Learning from Anonymized and Incomplete Tabular Data
This addresses the challenge of maintaining machine learning utility for datasets with mixed privacy controls, which is incremental as it builds on existing imputation and LLM-based approaches.
The paper tackled the problem of machine learning on tabular data with user-driven privacy, where data includes original, generalized, and missing values, by proposing novel data transformation strategies that account for heterogeneous anonymization. The results demonstrated that these methods reliably regain utility, with generalized values being preferable to suppression and the best strategy depending on the scenario.
User-driven privacy allows individuals to control whether and at what granularity their data is shared, leading to datasets that mix original, generalized, and missing values within the same records and attributes. While such representations are intuitive for privacy, they pose challenges for machine learning, which typically treats non-original values as new categories or as missing, thereby discarding generalization semantics. For learning from such tabular data, we propose novel data transformation strategies that account for heterogeneous anonymization and evaluate them alongside standard imputation and LLM-based approaches. We employ multiple datasets, privacy configurations, and deployment scenarios, demonstrating that our method reliably regains utility. Our results show that generalized values are preferable to pure suppression, that the best data preparation strategy depends on the scenario, and that consistent data representations are crucial for maintaining downstream utility. Overall, our findings highlight that effective learning is tied to the appropriate handling of anonymized values.