Evgeniya Ustinova

Learn More
We introduce a new representation learning approach for domain adaptation, in which data at training and test time come from similar but different distributions. Our approach is directly inspired by the theory on domain adaptation suggesting that, for effective domain transfer to be achieved, predictions must be made based on features that cannot(More)
In this work we explore the applicability of the recently proposed convolutional neural net architecture, called Bilinear CNN, and its new modification that we call Multiregion Bilinear CNN to the person re-identification problem. Originally, Bilinear CNNs were introduced for fine-grained classification and proved to be both simple and high-performing(More)
Review of various reduced-complexity min-sum based decoding algorithms for LDPC codes are presented. New BP-based approach for LDPC decoding was proposed, and one new a-min∗-min decoding algorithm was presented. Complexity, effectiveness, and average number of iteration for each algorithm are shown.
Face verification problem has been successfully approached in many recent papers [12, 11, 9, 8]. Unfortunately, recognition quality decreases significantly in the presence of different degradation factors inherent to real world data [5], for example surveillance videos, where faces can be very small in size, blurred and affected with compression artifacts.(More)
Благодаря перечисленным выше особенностям наборы пространственных данных достаточно легко переносятся из традиционного гипертекстового Веб в Семантический Веб. На практике это означает простоту включения геопространственных наборов данных в проект Linking Open Data [3], целью которого является наполнение сети Интернет данными в стандартных форматах Semantic(More)
  • 1