Research map / Adversarial robustness and security / Certified and empirical robustness
Hierarchical interpretations for neural network predictions
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 CD rather than occlusion as the group-scoring function for agglomeration | Agglomerative feature hierarchy · MNIST adversarial hierarchy across five attacks | correlation: 0.259 -> 0.762 (Saliency); 0.196 -> 0.662 (Gradient attack); 0.131 -> 0.590 (FGSM); 0.155 -> 0.684 (Boundary); 0.202 -> | Helped |
| Partition convolutional-layer bias between the feature-group and remainder contributions (ablation) | Generalized CD for CNNs · ImageNet CNN interpretations | qualitative score and heatmap quality | Helped |
| Replace the ReLU Shapley decomposition with the proposed ReLU decomposition (ablation) | Generalized CD for CNNs · ImageNet CNN interpretations | qualitative score and heatmap quality | Helped |
| Apply the bias-partition and ReLU-decomposition changes (ablation) | Generalized CD · SST and MNIST interpretations | interpretation quality | No clear effect |
| Use hierarchical ACD interpretations rather than non-hierarchical methods | Interpretation methods · ImageNet human model-selection study | binary accuracy in selecting the more accurate model | Helped |
| Use ACD interpretations rather than other interpretation methods | Interpretation methods · SST human trust study | average trust rank | Helped |
Closest papers by meaning
- Interpretable Part-Whole Hierarchies and Conceptual-Semantic Relationships in Neural Networks (CVPR 2022 · similarity 0.87)
- Explainable Neural Networks with Guarantee: A Sparse Estimation Approach (AAAI 2025 · similarity 0.85)
- Predictive Concept Decoders: Training Scalable End-to-End Interpretability Assistants (NeurIPS 2026 · similarity 0.84)
- Neurons to Words: A Novel Method for Automated Neural Network Interpretability and Alignment (AAAI 2025 · similarity 0.83)
- Understanding Generalization through Decision Pattern Shift (NeurIPS 2026 · similarity 0.83)
- Towards Modeling Uncertainties of Self-Explaining Neural Networks via Conformal Prediction (AAAI 2024 · similarity 0.83)
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.