28 October 2019 in IEEE Access
Similarity search is a foundational task in time series data mining. Although there are many ways to measure the similarity of time series, a lot of evidence indicates that dynamic time warping (DTW) has the best robustness in many applications. Unfortunately, the expensive computational cost limits its application in large-scale databases. To speed up similarity search under DTW, we design a framework of two-stage similarity search for time series. In the first stage, we propose an improved lower bounding distance, which can be used to discard plenty of dissimilar series to get a set of candidate sequences. In the second stage, to efficiently get the retrieval result from the set of candidate sequences, we explore early abandoning strategy to avoid the full calculation of DTW. Extensive experiments are conducted on real-world data sets. The experimental results indicate that the proposed method can improve the retrieval efficiency of similarity search under DTW and guarantee no false dismissals.
6 August 2019 in IEEE Access
In this paper, a discriminative autoencoding framework is proposed for semi-supervised anomaly detection using reconstruction errors. The framework only consists of a generator and a discriminative encoder, and the output of the latter is a vector. In the training process, the framework is trained as a generative adversarial network based on quadratic potential divergence. An extra loss added in the objective function enforces the discriminative encoder to use the mean value of the output vector for discrimination, which also empowers it with encoding ability. In the testing process, the trained framework can be used as an autoencoder to reconstruct test samples, where the trained discriminative encoder works as an encoder, and samples with reconstruction errors above a predefined threshold are determined as anomalies. The properties of quadratic potential divergence ensure a simple training process with comparable performance, meanwhile the discriminative encoder with two functions makes full use of training resources and reduces required network structures. Comparisons on benchmark datasets also show the efficiency and superiorities of the proposed methods.
5 April 2021 in IEEE Access
The maximum information coefficient (MIC) is a novel and widely-using measure of association detection in large datasets. The most outstanding feature of MIC is that it has both generality and equability. However, MIC can only deal with two variables and cannot precisely estimate coupling associations of multiple variables. In this paper, we propose an extension of MIC to deal with multi-variable datasets, called the multi-variable maximum information coefficient (MMIC). Some inherited and novel properties of MMIC are proved, including generality, equability, monotonicity, and subadditivity. We design an algorithm based on greedy stepwise strategy and upper confidence bound (UCB) for an approximate calculation of MMIC. The tests of MMIC on generated datasets and examples on real datasets are carried out to detect known and novel associations.
1 March 2016 in International Journal of Distributed Sensor Networks
Airborne highly dynamic ad hoc UAV network has features of high node mobility, fast changing network topology, and complex application environment. The performance of traditional routing algorithms is so poor over aspects such as end to end delay, data packet delivery ratio, and routing overhead that they cannot provide efficient communication for multi-UAVs carrying out missions synergistically. A bionic optimization based stability and congestion aware routing algorithm—BSCAR algorithm—is proposed to solve these problems. This algorithm integrates biological behavior and dynamic source routing algorithm, which can sense the congestion level of routes and the stability of routes. Ant colony optimization algorithm and the mathematical model of Physarum's behavior exert effort in the process of route discovery and maintenance. The level of pheromone in routes is chosen as a standard to choose route and calculated by the mathematical model of Physarum's behavior. A new volatilization mechanism of pheromone is also introduced into the algorithm. Meanwhile, the algorithm can make adjustment to the variance of UAV formation to prevent the compromise of the network performance. The simulation results show that the BSCAR algorithm has superiority over traditional algorithms and it is dependable in battlefield environment.
16 November 2026 in ACM Transactions on Multimedia Computing, Communications, and Applications
Real-scanned raw point clouds face dual challenges: sparse non-uniform distribution and significant noise/outliers. While many upsampling methods address sparsity, they remain highly noise-sensitive. That is, they fail to solve both issues. To fill this gap, we propose Robust-APU, an end-to-end unified framework integrating arbitrary upsampling and denoising. Our approach synergistically balances the two tasks: first, it interpolates the sparse point cloud to lay the upsampling foundation, then uses a dedicated noise-resistant module for preliminary denoising. Next, we design a weight-adaptive upsampling module to extract upsampled features from the original input. Though these features contain noise, they encode rich geometric information. Finally, a position refinement module fuses features from the denoising module and the upsampling module to enhance upsampled point accuracy and eliminates residual noise, boosting both upsampling quality and noise robustness. Extensive experiments show that Robust-APU outperforms state-of-the-arts at various upsampling rates. Specifically, on the PU-GAN dataset under the Gaussian noise level of 2%, it achieves 34.6% and 39.2% improvement in P2F metric, proving its ability to deliver high-quality upsampling while suppressing noise effectively.