LDTC: Lifelong deep temporal clustering for multivariate time series
This addresses the challenge of lifelong temporal clustering for real-world applications, offering a novel approach to prevent accuracy degradation in sequential learning, though it appears incremental in combining existing techniques like autoencoders and rehearsal methods.
The paper tackles the problem of clustering multivariate time series in lifelong learning settings, where existing methods fail to handle dynamic data and sequential tasks without catastrophic forgetting. The proposed LDTC algorithm integrates dimensionality reduction and clustering in an end-to-end framework, achieving high-quality results as demonstrated on seven real-world datasets.
Clustering temporal and dynamically changing multivariate time series from real-world fields, called temporal clustering for short, has been a major challenge due to inherent complexities. Although several deep temporal clustering algorithms have demonstrated a strong advantage over traditional methods in terms of model learning and clustering results, the accuracy of the few algorithms are not satisfactory. None of the existing algorithms can continuously learn new tasks and deal with the dynamic data effectively and efficiently in the sequential tasks learning. To bridge the gap and tackle these issues, this paper proposes a novel algorithm \textbf{L}ifelong \textbf{D}eep \textbf{T}emporal \textbf{C}lustering (\textbf{LDTC}), which effectively integrates dimensionality reduction and temporal clustering into an end-to-end deep unsupervised learning framework. Using a specifically designed autoencoder and jointly optimizing for both the latent representation and clustering objective, the LDTC can achieve high-quality clustering results. Moreover, unlike any previous work, the LDTC is uniquely equipped with the fully dynamic model expansion and rehearsal-based techniques to effectively learn new tasks and to tackle the dynamic data in the sequential tasks learning without the catastrophic forgetting or degradation of the model accuracy. Experiments on seven real-world multivariate time series datasets show that the LDTC is a promising method for dealing with temporal clustering issues effectively and efficiently.