We provide several strict convergence in measure theorems for the seminormed and semiconormed fuzzy integrals. These results are finer in comparison with the strict convergence in measure theorem which has been previously studied.

Thanks to the advancing mobile location services, people nowadays can post about places to share visiting experience on-the-go. A large place graph not only helps users explore interesting destinations, but also provides opportunities for understanding and modeling the real world. To improve coverage and flexibility of the place graph, many platforms import places data from multiple sources, which unfortunately leads to the emergence of numerous duplicated places that severely hinder subsequent location-related services. In this work, we take the anonymous place graph from Facebook as an example to systematically study the problem of place deduplication: We carefully formulate the problem, study its connections to various related tasks that lead to several promising basic models, and arrive at a systematic two-step data-driven pipeline based on place embedding with multiple novel techniques that works significantly better than the state-of-the-art.