Yu Wang, Shiwan Zhao, Zhihu Wang, Ming Fan, Xicheng Zhang, Yubo Zhang, Zhengfan Wang, Heyuan Huang, Ting Liu. Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing. 2025.

As a popular technique in cloud computing, multi-tenancy (MT) can significantly ease software maintenance, and improve resource utilization. To make use of the MT technique, an application may need to be transformed to be MT-enabled. This process involves finding and processing a special kind of data entities named global isolation points (GIPs). Practically, finding all GIPs of an application is challenging. Traditional method involves manually browsing the application code, requiring a great deal of human effort. To solve this problem, we introduce a toolkit named Auto-MT to help find and process GIPs of an application. Auto-MT is able to find new GIPs based on their relations to known GIPs. To characterize the relation, a novel graph called value flow graph (VFG) is introduced, which models the value flows of data entities. It can also be used in other scenarios, like taint analysis. We have implemented Auto-MT as an Eclipse Plug-in, and applied it to transform Roller, a widely used Java application. Experimental results show that Auto-MT saves substantial human effort, and accelerates the process of transforming applications to be MT-enabled.

Social services are more and more popular with the development of web and mobile internet. Behind social services, graph model is the key data structure representing the relationship among social entities. Partitioning social graphs according to graph connectivity is an important technique for the competency of social services including parallel computing, scalability, and analytics. This paper proposes a novel diffusion based approach for partitioning social graphs. A brand-new Random Walks model with Constant Source and Sink nodes (RWCSS) is devised to analogize a dynamic balanced diffusion procedure on graphs. The stationary state of RWCSS provides the closeness measurement between nodes which can serve as an effective global estimation for partitioning. Through incorporating the RWCSS model into the Bubble Framework, an efficient graph partitioning algorithm (Bubble-RWCSS) is implemented in a way gradually improving the partitioning quality with global consideration in each iteration. The evaluation on public well-known social graphs demonstrates the promising performance of the proposed method.

This letter presents an automatic moving object detection method in image sequences captured from videos. While we focus on extracting moving objects throughout a frame sequence, our approach does not require any prior knowledge such as the background modeling nor the interaction by users such as empirical thresholds tuning. Based on the continuous symmetric difference of the adjacent frames, we get the full resolution saliency map of the current frame, which highlights the moving objects with higher saliency values and meanwhile inhibits the saliency of the background. Then, the maximum entropy sum method is utilized to adaptively calculate the threshold to determine the candidate areas and get the reasonable attention seeds. After that, the ground truth is obtained based on the modified fuzzy growing of the attention seeds. The proposed algorithm is effective, robust and the experimental results demonstrate that it is promising for moving object detection.