Deep clustering by maximizing mutual information in variational auto-encoder

Xu, Chen; Dai, Yuanfei; Lin, Renjie; Wang, Shiping

doi:10.1016/j.knosys.2020.106260

Cited by 36 publications

(19 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Traditional unsupervised learning techniques such as Kmeans [12] cannot be easily applied in high-dimensional data due to a catastrophe of dimension [33]. Recent work on unsupervised deep learning is divided into the following notable aspects: generative learning, e.g., generating adversarial networks (GAN) [34], variational autoencoder (VAE) [35]) and contrastive learning such as simple framework for contrastive learning of visual representations (SimCLR) [36] and simple siamese (SimSiam) [37].…”

Section: Unsupervised Deep Learningmentioning

confidence: 99%

BisSiam: Bispectrum Siamese Network Based Contrastive Learning for UAV Anomaly Detection

Li¹,

Zhou²,

Cai³

et al. 2023

IEEE Trans. Knowl. Data Eng.

View full text Add to dashboard Cite

In recent years, a surging number of unmanned aerial vehicles (UAVs) are pervasively utilized in many areas. However, the increasing number of UAVs may cause privacy and security issues such as voyeurism and espionage. It is critical for individuals or organizations to manage their behaviors and proactively prevent the misbehaved invasion of unauthorized UAVs through effective anomaly detection. The UAV anomaly detection framework needs to cope with complex signals in the noisy-prone environments and to function with very limited labeled samples. This paper proposes BISSIAM, a novel framework that is capable of identifying UAV presence, types and operation modes. BISSIAM converts UAVs signals to bispectrum as the input and exploits a siamese network based contrastive learning model to learn the vector encoding. A sampling mechanism is proposed for optimizing the sample size involved in the model training whilst ensuring the model accuracy without compromising the training efficiency. Finally, we present a similarity-based fingerprint matching mechanism for detecting unseen UAVs without the need of retraining the whole model. Experiment results show that our approach outperforms other baselines and can reach 92.85% accuracy of UAV type detection in unsupervised learning scenarios. 91.4% accuracy can be achieved when BISSIAM is used for detecting the UAV type of the out-of-sample UAVs.

show abstract

Section: Unsupervised Deep Learningmentioning

confidence: 99%

BisSiam: Bispectrum Siamese Network Based Contrastive Learning for UAV Anomaly Detection

Li¹,

Zhou²,

Cai³

et al. 2023

IEEE Trans. Knowl. Data Eng.

View full text Add to dashboard Cite

show abstract

“…This necessitates dimensionality reduction methods, which the authors do as follows. First, Xu et al [28] develop a Deep Clustering via Variational Auto-Encoder (DC-VAE) of mutual information maximization. Deep clustering refers to the process of guiding clustering methods jointly with automatic learning representation from the high-semantic and high-dimensional data via deep neural networks.…”

Section: Pre-processing and Dimension Reductionmentioning

confidence: 99%

“…Most papers report using only one dataset (14 papers), but the amount varies from 1 to 45 datasets. The papers that perform experiments on more than three datasets are: Silva, Almeida, and Yamakami [24] (45 datasets), Liu et al [15] (12 datasets), Liu and Guo [36] and Tellez et al [18] (7 datasets), Zheng and Zheng [37] and Xu et al [28] (6 datasets), and Conover et al [21] (5 datasets), Shi and Lu [38] and Wei et al [25] (4 datasets). We also identify that eight papers analyze three datasets, and five papers analyze two datasets.…”

Section: Characteristics Of the Datasets Used In The Experimentsmentioning

confidence: 99%

See 1 more Smart Citation

On the use of Machine Learning and Deep Learning for Text Similarity and Categorization and its Application to Troubleshooting Automation

Couto¹,

Tomaz²,

Godoy³

et al. 2022

Proceedings of the Annual Hawaii International Conference on System Sciences

View full text Add to dashboard Cite

Troubleshooting is a labor-intensive task that includes repetitive solutions to similar problems. This task can be partially or fully automated using text-similarity matching to find previous solutions, lowering the workload of technicians. We develop a systematic literature review to identify the best approaches to solve the problem of troubleshooting automation and classify incidents effectively. We identify promising approaches and point in the direction of a comprehensive set of solutions that could be employed in solving the troubleshooting automation problem. 1 Available at https://archive.ics.uci.edu/ml/ datasets/reuters-21578+text+categorization+ collection. Accessed in May 2021. 2 Available at https://archive.ics.uci.edu/ml/ datasets/Twenty+Newsgroups. Accessed in May 2021.

show abstract

“…Machine learning methods have been used as the powerful tools for feature detection/extraction and trend estimation/forecasting in the distributed sensor network applications. Supervised machine learning methods, such as neural network (NN), 1–18 convolutional neural network (CNN), 19–35 and recurrent neural network (RNN), 36–47 can be applied to the prediction and classification, while unsupervised machine learning methods, such as restricted Boltzmann machine (RBM), 48 deep belief network (DBN), deep Boltzmann machine (DBM), 49,50 auto-encoder (AE), 51–56 and denoising auto-encoder (DAE), can be utilized for the data denoising and model generalization. Furthermore, reinforcement learning methods, including generative adversarial networks (GANs) 57–60 and deep Q-networks (DQNs), are widely used in tools for generative networks and discriminative networks to optimize the contesting process in a zero-sum...…”

mentioning

confidence: 99%