Cutpurse improves model extraction on tabular data
Original title: Cutpurse: Decomposing Explanation-Guided Model Extraction on Tabular Data
Cutpurse achieves 16.3 percentage point fidelity gain on continuous features without explanation APIs.
Explanation-guided extraction does not reveal privileged model information. Lime bin edges improve fidelity but are reproducible from adversary data. Cutpurse uses self-computed grids with diverse queries to exceed undefended published attacks by 7.0 points. Results concern default Lime and Shap configurations for tabular classifiers.
- 16.3 pp fidelity improvement on continuous targets
- 7.0 pp better than undefended attacks
- No need for explanation API bin edges
- Works under in-domain assumptions
From the original article
1 min read at arxiv.orgExplanation APIs are often treated as increasing model-extraction risk, but an end-to-end attack comparison cannot show whether an explanation reveals model-specific information or merely supplies a convenient next-query coordinate. We study this distinction for explanation-guided extraction of tabular classifiers.
By Seyma Selcan Magara · arxiv.org
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.