Outline
- Title: Geometric structure detector based on neural network
- Who: Names and logins
- Miles Miller-Dickson(mmillerd)
- Xiaoyu Wei (xwei10)
- Yihao Zhou (yzhou153)
- Introduction:
- We would like to examine how geometrical structure can be detected in networks (graphs).
- It has been shown that various geometries in a network can be associated with hierarchical structure and even network security.
- The problem is a classification.
- Related Work:
- Utilizing recently proposed geometric notions of curvature on weighted graphs, they investigate the features of gene co-expression networks derived from large-scale genomic studies of cancer. They find that the curvature of these networks reliably distinguishes between cancer and normal samples, with cancer networks exhibiting higher curvature than their normal counterparts.
- https://www.nature.com/articles/srep12323
- Data:
- The graph dataset.
- We need to do significant preprocessing.
- Methodology:
- The convolutional neural network is used to separate the feature points of the geometric structure, and then the feature points are sampled, and finally the geometric structure is classified by the classifier.
- Metrics:
- The neural network can divide the graph into three classes of geometry of surfaces: positive, negative, and zero curvature.
- Ethics:
- Various geometries in a network can be associated with hierarchical structure and even network security.
- Our work is mainly classification, and deep learning can solve classification tasks well.
Log in or sign up for Devpost to join the conversation.