Deep Learning / 28
Interpretability
Explore neural-network visualizations, attributions, probing, and the limits of interpretation.
Seeing what a model attends to is not the same as proving why it decided.
This is a knowledge index. Published titles open their articles; the remaining entries outline planned coverage.
Subtopics
- 28.01
Why Interpret Neural Networks?
Planned guide: Why Interpret Neural Networks. Focus on intuition, assumptions, and a worked example.
- 28.02
Feature Visualization
Planned guide: Feature Visualization. Focus on intuition, assumptions, and a worked example.
- 28.03
Activation Visualization
Planned guide: Activation Visualization. Focus on intuition, assumptions, and a worked example.
- 28.04
Saliency Maps
Planned guide: Saliency Maps. Focus on intuition, assumptions, and a worked example.
- 28.05
Grad-CAM
Planned guide: Grad-CAM. Focus on intuition, assumptions, and a worked example.
- 28.06
Integrated Gradients
Planned guide: Integrated Gradients. Focus on intuition, assumptions, and a worked example.
- 28.07
Attention Visualization
Planned guide: Attention Visualization. Focus on intuition, assumptions, and a worked example.
- 28.08
SHAP for Neural Networks
Planned guide: SHAP for Neural Networks. Focus on intuition, assumptions, and a worked example.
- 28.09
Representation Probing
Planned guide: Representation Probing. Focus on intuition, assumptions, and a worked example.
- 28.10
Interpretability Limitations
Planned guide: Interpretability Limitations. Focus on intuition, assumptions, and a worked example.