Local Clustering

Local clustering focuses on identifying groups of closely related data points within a larger dataset, without needing to analyze the entire dataset, thus improving efficiency. Current research emphasizes developing efficient algorithms, often based on graph representations and sparse solutions to linear systems, to perform this task effectively, particularly in the presence of noisy data or distributed data settings like federated learning. These advancements are crucial for handling large datasets in various applications, including image classification, social network analysis, and privacy-preserving data analysis, where processing the entire dataset is computationally prohibitive or violates privacy constraints. The development of robust and efficient local clustering methods is driving progress in these fields.

Papers