Facebook to Create a Dataset to Train Dialogue AI

The dataset includes 5 million characters and 700 million dialogues
19 September 2018   724

Researchers from Facebook created a set of training data to improve the effectiveness of neural networks training, specializing in communicating with live users. It includes 5 million characters and 700 million dialogues.

Persona-based network architecture
Persona-based network architecture

In the basis of the software, developers have put a set of PERSONA-CHAT, developed jointly by Facebook experts and scientists from the Montreal Institute of Learning Algorithms. First of all, the increase in the volume of data is notable - the basic data-set contained only about a thousand personalities. But researchers are paying attention to a more important aspect. The PERSONA-CHAT content was created artificially, and the new set was formed on the basis of the Reddit user dialogues.

An interactive neural network, trained on a new set of data, leads more engaging dialogues than networks that did not have access to a collection of personalities. Moreover, the training of systems based on the characters is faster.

Choosing the right data set for learning artificial intelligence is one of the key tasks for developers. The accuracy and productivity of the software being created depends on it. In September 2018, Google in the test mode launched a special tool to find suitable collections.

Nvidia to Open SPADE Source Code

SPADE machine learning system creates realistic landscapes based on rough human sketches
15 April 2019   858

NVIDIA has released the source code for the SPADE machine learning system (GauGAN), which allows for the synthesis of realistic landscapes based on rough sketches, as well as training models associated with the project. The system was demonstrated in March at the GTC 2019 conference, but the code was published only yesterday. The developments are open under the non-free license CC BY-NC-SA 4.0 (Creative Commons Attribution-NonCommercial-ShareAlike 4.0), allowing use only for non-commercial purposes. The code is written in Python using the PyTorch framework.

Sketches are drawn up in the form of a segmented map that determines the placement of exemplary objects on the scene. The nature of the generated objects is set using color labels. For example, a blue fill turns into sky, blue into water, dark green into trees, light green into grass, light brown into stones, dark brown into mountains, gray into snow, a brown line into a road, and a blue line into the river. Additionally, based on the choice of reference images, the overall style of the composition and the time of day are determined. The proposed tool for creating virtual worlds can be useful to a wide range of specialists, from architects and urban planners to game developers and landscape designers.

Objects are synthesized by a generative-adversarial neural network (GAN), which, based on a schematic segmented map, creates realistic images by borrowing parts from a model previously trained on several million photographs. In contrast to the previously developed systems of image synthesis, the proposed method is based on the use of adaptive spatial transformation followed by transformation based on machine learning. Processing a segmented map instead of semantic markup allows you to achieve an exact match of the result and control the style.

To achieve realism, two competing neural networks are used: the generator and the discriminator (Discriminator). The generator generates images based on mixing elements of real photos, and the discriminator identifies possible deviations from real images. As a result, a feedback is formed, on the basis of which the generator begins to assemble more and more qualitative samples, until the discriminator ceases to distinguish them from the real ones.