Tensor Clustering

Tensor clustering aims to discover underlying group structures within multi-dimensional data represented as tensors, leveraging the inherent high-order relationships between data points. Current research emphasizes developing efficient and scalable algorithms, often incorporating tensor decompositions like Tucker decomposition and novel architectures such as multi-scale entanglement renormalization ansatz (MERA) networks, to handle high-dimensional data and multiple data views. These advancements improve clustering accuracy and efficiency, particularly for large datasets with noise or anomalies, finding applications in diverse fields like passenger flow modeling and multi-tissue gene expression analysis. The resulting improved clustering techniques offer valuable insights and enhanced data analysis capabilities across various scientific domains.