Researchers to Develop Accent Detection AI

The team of scientists was from Cisco, the Moscow Institute of Physics and Technology and the Higher School of Economics
13 July 2018   989

A team of scientists used machine learning to develop an improved model for speech recognition. This is reported by Venture Beat.

Previously, scientists manually identified phonological similarities between units of language in general American English and the pronunciation dictionary of the Carnegie Mellon University. To create an improved model, they went non-standard way and allowed it to automatically form the rules. Then, it compared the resulting unique list with a set of examples from George Mason University's speech accents archive.

More non-native accented speech data is necessary to enhance the performance of … existing [speech recognition] models. However, its synthesis is still an open problem.
 

Researchers

Based on the received examples, the team created a phonetic data set, through which a neural network, often used for speech recognition, was trained. The accuracy of the definition of words, after overcoming the mark of 800,000 examples, was 59%.

The study was called preliminary due to fewer sounds in the Carnegie-Mellon University dictionary. Despite phonetic coincidences in 13 out of 20 dictionary comparisons, scientists managed to increase the data array from 103 thousand phonetic transcriptions with one accent to 1 million samples with several accents.

Nvidia to Open SPADE Source Code

SPADE machine learning system creates realistic landscapes based on rough human sketches
15 April 2019   875

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.