Reproduce “Efficient computation of counterfactual explanations of LVQ models” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
André Artelt, Barbara Hammer
The authors develop a method for finding counterfactuals in learning vector quantization (LVQ) classifiers. It searches target-class prototypes and formulates input changes as optimization problems tailored to the model and distance measure. In handwritten-digit experiments with GMLVQ, mean Manhattan distance was 4.40, versus 11.63 for CMA-ES. Results are specific to tested settings; localized metrics require approximate non-convex optimization.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.