Feature Visualization in CNNs
GPTKB entity
Statements (33)
| Predicate | Object |
|---|---|
| gptkbp:instanceOf |
Computer Vision Technique
|
| gptkbp:appliesTo |
Convolutional Neural Networks
|
| gptkbp:enables |
Feature Attribution
Bias Detection Model Debugging |
| https://www.w3.org/2000/01/rdf-schema#label |
Feature Visualization in CNNs
|
| gptkbp:method |
Deconvolution
Activation Maximization Class Activation Mapping Saliency Maps Feature Inversion |
| gptkbp:popularizedBy |
Zeiler and Fergus (2014)
|
| gptkbp:relatedTo |
gptkb:Explainable_AI
Backpropagation Gradient Ascent Image Synthesis Neural Network Interpretability Neural Network Transparency |
| gptkbp:supportedBy |
Hierarchical Feature Learning
Class-specific Features Spatial Patterns |
| gptkbp:usedFor |
Interpreting learned features
Understanding internal representations |
| gptkbp:uses |
gptkb:Lucid
gptkb:TensorBoard gptkb:DeepDream NetDissect |
| gptkbp:visualizes |
Filters
Feature Maps Layer Outputs Neuron Activations |
| gptkbp:bfsParent |
gptkb:Transformer_Circuits
|
| gptkbp:bfsLayer |
8
|