% IMPORTANT: The following is UTF-8 encoded.  This means that in the presence
% of non-ASCII characters, it will not work with BibTeX 0.99 or older.
% Instead, you should use an up-to-date BibTeX implementation like “bibtex8” or
% “biber”.

@INPROCEEDINGS{Zhao:1049202,
      author       = {Zhao, Xuan and Cao, Zhuo and Bangun, Arya and Scharr, Hanno
                      and Assent, Ira},
      title        = {{C}lassifier {R}econstruction {T}hrough
                      {C}ounterfactual-{A}ware {W}asserstein {P}rototypes},
      reportid     = {FZJ-2025-05284},
      year         = {2025},
      abstract     = {Counterfactual explanations provide actionable insights by
                      identifying minimal input changes required to achieve a
                      desired model prediction. Beyond their interpretability
                      benefits, counterfactuals can also be leveraged for model
                      reconstruction, where a surrogate model is trained to
                      replicate the behavior of a target model. In this work, we
                      demonstrate that model reconstruction can be significantly
                      improved by recognizing that counterfactuals, which
                      typically lie close to the decision boundary, can serve as
                      informative—though less representative—samples for both
                      classes. This is particularly beneficial in settings with
                      limited access to labeled data. We propose a method that
                      integrates original data samples with counterfactuals to
                      approximate class prototypes using the Wasserstein
                      barycenter, thereby preserving the underlying distributional
                      structure of each class. This approach enhances the quality
                      of the surrogate model and mitigates the issue of decision
                      boundary shift, which commonly arises when counterfactuals
                      are naively treated as ordinary training instances.
                      Empirical results across multiple datasets show that our
                      method improves fidelity between the surrogate and target
                      models, validating its effectiveness.},
      month         = {Jul},
      date          = {2025-07-13},
      organization  = {ICML 2025 AIW, Vancouver (Canada), 13
                       Jul 2025 - 19 Jul 2025},
      subtyp        = {Other},
      cin          = {IAS-8},
      cid          = {I:(DE-Juel1)IAS-8-20210421},
      pnm          = {5112 - Cross-Domain Algorithms, Tools, Methods Labs (ATMLs)
                      and Research Groups (POF4-511)},
      pid          = {G:(DE-HGF)POF4-5112},
      typ          = {PUB:(DE-HGF)6},
      doi          = {10.34734/FZJ-2025-05284},
      url          = {https://juser.fz-juelich.de/record/1049202},
}