Open World Autoencoding Drift Detection with Novel Class Recognition in Tabular Non-stationary Data Streams
About
Data stream processing has become a landmark in modern machine learning applications, with concept drifts and novel class appearances posing the primary challenges faced by sophisticated recognition methods. This work proposes an unsupervised concept drift detection method that identifies shifts in known class distributions based on the reconstruction errors of an autoencoder, while also enabling the recognition of novel class samples through density estimation of a proxy representation of samples. Using mirrored autoencoders allows for independent incremental adaptation to changing problem distributions for the two considered tasks, resulting in continuous adjustment to evolving concepts and reliable recognition of unknown samples. Conducted experiments used a diverse set of synthetic tabular data streams, where both concept drifts and the emergence of novelties were observed. The results show that the proposed approach is competitive with current state-of-the-art unsupervised drift detectors and novelty classifiers.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Drift Detection | Synthetic Data Streams various configurations | D1 Measure7.285 | 49 | |
| Novelty Recognition | Synthetic Data Stream N5 S1.0 | Balanced Acc0.614 | 5 | |
| Novelty Recognition | Synthetic Data Stream N5 S2.0 | Balanced Accuracy77.5 | 5 | |
| Novelty Recognition | Synthetic Data Stream N5 | S3.0 | Balanced Accuracy88.2 | 5 | |
| Novelty Recognition | Synthetic Data Stream N9 S1.0 | Balanced Accuracy62.2 | 5 | |
| Novelty Recognition | Synthetic Data Stream N9 S2.0 | Balanced Accuracy76.3 | 5 | |
| Novelty Recognition | Synthetic Data Stream N9 S3.0 | Balanced Accuracy89.9 | 5 |