Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2410.19607 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866910744524619776 |
|---|---|
| author | Tan, Shuhang Sia, Jayson Bogdan, Paul Ivanov, Radoslav |
| author_facet | Tan, Shuhang Sia, Jayson Bogdan, Paul Ivanov, Radoslav |
| contents | This paper presents a new look at the neural network (NN) robustness problem, from the point of view of graph theory analysis, specifically graph curvature. Graph curvature (e.g., Ricci curvature) has been used to analyze system dynamics and identify bottlenecks in many domains, including road traffic analysis and internet routing. We define the notion of neural Ricci curvature and use it to identify bottleneck NN edges that are heavily used to ``transport data" to the NN outputs. We provide an evaluation on MNIST that illustrates that such edges indeed occur more frequently for inputs where NNs are less robust. These results will serve as the basis for an alternative method of robust training, by minimizing the number of bottleneck edges. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2410_19607 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Analyzing Neural Network Robustness Using Graph Curvature Tan, Shuhang Sia, Jayson Bogdan, Paul Ivanov, Radoslav Machine Learning This paper presents a new look at the neural network (NN) robustness problem, from the point of view of graph theory analysis, specifically graph curvature. Graph curvature (e.g., Ricci curvature) has been used to analyze system dynamics and identify bottlenecks in many domains, including road traffic analysis and internet routing. We define the notion of neural Ricci curvature and use it to identify bottleneck NN edges that are heavily used to ``transport data" to the NN outputs. We provide an evaluation on MNIST that illustrates that such edges indeed occur more frequently for inputs where NNs are less robust. These results will serve as the basis for an alternative method of robust training, by minimizing the number of bottleneck edges. |
| title | Analyzing Neural Network Robustness Using Graph Curvature |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2410.19607 |