Prediction using the ground truth sounds like an oxymoron in machine learning. However, such an unrealistic setting was used in hundreds, if not thousands of papers in the area of finding graph representations. To evaluate the multi-label problem of node classification by using the obtained representations, many works assume that the number of labels of each test instance is known in the prediction stage. In practice such ground truth information is rarely available, but we point out that such an inappropriate setting is now ubiquitous in this research area. We detailedly investigate why the situation occurs. Our analysis indicates that with unrealistic information, the performance is likely over-estimated. To see why suitable predictions were not used, we identify difficulties in applying some multi-label techniques. For the use in future studies, we propose simple and effective settings without using practically unknown information. Finally, we take this chance to compare major graph representation learning methods on multi-label node classification.

The popularization of broadband access and Internet enables Live p2p streaming services. Live p2p streaming not only enriches the approaches and media of broadcasting but also frees the restriction on content providers as any user is able to be a content provider. The high scalability of p2p architecture significantly eases the burden of streaming server and bandwidth. This paper proposes a hybrid live p2p streaming system. A mesh in tree structure, denoted as MeTree, is suggested. Metree considers both the physical topology and peers' contributions to build the overlay structure. According each peer's contribution and location, Metree groups peers into different groups to provide differentiated services and reduce delay. Simulation results demonstrate that Metree significantly improve users' QoE and system performance while end-to-end delay, server loading, and network bandwidth are reduced.