Students in the Big Data Analytics MA Program won a prize at the NeuroHive Hackathon. Ivan Bondarenko, a teacher at the NSU Humanities Institute, led the team.
"NeuroHive 2018", an open source online hackathon, was held from April 20 to May 20, 2018. To participate you had to develop a project (application, site, bot in Telegram, mod in your favorite game), using neural networks and machine learning. The primary focus was computer vision, work with images, sound, video, texts, and other content. 18 projects were developed.
MA students Pyotr Gusev and Anastasia Malysheva developed a service "Soroka" and won fourth place in the hackathon for developers of artificial neural networks. "Soroka" is a neural network product that analyzes the reputation of a person or company on the Internet. The project uses convolutional neural networks that are successfully used for image analysis. Recently, this approach gained popularity in tasks related to processing natural language. A number of publications have demonstrated their advantages over other algorithms. For example, the quality of a convolutional network is comparable to the quality of a recurrent network such as LSTM or GRU, while the learning time of a convolutional network is several times less. "Soroka" will help monitor its own reputation and the reputation of its company so it will be the first to learn about significant public opinion trends and to react in a timely manner to any changes. Engineer-Developer Gusev reacted to their result, “We managed to create a significant product in a very short time. It is inspiring!”. Hackathon winners were chosen by the participants in online voting.