Most Probable Densest Subgraphs A summary of the IEEE International Conference on Data Engineering (ICDE) 2023 research paper by Arkaprava Saha, Xiangyu Ke, Arijit Khan, and Cheng Long Background: The discovery of dense subgraphs has attracted extensive attention in the data management community [1], [2], [3], [4], [5].They may correspond to communities [6], filter bubbles and echo chambers [7], [8] in social networks, brain regions responding to stimuli [9] or related to diseases [10], and commercial value motifs in financial domains [11]. They also have wide applications in graph compression and visualization [12], [13], [14], indexing for reachability and distance queries [15], [16], and social piggybacking [17]. Densest subgraphs usually maximize some notion of density in a given graph, e.g., the edge density [1], defined as the ratio of the number of induced edges to the number of nodes in a subgraph. Although there are an exponential number of subgraphs, a densest subgraph