Abstract | ||
---|---|---|
We present Tensor-Train RNN (TT-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 tensor recurrent architecture addresses these issues by learning the nonlinear dynamics directly using higher order moments and high-order state transition functions. Furthermore, we decompose the higher-order structure using the tensor-train (TT) decomposition to reduce the number of parameters while preserving the model performance. We theoretically establish the approximation properties of Tensor-Train RNNs for general sequence inputs, and such guarantees are not available for usual RNNs. We also demonstrate significant long-term prediction improvements over general RNN and LSTM architectures on a range of simulated environments with nonlinear dynamics, as well on real-world climate and traffic data. |
Year | Venue | Field |
---|---|---|
2017 | arXiv: Learning | Time series,Higher order moments,Mathematical optimization,Propagation of uncertainty,Nonlinear system,Tensor,Computer science,Multivariate statistics,Artificial intelligence,Tensor train,Machine learning |
DocType | Volume | Citations |
Journal | abs/1711.00073 | 6 |
PageRank | References | Authors |
0.49 | 15 | 4 |
Name | Order | Citations | PageRank |
---|---|---|---|
Qi Yu | 1 | 188 | 12.87 |
Qi Yu | 2 | 188 | 12.87 |
Stephan Zheng | 3 | 14 | 2.63 |
Animashree Anandkumar | 4 | 1629 | 116.30 |