Sequential Drift Detection in Deep Learning Classifiers
This work addresses data drift detection for deep learning systems, presenting an incremental improvement with a new loss function for balancing trade-offs.
The paper tackles the problem of detecting data drift in deep learning classifiers by using neural network embeddings within a sequential decision framework, achieving control over false alarm rates while balancing detection speed and accuracy.
We utilize neural network embeddings to detect data drift by formulating the drift detection within an appropriate sequential decision framework. This enables control of the false alarm rate although the statistical tests are repeatedly applied. Since change detection algorithms naturally face a tradeoff between avoiding false alarms and quick correct detection, we introduce a loss function which evaluates an algorithm's ability to balance these two concerns, and we use it in a series of experiments.