Visualization and Interpretability

Plot training progress, assess accuracy, explain predictions, and visualize features learned by a network

Monitor training progress using built-in plots of network accuracy and loss. Investigate trained networks using visualization techniques such as Grad-CAM, occlusion sensitivity, LIME, and deep dream.

Deep Learning Visualization Methods

Flow chart showing methods used to visualize the predictions made by a trained network. The visualization methods can be used for verification, debugging, learning, detecting bias, and model selection.

Deep Learning Visualization Methods

Apps

Deep Network Designer

Design and visualize deep learning networks

Objekte

trainingProgressMonitor Monitor and plot training progress for deep learning custom training loops (Seit R2022b)

Funktionen

alle erweitern

Network

`analyzeNetwork`	Analyze deep learning network architecture
`plot`	Plot neural network architecture

Training Progress

`updateInfo`	Update information values for custom training loops (Seit R2022b)
`recordMetrics`	Record metric values for custom training loops (Seit R2022b)
`groupSubPlot`	Group metrics in training plot (Seit R2022b)
`yscale`	Set training plot y-axis scale (linear or logarithmic) (Seit R2024a)

Metrics

`accuracyMetric`	Deep learning accuracy metric (Seit R2023b)
`aucMetric`	Deep learning area under ROC curve (AUC) metric (Seit R2023b)
`fScoreMetric`	Deep learning F-score metric (Seit R2023b)
`precisionMetric`	Deep learning precision metric (Seit R2023b)
`recallMetric`	Deep learning recall metric (Seit R2023b)
`rmseMetric`	Deep learning root mean squared error metric (Seit R2023b)

Prediction

`predict`	Compute deep learning network output for inference (Seit R2019b)
`minibatchpredict`	Mini-batched neural network prediction (Seit R2024a)
`scores2label`	Convert prediction scores to labels (Seit R2024a)
`confusionchart`	Create confusion matrix chart for classification problem
`sortClasses`	Sort classes of confusion matrix chart
`rocmetrics`	Receiver operating characteristic (ROC) curve and performance metrics for binary and multiclass classifiers (Seit R2022b)
`addMetrics`	Compute additional classification performance metrics (Seit R2022b)
`average`	Compute performance metrics for average receiver operating characteristic (ROC) curve in multiclass problem (Seit R2022b)
`plot`	Plot receiver operating characteristic (ROC) curves and other performance curves (Seit R2022b)

Interpretability

`imageLIME`	Explain network predictions using LIME (Seit R2020b)
`occlusionSensitivity`	Explain network predictions by occluding the inputs (Seit R2019b)
`deepDreamImage`	Visualize network features using deep dream
`gradCAM`	Explain network predictions using Grad-CAM (Seit R2021a)
`drise`	Explain object detection network predictions using D-RISE (Seit R2024a)

Eigenschaften

ConfusionMatrixChart Properties	Confusion matrix chart appearance and behavior
ROCCurve Properties	Receiver operating characteristic (ROC) curve appearance and behavior (Seit R2022b)

Themen

Training Progress and Performance

Monitor Deep Learning Training Progress
This example shows how to monitor the training progress of deep learning networks.
Monitor Custom Training Loop Progress
Track and plot custom training loop progress.
Monitor GAN Training Progress and Identify Common Failure Modes
Learn how to diagnose and fix some of the most common failure modes in GAN training.
ROC Curve and Performance Metrics
Use rocmetrics to examine the performance of a classification algorithm on a test data set.
Compare Deep Learning Models Using ROC Curves
This example shows how to use receiver operating characteristic (ROC) curves to compare the performance of deep learning models.
Define Custom Metric Function
Define custom deep learning metrics using functions.
Define Custom Deep Learning Metric Object
Learn how to define custom deep learning metrics using custom classes.
Define Custom F-Beta Score Metric Object
This example shows how to define a generic F-beta score metric.

Interpretability

Explore Network Predictions Using Deep Learning Visualization Techniques
This example shows how to investigate network predictions using deep learning visualization techniques.
Understand Network Predictions Using Occlusion
This example shows how to use occlusion sensitivity maps to understand why a deep neural network makes a classification decision.
Interpret Deep Network Predictions on Tabular Data Using LIME
This example shows how to use the locally interpretable model-agnostic explanations (LIME) technique to understand the predictions of a deep neural network classifying tabular data.
Investigate Spectrogram Classifications Using LIME
This example shows how to use locally interpretable model-agnostic explanations (LIME) to investigate the robustness of a deep convolutional neural network trained to classify spectrograms.
Investigate Classification Decisions Using Gradient Attribution Techniques
This example shows how to use gradient attribution maps to investigate which parts of an image are most important for classification decisions made by a deep neural network.
Investigate Network Predictions Using Class Activation Mapping
This example shows how to use class activation mapping (CAM) to investigate and explain the predictions of a deep convolutional neural network for image classification.
Visualize Image Classifications Using Maximal and Minimal Activating Images
This example shows how to use a data set to find out what activates the channels of a deep neural network.
View Network Behavior Using tsne
This example shows how to use the tsne function to view activations in a trained network.
Visualize Activations of LSTM Network
This example shows how to investigate and visualize the features learned by LSTM networks by extracting the activations.
Visualize Activations of a Convolutional Neural Network
This example shows how to feed an image to a convolutional neural network and display the activations of different layers of the network.
Visualize Features of a Convolutional Neural Network
This example shows how to visualize the features learned by convolutional neural networks.
Deep Learning Visualization Methods
Learn about and compare deep learning visualization methods.

Enthaltene Beispiele

Detect Vanishing Gradients in Deep Neural Networks by Plotting Gradient Distributions

Monitor vanishing gradients while training a deep neural network.

Live Script öffnen

Grad-CAM Reveals the Why Behind Deep Learning Decisions

Use the gradient-weighted class activation mapping (Grad-CAM) technique to understand why a deep learning network makes its classification decisions. Grad-CAM, invented by Selvaraju and coauthors [1], uses the gradient of the classification score with respect to the convolutional features determined by the network in order to understand which parts of the image are most important for classification. This example uses the GoogLeNet pretrained network for images.

Live Script öffnen

Interpret Deep Learning Time-Series Classifications Using Grad-CAM

Use the gradient-weighted class activation mapping (Grad-CAM) technique to understand the classification decisions of a 1-D convolutional neural network trained on time-series data.

Live Script öffnen

Understand Network Predictions Using LIME

Use locally interpretable model-agnostic explanations (LIME) to understand why a deep neural network makes a classification decision.

Live Script öffnen

Explore Semantic Segmentation Network Using Grad-CAM

Explore the predictions of a pretrained semantic segmentation network using Grad-CAM.

Live Script öffnen

Investigate Audio Classifications Using Deep Learning Interpretability Techniques

Use interpretability techniques to investigate the predictions of a deep neural network trained to classify audio data.

Live Script öffnen

Deep Dream Images Using GoogLeNet

Generate images using deepDreamImage with the pretrained convolutional neural network GoogLeNet.

Live Script öffnen

Detect Issues During Deep Neural Network Training

Automatically detect issues while training a deep neural network.

Live Script öffnen

Explore Quantized Semantic Segmentation Network Using Grad-CAM

Compare the predictions of a quantized semantic segmentation network to the original network using the gradient-weighted class activation mapping (Grad-CAM) interpretability method.

Live Script öffnen