The document summarizes a research paper that proposes a new algorithm called the leader-follower algorithm for detecting community structure in networks. The algorithm is based on identifying the internal structure of communities rather than external connectivity properties used by traditional methods like spectral clustering. It first distinguishes between leader nodes that connect different communities and loyal follower nodes that only have neighbors within their community. It then assigns loyal followers to leaders to form the communities. The algorithm is able to detect communities of any size and learn their number without requiring this as input, offering an improvement over spectral clustering especially for densely connected networks. Experiments on synthetic and real-world networks demonstrate its effectiveness.