Nvidia to Represent Self-Driving Safety Report

In particular, the creators said that the main aspect of security is the autonomy of an unmanned vehicle
24 October 2018   499

NVIDIA published a report on the safety of unmanned vehicles, in which it tried to highlight all aspects of the new machines. The main thing that made the emphasis in the document - is proof of the safety of autopilots.

In the report, the company described some technical development processes. In particular, the creators said that the main aspect of security is the autonomy of an unmanned vehicle. In other words, the onboard AI must have sufficient performance not to apply for commands to the control room, but to make a decision independently in most cases. However, this is not always the case.

The report states that the safety of travel on unmanned vehicles is based on four fundamental principles.

The first is a powerful platform of artificial intelligence with the possibility of deep learning. Such a platform should be able to work with sensors, surround vision systems and other devices. Of course, this will require a powerful hardware.

The second important aspect is data processing. It is assumed that each vehicle will generate petabytes of data, so some of the information will need to be processed remotely in data centers. All this will require a completely new computing architecture and infrastructure.

Self-Driving Safety Report Illustration
Self-Driving Safety Report Illustration

The third principle is testing of unmanned vehicles in real conditions, on ordinary roads and under various driving conditions. This will allow to collect more information about the trip and supplement it with conventional computer simulation.

The fourth and last are security measures for drivers involved in testing. The NVIDIA report noted that test participants must be trained before operating such machines. 

In 2017, NVIDIA introduced a new version of the Drive PX platform for unmanned vehicles. The company claimed that it would become the basis for autonomy of the fifth level (full autonomy of unmanned vehicles), but so far this has not happened.

Neural Network to Create Landscapes from Sketches

Nvidia created GauGAN model that uses generative-competitive neural networks to process segmented images and create beautiful landscapes from peoples' sketches
20 March 2019   172

At the GTC 2019 conference, NVIDIA presented a demo version of the GauGAN neural network, which can turn sketchy drawings into photorealistic images.

The GauGAN model, named after the famous artist Paul Gauguin, uses generative-competitive neural networks to process segmented images. The generator creates an image and transfers it to the discriminator trained in real photographs. He in turn pixel-by-pixel tells the generator what to fix and where.

Simply put, the principle of the neural network is similar to the coloring of the coloring, but instead of children's drawings, it produces beautiful landscapes. Its creators emphasize that it does not just glue pieces of images, but generates unique ones, like a real artist.

Among other things, the neural network is able to imitate the styles of various artists and change the times of the day and year in the image. It also generates realistic reflections on water surfaces, such as ponds and rivers.

So far, GauGAN is configured to work with landscapes, but the neural network architecture allows us to train it to create urban images as well. The source text of the report in PDF is available here.

GauGAN can be useful to both architects and city planners, and landscape designers with game developers. An AI that understands what the real world looks like will simplify the implementation of their ideas and help you quickly change them. Soon the neural network will be available on the AI ​​Playground.