Research map / Adversarial robustness and security / Certified and empirical robustness
FIMAP: Feature Importance by Minimal Adversarial Perturbation
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use FIMAP instead of LIME to generate 2000 explanations | Half moons explanations for a Random Forest · Half moons | explanation generation time: 214 seconds -> 1.32e-2 seconds | Helped |
| Use FIMAP explanations instead of LIME explanations | Half moons instances · Half moons | mean cosine similarity: 0.936; 0.964 | Mixed |
| Train FIMAP to approximate the Random Forest's labels and produce minimal adversarial perturbations | Half moons classification · Half moons test | adversarial accuracy: 0.977 | Helped |
| Use FIMAP to generate minimal perturbations | MNIST 8 vs 3 classification · MNIST 8 vs 3 test | adversarial accuracy: 0.975 | Helped |
| Use FIMAP with categorical and continuous feature handling | UCI Adult subset classification · UCI Adult subset | adversarial accuracy: 0.853 | Helped |
Closest papers by meaning
- Conditional Feature Importance with Generative Modeling Using Adversarial Random Forests (AAAI 2025 · similarity 0.87)
- Feature Attribution with Necessity and Sufficiency via Dual-stage Perturbation Test for Causal Explanation (ICML 2024 · similarity 0.85)
- Keep the Faith: Faithful Explanations in Convolutional Neural Networks for Case-Based Reasoning (AAAI 2024 · similarity 0.84)
- Testing Semantic Importance via Betting (NeurIPS 2024 · similarity 0.84)
- Counterfactual Maps: What They Are and How to Find Them (NeurIPS 2026 · similarity 0.84)
- Counterfactual Explanations for Misclassified Images: How Human and Machine Explanations Differ (Abstract Reprint) (AAAI 2024 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.