Multi-scale Long Short-Term Memory Research Articles

An automatic and integrated self-diagnosing system for the silting disease of drainage pipelines based on SSAE-TSNE and MS-LSTM

The regular detection and diagnosis mechanism for the silting disease of drainage pipelines (SDP) is critical for making dredging decisions and flood forecasting. Simultaneously, there is a complex coupling relationship between the pipeline siltation and the dynamic changes of the inlet and outlet flows. However, the traditional silting detection frameworks of drainage pipelines face difficulties in two aspects, accurate flow rate measurements and efficient and intelligent siltation diagnoses. In this paper, an automatic and integrated self-diagnosing system for the silting disease of drainage pipelines is proposed. First, a gradient penalty generative adversarial network (GPGAN), a target recognition algorithm called guided faster R-CNN (GF R-CNN) and an image segmentation method called parallel multiscale U-Net (PMSU-Net) are introduced to dynamically measure the flow of water in the drainage pipelines. Second, an automatic siltation detection and diagnosis algorithm of SDP is constructed based on a t-SNE stack sparse autoencoder (SSAE-TSNE) and multiscale long short-term memory (MS-LSTM). Then, a full-scale prototype verifies the feasibility of the system. This improved flow measurement algorithm achieves the highest accuracy of 90.32%, and an F1 score of 0.963 in comparison with the other algorithms, and the precision-recall curve was the closest to the top right corner. The error rate of the proposed self-diagnosing system is controlled within 8%, which is far better than the other algorithms. Finally, the system algorithms are embedded in an intelligent platform consisting of an unmanned pipeline measuring device (UPM), an unmanned aerial vehicle (UAV) and a server diagnostic centre. The practical application and test show that the accuracy, precision and response speed of the proposed integrated self-diagnosing system have obvious advantages in the function of siltation detection and diagnosis in drainage pipelines.

Multi-Scale LSTM Model for BGP Anomaly Classification

As a policy-based routing protocol, the primary purpose of Border Gateway Protocol (BGP) is to exchange routing reachability information to provide sufficient end-to-end Quality-of-Service (QoS). The constant increase of anomalous traffic of BGP affects the connectivity and reachability of routing information among different Autonomous Systems (ASs), which calls for building accurate alerting models to provide stable routing services in the Internet. The previous works classify anomalies without considering the characteristic of multiple time scales, which may lead to inaccurate classification. In this paper, we propose a novel Multi-Scale Long Short-Term Memory (MSLSTM) model to capture the anomalous behaviors from BGP traffic. In our model, a Discrete Wavelet Transform is used to obtain temporal information on multiple scales, and a hierarchical two-layer LSTM architecture is devised where the first layer learns the attentions of different time scales to generate an integrated historical representation, and the second layer captures the temporal dependency in the learned representation. To evaluate the feasibility in different alerting scenarios, we conduct comprehensive experiments based on several BGP data sets collected from real world applications. The results demonstrate that our model achieves a promising performance compared with the state-of-the-art approaches.