Visualizar redes neuronales profundas

Represente el progreso del entrenamiento, evalúe la precisión, explique predicciones y visualice las características aprendidas por una red de imágenes

Monitorice el progreso del entrenamiento usando gráficas integradas de precisión y pérdida de red. Investigue redes entrenadas usando técnicas de visualización como Grad-CAM, sensibilidad de oclusión, LIME y Deep Dream.

Apps

Deep Network Designer

Diseñar y visualizar redes de deep learning

Funciones

expandir todo

Gráficas

`analyzeNetwork`	Analyze deep learning network architecture
`trainingProgressMonitor`	Monitor and plot training progress for deep learning custom training loops (desde R2022b)
`updateInfo`	Update information values for custom training loops (desde R2022b)
`recordMetrics`	Record metric values for custom training loops (desde R2022b)
`groupSubPlot`	Group metrics in training plot (desde R2022b)
`plot`	Representar una arquitectura de red neuronal

Visualización de características

`predict`	Compute deep learning network output for inference (desde R2019b)
`minibatchpredict`	Mini-batched neural network prediction (desde R2024a)
`scores2label`	Convert prediction scores to labels (desde R2024a)
`deepDreamImage`	Visualize network features using deep dream
`occlusionSensitivity`	Explain network predictions by occluding the inputs (desde R2019b)
`imageLIME`	Explain network predictions using LIME (desde R2020b)
`gradCAM`	Explain network predictions using Grad-CAM (desde R2021a)
`confusionchart`	Crear una gráfica de matriz de confusión para un problema de clasificación
`sortClasses`	Sort classes of confusion matrix chart
`rocmetrics`	Receiver operating characteristic (ROC) curve and performance metrics for binary and multiclass classifiers (desde R2022b)
`addMetrics`	Compute additional classification performance metrics (desde R2022b)
`average`	Compute performance metrics for average receiver operating characteristic (ROC) curve in multiclass problem (desde R2022b)

Propiedades

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

Temas

Interpretabilidad

Deep Learning Visualization Methods
Learn about and compare deep learning visualization methods.
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.
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 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.

Progreso y rendimiento del entrenamiento

Clasificar imágenes de una webcam mediante deep learning
Este ejemplo muestra cómo clasificar imágenes de una webcam en tiempo real usando una red neuronal convolucional profunda preentrenada, GoogLeNet.
Monitorizar el progreso del entrenamiento de deep learning
Este ejemplo muestra cómo monitorizar el proceso de entrenamiento en redes de deep learning.
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.

Ejemplos destacados

Detect Vanishing Gradients in Deep Neural Networks by Plotting Gradient Distributions

Monitor vanishing gradients while training a deep neural network.

Abrir script en vivo

Explore Network Predictions Using Deep Learning Visualization Techniques

Investigate network predictions using deep learning visualization techniques.

Abrir script en vivo

Grad-CAM revela cómo se realiza la toma de decisiones en deep learning

En este ejemplo se muestra cómo utilizar la técnica de aplicación de activación de clase ponderada por gradiente (Grad-CAM) para entender por qué una red de deep learning toma unas decisiones de clasificación concretas. Grad-CAM, inventado por Selvaraju y otros autores [1], utiliza el gradiente de la puntuación de clasificación con respecto a las características convolucionales determinadas por la red para entender qué partes de la imagen son más importantes para la clasificación. En este ejemplo se usa la red preentrenada GoogLeNet para las imágenes.

Abrir script en vivo

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.

Abrir script en vivo

Understand Network Predictions Using LIME

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

Abrir script en vivo

Explore Semantic Segmentation Network Using Grad-CAM

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

Abrir script en vivo

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.

Abrir script en vivo

Deep Dream Images Using GoogLeNet

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

Abrir script en vivo