MathWorks annoncerer release 2019b af MATLAB og Simulink
Mathworks har netop præsenteret release 2019b af MATLAB og Simulink værktøjerne, der begge inkluderer markante nyheder (in english).MathWorks has introduced Release 2019b with a range of new capabilities in MATLAB and Simulink, including those in support of artificial intelligence, deep learning and the automotive industry. In addition, R2019b introduces new products in support of robotics, new training resources for event-based modeling, and updates and bug fixes across the MATLAB and Simulink product families.
Among the MATLAB highlights in R2019b is the introduction of Live Editor Tasks, which enables users to interactively explore parameters, preprocess data, and generate MATLAB code that becomes part of the live script. Now, MATLAB users can focus on the task instead of the syntax or complex code, and automatically run generated code to quickly iterate on parameters through visualization.
R2019b highlights of Simulink include the new Simulink Toolstrip, which helps users access and discover capabilities as they are needed. In the Simulink Toolstrip tabs are arranged according to workflow and sorted by frequency of use, saving navigation and search time.
Artificial intelligence and deep learning
In R2019b, Deep Learning Toolbox builds on the flexible training loops and networks introduced earlier this year. New capabilities enable users to train advanced network architectures using custom training loops, automatic differentiation, shared weights, and custom loss functions.
In addition, users can now build generative adversarial networks (GANs), Siamese networks, variational autoencoders, and attention networks. Deep Learning Toolbox also can now export to ONNX format networks that combine CNN and LSTM layers and networks that include 3D CNN layers.
R2019b also introduces significant capabilities in support of the automotive industry across multiple products, including:
Automated Driving Toolbox: Support for 3D simulation, including the ability to develop, test, and verify driving algorithms in a 3D environment, and a block that enables users to generate the velocity profile of a driving patch given kinematic constraints.
Powertrain Blockset: Ability to generate a deep learning SI engine model for algorithm design and performance, fuel economy, and emissions analysis.
Sensor Fusion and Tracking Toolbox: Ability to perform track-to-track fusion and architect decentralized tracking systems.
Polyspace Bug Finder: increased support of AUTOSAR C++14 coding guidelines to check for misuse of lambda expressions, potential problems with enumerations, and other issues.