Research map / Architectures and efficiency / RNNs and normalization
What does LIME really see in images?
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 averaged gradients along the path from the image to its replacement to approximate LIME coefficients | LIME explanations · CIFAR-10 test | Jaccard similarity of top positive superpixels: J5: 0.99; J10: 0.99 | Helped |
| Increase model complexity from the 1-layer network to VGG-like networks | LIME versus summed integrated gradients · CIFAR-10 test | Jaccard similarity of top positive superpixels: J5: 0.99 -> 0.81 -> 0.75 -> 0.71; J10: 0.99 -> 0.85 -> 0.81 -> 0.76 | Hurt |
| Use larger and more complex pretrained vision models | LIME versus summed integrated gradients · ILSVRC2017 test | Jaccard similarity of top positive superpixels: J5: 0.43 -> 0.42 -> 0.35; J10: 0.54 -> 0.44 -> 0.36 | Hurt |
| Replace the gradient-based approximation with a zero-gradient tree model | LIME versus integrated gradients · CART random forest | Approximate explanation coefficient: 0 | Hurt |
Closest papers by meaning
- SLICE: Stabilized LIME for Consistent Explanations for Image Classification (CVPR 2024 · similarity 0.89)
- Using Stratified Sampling to Improve LIME Image Explanations (AAAI 2024 · similarity 0.88)
- PerceptionCLIP: Visual Classification by Inferring and Conditioning on Contexts (ICLR 2024 · similarity 0.82)
- The Devil is in the Fine-Grained Details: Evaluating Open-Vocabulary Object Detectors for Fine-Grained Understanding (CVPR 2024 · similarity 0.82)
- Linear Differential Vision Transformer: Learning Visual Contrasts via Pairwise Differentials (NeurIPS 2025 · similarity 0.82)
- Probabilistic Conceptual Explainers: Trustworthy Conceptual Explanations for Vision Foundation Models (ICML 2024 · similarity 0.82)
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.