What’s New in MATLAB for Deep Learning?

MATLAB makes deep learning easy and accessible for everyone, even if you’re not an expert. Check out the latest features for designing and building your own models, network training and visualization, and deployment.


Experiment Manager App :

Manage multiple deep learning experiments, keep track of training parameters, and analyze and compare results and code

>> Learn more


Data Preparation and Labeling

  • Video Labeler: Label ground-truth data in a video or image sequences

  • Audio Labeler: Interactively define and visualize ground-truth labels for audio datasets

  • New Signal Labeler: Visualize and label signals interactively

  • New Pixel label datastore: Store pixel information for 2D and 3D semantic segmentation data

  • New Audio datastore: Manage large collections of audio recordings

  • New Image datastore: Support for 3D data


Network Architectures

  • New Build advanced network architectures like GANs, Siamese networks, attention networks, and variational autoencoders

  • Train a “you-only-look-once” (YOLO) v2 deep learning object detector and generate C and CUDA code

  • Deep Network Designer: Graphically design and analyze deep networks and generate MATLAB code

  • Custom layers support: Define new layers with multiple inputs and outputs, and specify loss functions for classification and regression

  • Combine LSTM and convolutional layers for video classification and gesture recognition


Deep Learning Interoperability

  • Import and export models with other deep learning frameworks using the ONNX model format and generate CUDA code

  • New Ability to work with MobileNet-v2, ResNet-101, Inception-v3, SqueezeNet, NASNet-Large, and Xception

  • Import TensorFlow-Keras models and generate C, C++ and CUDA code

  • Import DAG networks in Caffe model importer

See a comprehensive list of pretrained models supported in MATLAB.

Networking training.jpg

Network Training

  • Automatically validate network performance, and stop training when the validation metrics stop improving

  • New Train deep learning networks on 3D image data

  • Perform hyperparameter tuning using Bayesian optimization

  • Additional optimizers for training: Adam and RMSProp

  • Train DAG networks in parallel and on multiple GPUs

  • Train deep learning models on NVIDIA DGX and cloud platforms


Debugging and Visualization

  • DAG activations: Visualize intermediate activations for networks like ResNet-50, ResNet-101, GoogLeNet, and Inception-v3

  • Monitor training progress with plots for accuracy, loss, and validation metrics

  • Network Analyzer: Visualize, analyze, and find problems in network architectures before training

  • New Visualize activations of LSTM networks and use Grad-CAM to understand classification decisions

Debuging visualization.jpg


  • New Generate code for networks such as YOLO V2 object detector, DeepLab-v3+, MobileNet-v2, Xception, DenseNet-201, and recurrent networks

  • New Deploy deep learning networks to ARM Mali GPUs

  • New Automated deployment to Jetson AGX Xavier and Jetson Nano platforms

  • Apply CUDA optimized transposes using shared memory for improved performance


Reinforcement Learning

  • New Reinforcement Learning Algorithms: Train deep neural network policies using DQN, DDPG, A2C, PPO, and other algorithms

  • Environment Modeling: Create MATLAB and Simulink models to represent environments and provide observation and reward signals for training policies

  • Training Acceleration: Parallelize policy training on GPUs and multicore CPUs

  • New Reference Examples: Implement policies for automated driving, robotics, and control design applications

Reinforcement learning image.jpg
Related Resources

Introducing Deep Learning with MATLAB

Practical Deep Learning Examples with MATLAB

Deep Learning for Signal Processing with MATLAB

Have Questions About Using MATLAB for Deep  Learning

30 Days of Exploration at Your Fingertips