arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2403.07124stat.MEcs.SI
收藏

基于随机梯度下降的带吸引子动态网络模型推断

Stochastic gradient descent-based inference for dynamic network models with attractors

Hancong Pan, Xiaojing Zhu, Cantay Caliskan, Dino P. Christenson, Konstantinos Spiliopoulos, Dylan Walker, Eric D. Kolaczyk

AI总结:

针对带吸引子的协同演化潜空间网络模型依赖MCMC难以扩展、节点需全程存在的问题,提出基于SGD的参数估计与不确定性量化方法并扩展支持节点动态进出,精度损失小且可扩展至大规模网络,应用于美国国会议员X平台社交网络克服选择偏差并发现共和党内部排斥力变化。

摘要:

在带吸引子的协同演化潜空间网络(Coevolving Latent Space Networks with Attractors,CLSNA)模型中,潜空间中的节点代表社会参与者,边表示它们的动态交互。模型借鉴动力系统理论,在潜层加入吸引子以刻画节点间的吸引力和排斥力概念。然而,CLSNA依赖马尔可夫链蒙特卡洛(MCMC)估计,难以扩展规模,且要求节点在整个研究周期内都存在,限制了实际应用。我们通过三方面解决这些问题:(1)提出一种随机梯度下降(Stochastic gradient descent,SGD)参数估计方法;(2)开发了一种利用SGD进行不确定性量化的新方法;(3)扩展模型以允许节点随时间加入和退出。仿真结果表明,与MCMC相比,我们的扩展方案精度损失很小,但可扩展到规模大得多的网络。我们将该方法应用于社交媒体平台X上美国国会议员的纵向社交网络。考虑节点动态性克服了网络中的选择偏差,并揭示了共和党内独特且日益增强的排斥力。

英文摘要:

In Coevolving Latent Space Networks with Attractors (CLSNA) models, nodes in a latent space represent social actors, and edges indicate their dynamic interactions. Attractors are added at the latent level to capture the notion of attractive and repulsive forces between nodes, borrowing from dynamical systems theory. However, CLSNA reliance on MCMC estimation makes scaling difficult, and the requirement for nodes to be present throughout the study period limit practical applications. We address these issues by (i) introducing a Stochastic gradient descent (SGD) parameter estimation method, (ii) developing a novel approach for uncertainty quantification using SGD, and (iii) extending the model to allow nodes to join and leave over time. Simulation results show that our extensions result in little loss of accuracy compared to MCMC, but can scale to much larger networks. We apply our approach to the longitudinal social networks of members of US Congress on the social media platform X. Accounting for node dynamics overcomes selection bias in the network and uncovers uniquely and increasingly repulsive forces within the Republican Party.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑