Summary of Inferring Dynamic Networks From Marginals with Iterative Proportional Fitting, by Serina Chang et al.
Inferring Dynamic Networks from Marginals with Iterative Proportional Fitting
by Serina Chang, Frederic Koehler, Zhaonan Qu, Jure Leskovec, Johan Ugander
First submitted to arxiv on: 28 Feb 2024
Categories
- Main: Machine Learning (stat.ML)
- Secondary: Machine Learning (cs.LG); Social and Information Networks (cs.SI); Optimization and Control (math.OC); Statistics Theory (math.ST)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The proposed paper addresses the challenge of inferring dynamic networks from time-aggregated adjacency matrices and time-varying marginals. The authors build upon prior work that utilizes the iterative proportional fitting (IPF) procedure to estimate dynamic networks, but lack a statistical foundation for its use. The researchers establish a generative network model whose maximum likelihood estimates are recovered by IPF, revealing implicit assumptions on its usage and enabling new analyses such as structure-dependent error bounds. They also introduce an algorithm that guarantees IPF convergence under minimal changes to the network structure when it fails to converge on sparse data. Experiments with synthetic and real-world data demonstrate the practical value of the theoretical and algorithmic contributions. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary A group of scientists is trying to figure out how to create a picture of a network that changes over time, based on some limited information about the network’s edges and nodes. They’re using an old technique called IPF (or Sinkhorn’s algorithm) but they didn’t know if it was actually doing a good job or what assumptions were behind its use. So, they came up with a new way to understand how IPF works and even found some limitations where it doesn’t work well. They also created a new method that can help when IPF gets stuck on really sparse networks (where most of the edges are missing). Finally, they tested their ideas on fake data and real-world data to see if they were useful. |
Keywords
* Artificial intelligence * Likelihood
