National Research University – Higher School of Economics

About: National Research University – Higher School of Economics is a education organization based out in Moscow, Russia. It is known for research contribution in the topics: Population & Politics. The organization has 12873 authors who have published 23376 publications receiving 256396 citations.

Investigation and development of the intelligent voice assistant for the Internet of Things using machine learning

Abstract: Artificial intelligence technologies are beginning to be actively used in human life, this is facilitated by the appearance and wide dissemination of the Internet of Things (IoT). Autonomous devices are becoming smarter in their way to interact with both a human and themselves. New capacities lead to creation of various systems for integration of smart things into Social Networks of the Internet of Things. One of the relevant trends in artificial intelligence is the technology of recognizing the natural language of a human. New insights in this topic can lead to new means of natural human-machine interaction, in which the machine would learn how to understand human's language, adjusting and interacting in it. One of such tools is voice assistant, which can be integrated into many other intelligent systems. In this paper, the principles of the functioning of voice assistants are described, its main shortcomings and limitations are given. The method of creating a local voice assistant without using cloud services is described, which allows to significantly expand the applicability of such devices in the future.

Concept-Based Biclustering for Internet Advertisement

Abstract: The problem of detecting terms that can be interesting to the advertiser is considered. If a company has already bought some advertising terms which describe certain services, it is reasonable to find out the terms bought by competing companies. On the other hand, the company, that provides context advertisement, wants to discover prospective markets, the advertisers. It can be done by means of so called biclustering. For binary relation firms terms the most natural bicluster definition is a tuple of two subsets of firms and terms respectively, where each firm from the first component buys each term from the second one. To solve this task there is a well-developed notion of formal concept which has almost equivalent definition to such a bicluster in terms of object-attribute tables in Formal Concept Analysis. However, the number of formal concepts (biclusters) for a given dataset can be of exponential size in the worst case. To avoid this difficulty we proposed a new concept-based biclustering method. The new bicluster definition, (dense) object-attribute bicluster or simply oa-bicluster, is a relaxation of formal concept notion. Our findings shows that the number of (dense) oa-biclusters is no greater than the number of non-empty cells of initial binary relation. The paper contains experimental results on applying the proposed algorithm to contextual Internet advertisement data in comparison with some FCA algorithms and additional results on so-called morphological metarules for term recommendation task on the same data.

Decoding Movement From Electrocorticographic Activity: A Review.

Abstract: Electrocorticography (ECoG) holds promise to provide efficient neuroprosthetic solutions for people suffering from neurological disabilities. This recording technique combines adequate temporal and spatial resolution with the lower risks of medical complications compared to the other invasive methods. ECoG is routinely used in clinical practice for preoperative cortical mapping in epileptic patients. During the last two decades, research utilizing ECoG has considerably grown, including the paradigms where behaviorally relevant information is extracted from ECoG activity with decoding algorithms of different complexity. Several research groups have advanced toward the development of assistive devices driven by brain-computer interfaces (BCIs) that decode motor commands from multichannel ECoG recordings. Here we review the evolution of this field and its recent tendencies, and discuss the potential areas for future development.

Two-step classification method based on genetic algorithm for bankruptcy forecasting

Abstract: By present, many models of bankruptcy forecasting have been developed, but this area remains a field of research activity; little is known about the practical application of existing models. In our opinion, this is because the use of existing models is limited by the conditions in which they are developed. Another question concerns the factors that can be significant for forecasting. Many authors suggest that indicators of the external environment, corporate governance as well as firm size contain important information; on the other hand, the large number of factors does not necessary increase predictive ability of a model. In this paper, we suggest the genetic algorithm based two-step classification method (TSCM) that allows both selecting the relevant factors and adapting the model itself to application. Classifiers of various models are trained at the first step and combined into the voting ensemble at the second step. The combination of random sampling and feature selection techniques were used to ensure the necessary diversity level of classifiers at the first step. The genetic algorithms are applied at the step of features selection and then at the step of weights determination in ensemble. The characteristics of the proposed method have been tested on the balanced set of data. It included 912 observations of Russian companies (456 bankrupts and 456 successful) and 55 features (financial ratios and macro/micro business environment factors). The proposed method has shown the best accuracy (0.934) value among tested models. It has also shown the most balanced precision-recall ratio. It found bankrupts (recall = 0.953) and not bankrupts (precision = 0.910) rather accurately than other tested models. The ability of method to select the task-relevant features has been also tested. Excluding the features that are significant for less than 50% of the classifiers in the ensemble improved the all performance metrics (accuracy = 0.951, precision = 0.932, recall = 0.965). So, the proposed method allows to improve the advantages and alleviate the weaknesses inherent in ordinary classifiers, enabling the business decisions support with a higher reliability.