Research map / Label-efficient and robust learning / Recognition and long-tail
Beyond Single Path Integrated Gradients for Reliable Input Attribution via Randomized Path Sampling
Giyoung Jeon, Haedong Jeong, Jaesik Choi
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| aggregate attributions over multiple stick-breaking sampled paths instead of a single straight path | Integrated Gradients · ImageNet val, VGG-16 | insertion: 0.096 -> 0.443 | Improved |
| aggregate attributions over sampled paths | Integrated Gradients · ImageNet val, VGG-16 | deletion: 0.036 -> 0.019 | Improved |
| SPI versus IG under ROAR with 10% pixels removed | Integrated Gradients · CIFAR-10 ROAR ResNet-18 | test accuracy: 39.70% -> 31.00% | Improved |
| sample paths with stick-breaking process with uniform base distribution instead of hypergeometric distribution (ablation) | SPI path sampling · path distribution analysis | representation power of sampled paths | Improved |
| use concentration alpha = 10 rather than low alpha (ablation) | SPI · ImageNet insertion/deletion | insertion/deletion game | Improved |
| FullGrad versus other baselines on ROAR at low removal | FullGrad · CIFAR-10 ROAR 10% removed | test accuracy: 67.49% at 10% removed vs IG 39.70% | Hurt |
Closest papers by meaning
- Integrated Decision Gradients: Compute Your Attributions Where the Model Makes Its Decision (AAAI 2024 · similarity 0.91)
- Manifold Integrated Gradients: Riemannian Geometry for Feature Attribution (ICML 2024 · similarity 0.88)
- Denoising Diffusion Path: Attribution Noise Reduction with An Auxiliary Diffusion Model (NeurIPS 2024 · similarity 0.88)
- Iterative Search Attribution for Deep Neural Networks (ICML 2024 · similarity 0.87)
- Path Choice Matters for Clear Attributions in Path Methods (ICLR 2024 · similarity 0.86)
- From Baselines to Transport Geodesics: Axiomatic Attribution via Optimal Generative Flows (NeurIPS 2026 · similarity 0.86)
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.
