Summary of Exploring Higher-order Neural Network Node Interactions with Total Correlation, by Thomas Kerby et al.
Exploring higher-order neural network node interactions with total correlation
by Thomas Kerby, Teresa White, Kevin Moon
First submitted to arxiv on: 6 Feb 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Machine Learning (stat.ML)
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 paper proposes a novel method called Local Correlation Explanation (CorEx) to accurately capture higher-order variable interactions (HOIs) in complex systems. HOIs are notoriously difficult to characterize, especially when they change across data. CorEx tackles this challenge by first clustering data points based on their proximity and then using multivariate mutual information to learn local HOIs within each cluster. The authors demonstrate the effectiveness of Local CorEx on synthetic and real-world datasets, revealing hidden insights about data structure. Additionally, they show how Local CorEx can be used to interpret the inner workings of trained neural networks. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Imagine trying to understand how different things in an ecosystem or a team work together. This is hard because many factors interact with each other in complex ways. A new method called Local Correlation Explanation helps solve this problem by grouping similar data points together and then understanding how these groups relate to each other. The authors tested this method on some datasets and found that it can uncover hidden patterns. They also showed that this method can be used to understand what’s happening inside trained computer networks. |
Keywords
* Artificial intelligence * Clustering
