Summary of Treemil: a Multi-instance Learning Framework For Time Series Anomaly Detection with Inexact Supervision, by Chen Liu et al.
TreeMIL: A Multi-instance Learning Framework for Time Series Anomaly Detection with Inexact Supervision
by Chen Liu, Shibo He, Haoyu Liu, Shizhong Li
First submitted to arxiv on: 20 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 This paper proposes a tree-based multi-instance learning (MIL) framework, called TreeMIL, for time series anomaly detection (TSAD). The traditional MIL approach focuses on individual point anomalies, but TreeMIL addresses collective anomalies that exhibit abnormal patterns over subsequences. The framework consists of an N-ary tree structure to divide the entire series into nodes with different levels and lengths. Subsequence features are extracted to determine the presence of collective anomalies, and point-level anomaly scores are calculated by aggregating features from nodes at different levels. Experimental results on seven public datasets and eight baselines show that TreeMIL achieves an average 32.3% improvement in F1-score compared to previous state-of-the-art methods. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper helps us detect unusual patterns in time series data, like medical records or network traffic. Right now, it’s hard to get precise labels for these patterns, so we use a special kind of training that only looks at the overall pattern, not individual points. The new method, called TreeMIL, can spot both small and big anomalies, which is important because sometimes unusual things happen over time, not just at one point. |
Keywords
* Artificial intelligence * Anomaly detection * F1 score * Time series
