Research map / Time series / Anomaly detection
MEMTO: Memory-guided Transformer for Multivariate Time Series Anomaly Detection
Junho Song, Keonwoo Kim, Jeonglyul Oh, Sungzoon Cho
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add LSD to ISD in anomaly criterion (ablation) | MEMTO · 5 datasets avg | F1-score: 69.23 -> 95.73 | Improved |
| add ISD to LSD in anomaly criterion (ablation) | MEMTO · 5 datasets avg | F1-score: 60.16 -> 95.73 | Improved |
| add Gated memory module (ablation) | MEMTO · 5 datasets avg | F1-score: 69.46 -> 95.73 | Improved |
| use Gated memory instead of MemAE memory (with K-means) (ablation) | MEMTO · 5 datasets avg | F1-score: 94.21 -> 95.73 | Improved |
| use Gated memory instead of MNAD memory (with K-means) (ablation) | MEMTO · 5 datasets avg | F1-score: 94.64 -> 95.73 | Improved |
| initialize memory with K-means (two-phase training) (ablation) | MEMTO · 5 datasets avg | F1-score: 87.33 -> 95.73 | Improved |
| initialize memory with K-means (two-phase training) (ablation) | MemAE · 5 datasets avg | F1-score: 79.50 -> 94.21 | Improved |
| initialize memory with K-means (two-phase training) (ablation) | MNAD · 5 datasets avg | F1-score: 79.23 -> 94.64 | Improved |
| add entropy loss to reconstruction loss (ablation) | MEMTO · 5 datasets avg | F1-score: 94.74 -> 95.73 | Improved |
| add reconstruction loss to entropy loss (ablation) | MEMTO · 5 datasets avg | F1-score: 93.75 -> 95.73 | Improved |
| use MEMTO instead of Anomaly Transformer (ablation) | Anomaly Transformer · 5 datasets avg | F1-score: 93.62 -> 95.74 | Improved |
| two-phase training with K-means adds training time (ablation) | MEMTO · SWaT | training time (s): 49.62 -> 121.7 | Hurt |
Later papers that compared with this method
3 later accepted papers report MEMTO as a baseline in their results tables; 3 of them report a better number on most of the results they share with it.
- PGRF-Net: A Prototype-Guided Relational Fusion Network for Diagnostic Multivariate Time-Series Anomaly Detection (ICLR 2026 · better on 2 of 2 shared results)
- Low Rank Transformer for Multivariate Time Series Anomaly Detection and Localization (ICLR 2026 · better on 1 of 1 shared result)
- GCAD: Anomaly Detection in Multivariate Time Series from the Perspective of Granger Causality (AAAI 2025 · better on 2 of 2 shared results)
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- Anomaly Transformer — Anomaly Transformer: Time Series Anomaly Detection with Association Discrepancy (ICLR 2022) · better on 7 of 8 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- The Elephant in the Room: Towards A Reliable Time-Series Anomaly Detection Benchmark (NeurIPS 2024 · similarity 0.87)
- Structured Temporal Causality for Interpretable Multivariate Time Series Anomaly Detection (NeurIPS 2025 · similarity 0.87)
- TAIL-MIL: Time-Aware and Instance-Learnable Multiple Instance Learning for Multivariate Time Series Anomaly Detection (AAAI 2025 · similarity 0.86)
- Detect, Explain, Interpret: An End-to-End Benchmark for Time Series Anomaly Detection, Explainability and Interpretability. (NeurIPS 2026 · similarity 0.86)
- Towards a General Time Series Anomaly Detector with Adaptive Bottlenecks and Dual Adversarial Decoders (ICLR 2025 · similarity 0.86)
- Position: Quo Vadis, Unsupervised Time Series Anomaly Detection? (ICML 2024 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
