The rapid growth of the Internet of Things (IoT) has intensified the demand for collaborative and decentralized learning frameworks. Since these devices high-dimensional data, centralized transmission is often impractical, while purely local learning suffers from slow convergence. Collaborative approaches can alleviate these issues by allowing agents to use information from one another to improve estimation. Each agent typically faces a personalized learning problem, and collaboration is only beneficial among agents whose data are generated from the same distributions. This thesis studies the problem of personalized online mean estimation in heterogeneous environments. To address the challenge of heterogeneity, collaborative algorithms are introduced that enable agents to identify similarity classes in real time and exploit information from agents belonging to the similarity class within the same class to accelerate convergence and improve accuracy.