Nonlinguistic Information Extraction by Semi-Supervised Techniques : доклад, тезисы доклада | Научно-инновационный портал СФУ

Nonlinguistic Information Extraction by Semi-Supervised Techniques : доклад, тезисы доклада

Тип публикации: доклад, тезисы доклада, статья из сборника материалов конференций

Конференция: International Conference on Informatics in Control, Automation and Robotics, ICINCO 2017; Madrid; Madrid

Год издания: 2017

Ключевые слова: Nonlinguistic Information Extraction, Semi-supervised learning, bio-inspired algorithms, evolutionary algorithms

Аннотация: The concept of nonlinguistic information includes all types of extra linguistic information such as factors of age, emotion and physical states, accent and others. Semi-supervised techniques based on using both labelled and unlabelled examples can be an efficient tool for solving nonlinguistic information extraction problems with large amounts of unlabelled data. In this paper a new cooperation of biology related algorithms (COBRA) for semi-supervised support vector machines (SVM) training and a new selfconfiguring genetic algorithm (SelfCGA) for the automated design of semi-supervised artificial neural networks (ANN) are presented. Firstly, the performance and behaviour of the proposed semi-supervised SVMs and semi-supervised ANNs were studied under common experimental settings; and their workability was established. Then their efficiency was estimated on a speech-based emotion recognition problem.

Ссылки на полный текст

Издание

Журнал: ICINCO 2017

Выпуск журнала: 1

Номера страниц: 312-317

Издатель: SCITEPRESS – Science and Technology Publications

Авторы

  • Semenkina M. (Reshetnev Siberian State University of Science and Technology)
  • Akhmedova Sh. (Reshetnev Siberian State University of Science and Technology)
  • Semenkin E. (Reshetnev Siberian State University of Science and Technology)

Вхождение в базы данных

Информация о публикациях загружается с сайта службы поддержки публикационной активности СФУ. Сообщите, если заметили неточности.

Вы можете отметить интересные фрагменты текста, которые будут доступны по уникальной ссылке в адресной строке браузера.