Summary of Non-neighbors Also Matter to Kriging: a New Contrastive-prototypical Learning, by Zhishuai Li et al.
Non-Neighbors Also Matter to Kriging: A New Contrastive-Prototypical Learning
by Zhishuai Li, Yunhao Nie, Ziyue Li, Lei Bai, Yisheng Lv, Rui Zhao
First submitted to arxiv on: 23 Jan 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI)
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 “Contrastive-Prototypical” self-supervised learning for Kriging (KCP) aims to refine valuable information from neighbors and recycle that from non-neighbors. This approach learns robust and general representations through a neighboring contrast module, which narrows the representation distance between the target and its neighbors while pushing away non-neighbors. A prototypical module identifies similar representations via exchanged prediction, refining misleading neighbors and recycling useful non-neighbors. The adaptive augmentation module incorporates data-driven attribute augmentation and centrality-based topology augmentation over spatiotemporal Kriging graph data. This approach demonstrates superior performance on real-world datasets with 6% improvements and exceptional transferability and robustness. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Kriging is a way to estimate things about places that haven’t been measured yet, based on information from nearby locations or connections. But sometimes, this method can be affected by not having enough sensors in certain areas. To fix this, we created a new way of doing Kriging called “Contrastive-Prototypical” self-supervised learning for Kriging (KCP). This approach helps refine information from neighbors and use helpful information from non-neighbors too. It does this by using two special modules: one that makes the representations of places more similar to each other, and another that identifies which representations are most important. We also added a way to make sure these modules learn good representations by using data-driven attribute augmentation and centrality-based topology augmentation over spatiotemporal Kriging graph data. |
Keywords
* Artificial intelligence * Self supervised * Spatiotemporal * Transferability
