Summary of Slade: Detecting Dynamic Anomalies in Edge Streams Without Labels Via Self-supervised Learning, by Jongha Lee et al.
SLADE: Detecting Dynamic Anomalies in Edge Streams without Labels via Self-Supervised Learning
by Jongha Lee, Sunwoo Kim, Kijung Shin
First submitted to arxiv on: 19 Feb 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Social and Information Networks (cs.SI)
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 This paper introduces SLADE (Self-supervised Learning for Anomaly Detection in Edge Streams), a deep neural network-based approach that detects anomalies in dynamically changing graphs, such as social networks or email communications. Unlike previous methods, which assume static input graphs, SLADE is designed to instantly detect anomalies as they occur and adapt to shifting graph states. The proposed method trains a self-supervised neural network to minimize drift in node representations and generate long-term interaction patterns from short-term ones. By detecting deviations in nodes’ interaction patterns over time, SLADE identifies abnormal behavior without relying on labeled data. In experiments across four real-world datasets, SLADE outperforms nine competing methods, including those leveraging label supervision. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper is about finding strange behaviors in big networks like social media or email. These networks are constantly changing, and it’s hard to detect when something abnormal happens. The authors created a new way to do this using special computers called neural networks. They trained the network to learn how normal behavior looks like, so it can spot when something goes wrong. This method is important because it doesn’t need labeled data (like “this is an anomaly” or “this is not”) to work well. In fact, their approach even outperformed other methods that did use labeled data! |
Keywords
* Artificial intelligence * Anomaly detection * Neural network * Self supervised
