Summary of Latent Conditional Diffusion-based Data Augmentation For Continuous-time Dynamic Graph Model, by Yuxing Tian and Yiyan Qi and Aiwen Jiang and Qi Huang and Jian Guo
Latent Conditional Diffusion-based Data Augmentation for Continuous-Time Dynamic Graph Model
by Yuxing Tian, Yiyan Qi, Aiwen Jiang, Qi Huang, Jian Guo
First submitted to arxiv on: 11 Jul 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 Continuous-Time Dynamic Graph (CTDG) models have gained popularity for capturing evolving relationships in real-world networks. However, existing CTDG models face challenges due to noise and limited historical data. To address these issues, we propose Conda, a novel latent diffusion-based graph data augmentation (GDA) method tailored for CTDGs. Conda features a sandwich-like architecture combining Variational Auto-Encoder (VAE) and conditional diffusion models to generate enhanced historical neighbor embeddings for target nodes. Unlike conventional diffusion models, Conda requires only historical neighbor sequence embeddings of target nodes for training, facilitating targeted augmentation. We integrate Conda into the CTDG model and adopt an alternating training strategy to optimize performance. Experimental results on six real-world datasets demonstrate consistent performance improvements, particularly in scenarios with limited historical data. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Imagine trying to understand how relationships change over time. This is hard because there’s noise and not enough information. To make it easier, we created a new way to enhance old data called Conda. Conda uses special techniques to generate more historical information about relationships. Unlike other methods that try to learn everything at once, Conda focuses on what matters most – understanding how relationships change over time. We tested Conda on six real-world datasets and it worked well, especially when there was limited information. |
Keywords
» Artificial intelligence » Data augmentation » Diffusion » Encoder
