Long-term Forecasting using Higher Order Tensor RNNs. Yu, R., Zheng, S., Anandkumar, A., & Yue, Y. arXiv:1711.00073 [cs], August, 2019. arXiv: 1711.00073
Long-term Forecasting using Higher Order Tensor RNNs [link]Paper  abstract   bibtex   
We present Higher-Order Tensor RNN (HOT-RNN), a novel family of neural sequence architectures for multivariate forecasting in environments with nonlinear dynamics. Long-term forecasting in such systems is highly challenging, since there exist long-term temporal dependencies, higher-order correlations and sensitivity to error propagation. Our proposed recurrent architecture addresses these issues by learning the nonlinear dynamics directly using higher-order moments and higher-order state transition functions. Furthermore, we decompose the higher-order structure using the tensor-train decomposition to reduce the number of parameters while preserving the model performance. We theoretically establish the approximation guarantees and the variance bound for HOT-RNN for general sequence inputs. We also demonstrate 5% \textasciitilde 12% improvements for long-term prediction over general RNN and LSTM architectures on a range of simulated environments with nonlinear dynamics, as well on real-world time series data.
@article{yu_long-term_2019,
	title = {Long-term {Forecasting} using {Higher} {Order} {Tensor} {RNNs}},
	url = {http://arxiv.org/abs/1711.00073},
	abstract = {We present Higher-Order Tensor RNN (HOT-RNN), a novel family of neural sequence architectures for multivariate forecasting in environments with nonlinear dynamics. Long-term forecasting in such systems is highly challenging, since there exist long-term temporal dependencies, higher-order correlations and sensitivity to error propagation. Our proposed recurrent architecture addresses these issues by learning the nonlinear dynamics directly using higher-order moments and higher-order state transition functions. Furthermore, we decompose the higher-order structure using the tensor-train decomposition to reduce the number of parameters while preserving the model performance. We theoretically establish the approximation guarantees and the variance bound for HOT-RNN for general sequence inputs. We also demonstrate 5\% {\textasciitilde} 12\% improvements for long-term prediction over general RNN and LSTM architectures on a range of simulated environments with nonlinear dynamics, as well on real-world time series data.},
	urldate = {2019-12-12},
	journal = {arXiv:1711.00073 [cs]},
	author = {Yu, Rose and Zheng, Stephan and Anandkumar, Anima and Yue, Yisong},
	month = aug,
	year = {2019},
	note = {arXiv: 1711.00073},
	keywords = {Computer Science - Machine Learning, forecasting, long-lead forescasting}
}

Downloads: 0