Abstract
Time series prediction can be generalized as a process that extracts useful information from historical records and then determines future values. Learning long-range dependencies that are embedded in time series is often an obstacle for most algorithms, whereas LSTM solutions, as a specific kind of scheme in deep learning, promise to effectively overcome the problem. In this article, we first give a brief introduction to the structure and forward propagation mechanism of LSTM. Then, aiming at reducing the considerable computing cost of LSTM, we put forward a RCLSTM model by introducing stochastic connectivity to conventional LSTM neurons. Therefore, RCLSTM exhibits a certain level of sparsity and leads to a decrease in computational complexity. In the field of telecommunication networks, the prediction of traffic and user mobility could directly benefit from this improvement as we leverage a realistic dataset to show that for RCLSTM, the prediction performance comparable to LSTM is available, whereas considerably less computing time is required. We strongly argue that RCLSTM is more competent than LSTM in latency-stringent or power-constrained application scenarios.
| Original language | English |
|---|---|
| Article number | 8663965 |
| Pages (from-to) | 114-119 |
| Number of pages | 6 |
| Journal | IEEE Communications Magazine |
| Volume | 57 |
| Issue number | 6 |
| DOIs | |
| Publication status | Published - 1 Jun 2019 |
| MoE publication type | A1 Journal article-refereed |
Fingerprint
Keywords
- logic gates
- computer architecture
- biological neural networks
- computational modeling
- deep learning
Cite this
}
Deep Learning with Long Short-Term Memory for Time Series Prediction. / Hua, Yuxiu; Zhao, Zhifeng; Li, Rongpeng; Chen, Xianfu; Liu, Zhiming; Zhang, Honggang.
In: IEEE Communications Magazine, Vol. 57, No. 6, 8663965, 01.06.2019, p. 114-119.Research output: Contribution to journal › Article › Scientific › peer-review
TY - JOUR
T1 - Deep Learning with Long Short-Term Memory for Time Series Prediction
AU - Hua, Yuxiu
AU - Zhao, Zhifeng
AU - Li, Rongpeng
AU - Chen, Xianfu
AU - Liu, Zhiming
AU - Zhang, Honggang
PY - 2019/6/1
Y1 - 2019/6/1
N2 - Time series prediction can be generalized as a process that extracts useful information from historical records and then determines future values. Learning long-range dependencies that are embedded in time series is often an obstacle for most algorithms, whereas LSTM solutions, as a specific kind of scheme in deep learning, promise to effectively overcome the problem. In this article, we first give a brief introduction to the structure and forward propagation mechanism of LSTM. Then, aiming at reducing the considerable computing cost of LSTM, we put forward a RCLSTM model by introducing stochastic connectivity to conventional LSTM neurons. Therefore, RCLSTM exhibits a certain level of sparsity and leads to a decrease in computational complexity. In the field of telecommunication networks, the prediction of traffic and user mobility could directly benefit from this improvement as we leverage a realistic dataset to show that for RCLSTM, the prediction performance comparable to LSTM is available, whereas considerably less computing time is required. We strongly argue that RCLSTM is more competent than LSTM in latency-stringent or power-constrained application scenarios.
AB - Time series prediction can be generalized as a process that extracts useful information from historical records and then determines future values. Learning long-range dependencies that are embedded in time series is often an obstacle for most algorithms, whereas LSTM solutions, as a specific kind of scheme in deep learning, promise to effectively overcome the problem. In this article, we first give a brief introduction to the structure and forward propagation mechanism of LSTM. Then, aiming at reducing the considerable computing cost of LSTM, we put forward a RCLSTM model by introducing stochastic connectivity to conventional LSTM neurons. Therefore, RCLSTM exhibits a certain level of sparsity and leads to a decrease in computational complexity. In the field of telecommunication networks, the prediction of traffic and user mobility could directly benefit from this improvement as we leverage a realistic dataset to show that for RCLSTM, the prediction performance comparable to LSTM is available, whereas considerably less computing time is required. We strongly argue that RCLSTM is more competent than LSTM in latency-stringent or power-constrained application scenarios.
KW - logic gates
KW - computer architecture
KW - biological neural networks
KW - computational modeling
KW - deep learning
UR - http://www.scopus.com/inward/record.url?scp=85062957095&partnerID=8YFLogxK
U2 - 10.1109/MCOM.2019.1800155
DO - 10.1109/MCOM.2019.1800155
M3 - Article
VL - 57
SP - 114
EP - 119
JO - IEEE Communications Magazine
JF - IEEE Communications Magazine
SN - 0163-6804
IS - 6
M1 - 8663965
ER -