In this article, we design algorithms to maintain approximate core values in dynamic hypergraphs. This notion has been well studied for normal graphs in both static and dynamic setting. We generalize the problem to hypergraphs when edges can be inserted or deleted by an adversary. We consider two dynamic scenarios. In the first case, there are only insertions; and in the second case, there can be both insertions and deletions. In either case, the update time is poly-logarithmic in the number of nodes, with the insertion-only case boasting a better approximation ratio. We also perform extensive experiments on large real-world datasets, which demonstrate the accuracy and efficiency of our algorithms.

中文翻译：

---

超图中的完全动态近似 k 核分解

在本文中，我们设计算法以在动态超图中保持近似的核心值。对于静态和动态设置中的正态图，这个概念已经得到了很好的研究。当对手可以插入或删除边时，我们将问题推广到超图。我们考虑两个动态场景。在第一种情况下，只有插入；在第二种情况下，可以同时存在插入和删除。在任何一种情况下，更新时间都是节点数量的多对数，仅插入的情况具有更好的近似率。我们还对大型现实世界数据集进行了广泛的实验，证明了我们算法的准确性和效率。