Summary of Warped Time Series Anomaly Detection, by Charlotte Lacoquelle et al.
Warped Time Series Anomaly Detection
by Charlotte Lacoquelle, Xavier Pucel, Louise Travé-Massuyès, Axel Reymonet, Benoît Enaux
First submitted to arxiv on: 18 Apr 2024
Categories
- Main: Artificial Intelligence (cs.AI)
- Secondary: Signal Processing (eess.SP)
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 proposes WarpEd Time Series ANomaly Detection (WETSAND), a novel approach for detecting time series outliers in systems with repetitive behavior. The method consists of three stages: first, identifying repetitive cycles and segmenting lengthy time series into individual task cycles; second, computing a prototype using a GPU-based barycenter algorithm; and third, detecting abnormal cycles by calculating an anomaly score. WETSAND leverages the Dynamic Time Warping algorithm and its variants to handle distorted time series data. Experimental results demonstrate that WETSAND scales well with large signals, produces human-friendly prototypes, requires minimal data, and outperforms general-purpose anomaly detection methods like autoencoders. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary WarpEd Time Series ANomaly Detection (WETSAND) is a new way to find unusual patterns in time series data. This kind of data comes from things like industrial robots that perform tasks many times. The problem with these systems is that the same task can take different amounts of time each time it’s done, and there are gaps in the data. WETSAND has three steps: first, find the repeated patterns and break them down into smaller parts; second, create a prototype using special math to handle big data; and third, use this prototype to find unusual events. This approach uses Dynamic Time Warping to make it work well with distorted data. It was tested on large datasets and worked better than some other methods. |
Keywords
» Artificial intelligence » Anomaly detection » Time series
