ChronosAD is a two-stage anomaly detection architecture that applies a time series foundation model for zero-shot embeddings followed by a custom Temporal Block, reporting average gains of 4.72% AUC and 6.60% AP over prior methods on 11 benchmarks.
Incorporating Transformer and LSTM to Kalman Filter with EM algorithm for state estimation,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
ChronosAD: Leveraging Time Series Foundation Models for Accurate Anomaly Detection
ChronosAD is a two-stage anomaly detection architecture that applies a time series foundation model for zero-shot embeddings followed by a custom Temporal Block, reporting average gains of 4.72% AUC and 6.60% AP over prior methods on 11 benchmarks.