial Brain Research Lab., School of Electronics Engineering, Kyungpook National University, Daegu, South Korea

Figure: A Multiple Timescale Gated Recurrent Unit.

Equation

Requirements

• Tensorflow 0.9
• python 2.x

References

• Minsoo Kim, Dennis Singh Moirangthem, and MinhoLee. 2016. Towards Abstraction from Extraction: Multiple Timescale Gated Recurrent Unit for Summarization. In Proceedings of the 1st Workshop on Representation Learning for NLP, Association for Computational Linguistics. pages 70–77.
• Dennis Singh Moirangthem and Minho Lee. 2017. Temporal hierarchies in multilayer gated recurrent neural networks for language models. In Neural Networks (IJCNN), 2017 International Joint Conference on. IEEE, pages 2152–2157.
• Dennis Singh Moirangthem, Jegyung Son, and Minho Lee. 2017. Representing compositionality based on multiple timescales gated recurrent neural networks with adaptive temporal hierarchy for character-level language models. In Proceedings of the 2nd Workshop on Representation Learning for NLP. Association for Computational Linguistics, pages 131–138.