Proceedings

Constraint-Data-Value-Maximization: Utilizing Data Attribution for Effective Data Pruning in Low-Data Environments

Danilo Brajovic, David A. Kreplin, Marco F. Huber

Download BibTeX Cite

Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence. Main Track. Pages 2438-2446.

https://doi.org/10.24963/ijcai.2026/271

Keywords

  • Machine Learning: Other
  • Machine Learning: Active learning
  • Data Mining: Anomaly/outlier detection
  • Machine Learning: Explainable/Interpretable machine learning
  • Data Mining: Big data and scalability

Abstract

Attributing model behavior to training data is an evolving research field.

A common benchmark is data removal, which involves eliminating data instances with either low or high values, then assessing a model's performance trained on the modified dataset. Many existing studies leverage Shapley-based data values for this task. In this paper, we demonstrate that these data values are not optimally suited for pruning low-value data when only a limited amount of data remains. To address this limitation, we introduce the Constraint-Data-Value-Maximization (CDVM) approach, which effectively utilizes data attributions for pruning in low-data scenarios. By casting pruning as a constrained optimization that both maximizes total influence and penalizes excessive per-test contributions, CDVM delivers robust performance when only a small fraction of the data is retained. On the OpenDataVal benchmark, CDVM shows strong performance and competitive runtime.

Danilo Brajovic, David A. Kreplin, and Marco F. Huber. 2026. Constraint-Data-Value-Maximization: Utilizing Data Attribution for Effective Data Pruning in Low-Data Environments. In Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence, IJCAI-26. IJCAI, 2438–2446. https://doi.org/10.24963/ijcai.2026/271
Welcome Future Conferences AI Journal Division IJCAI Proceedings About IJCAI IJCAI Awards