Cross-Domain Data Selection and Augmentation for Automatic Compliance Detection
For practitioners automating regulatory compliance, this provides a practical method to improve model generalization across different legal texts without requiring new annotations.
The paper tackles cross-domain transfer in automatic compliance detection, showing that targeted data selection from a source domain reduces negative transfer and improves performance across heterogeneous regulations.
Automating the detection of regulatory compliance remains a challenging task due to the complexity and variability of legal texts. Models trained on one regulation often fail to generalise to others. This limitation underscores the need for principled methods to improve cross-domain transfer. We study data selection as a strategy to mitigate negative transfer in compliance detection framed as a natural language inference (NLI) task. Specifically, we evaluate four approaches for selecting augmentation data from a larger source domain: random sampling, Moore-Lewis's cross-entropy difference, importance weighting, and embedding-based retrieval. We systematically vary the proportion of selected data to analyse its effect on cross-domain adaptation. Our findings demonstrate that targeted data selection substantially reduces negative transfer, offering a practical path toward scalable and reliable compliance automation across heterogeneous regulations.