Research map / Architectures and efficiency / CNN design
Interpretable Convolutional Neural Networks
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 |
|---|---|---|---|
| Add interpretable filters and local filter loss | VGG-M · ILSVRC Part multi-category, logistic | classification accuracy: 96.73 -> 97.99 | Helped |
| Add interpretable filters and local filter loss | VGG-M · Pascal VOC Part multi-category, logistic and softmax | classification accuracy: 93.88 -> 96.19; 81.93 -> 88.03 | Helped |
| Add interpretable filters and local filter loss | VGG-S · Pascal VOC Part multi-category, logistic and softmax | classification accuracy: 94.05 -> 96.78; 78.15 -> 86.13 | Helped |
| Add interpretable filters and local filter loss | VGG-16 · Pascal VOC Part multi-category, logistic and softmax | classification accuracy: 97.97 -> 98.50; 89.71 -> 91.60 | Helped |
| Add interpretable filters and local filter loss | VGG-16 · Pascal VOC Part single-category | classification accuracy: 98.66 -> 95.39 | Hurt |
| Add interpretable filters and local filter loss | VGG-16 · CUB200 single-category | classification accuracy: 98.91 -> 96.51 | Hurt |
| Add interpretable filters and local filter loss | VGG-16 · Pascal VOC Part single-category | part interpretability: 0.495 -> 0.770 | Helped |
| Add interpretable filters and local filter loss | AlexNet · ILSVRC 2013 DET Animal-Part single-category | location instability: 0.146 -> 0.091 | Helped |
| Add interpretable filters and local filter loss | VGG-M · CUB200-2011 single-category | location instability: 0.148 -> 0.065 | Helped |
| Use the interpretable CNN rather than the ordinary CNN | CNNs across the tested structures · Single-category classification | classification accuracy | Hurt |
Closest papers by meaning
- PICNN: A Pathway towards Interpretable Convolutional Neural Networks (AAAI 2024 · similarity 0.88)
- NeSyFOLD: A Framework for Interpretable Image Classification (AAAI 2024 · similarity 0.85)
- Interpretable Image Classification via Non-parametric Part Prototype Learning (CVPR 2025 · similarity 0.84)
- B-cosification: Transforming Deep Neural Networks to be Inherently Interpretable (NeurIPS 2024 · similarity 0.84)
- Measuring Per-Unit Interpretability at Scale Without Humans (NeurIPS 2024 · similarity 0.83)
- Automated Natural Language Explanation of Deep Visual Neurons with Large Models (Student Abstract) (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.