Перспективы интеллектуальной обработки больших объемов астрономических данных с помощью нейронных сетей
Companies are increasingly paying close attention to the IP portfolio, which is a key competitive advantage, so patents and patent applications, as well as analysis and identification of future trends, become one of the important and strategic components of a business strategy. We argue that the problems of identifying and predicting trends or entities, as well as the search for technical features, can be solved with the help of easily accessible Big Data technologies, machine learning and predictive analytics, thereby offering an effective plan for development and progress. The purpose of this study is twofold, the first is an identification of technological trends, the second is an identification of application areas and/or that are most promising in terms of technology development and investment. The research was based on methods of clustering, processing of large text files and search queries in patent databases. The suggested approach is considered on the basis of experimental data in the field of moving connected UAVs and passive acoustic ecology control.
We demonstrate that classical quadratic forms are not able to solve the problem of recognizing highdimensional images. The "deep" GalushkinHinton neural networks can solve the problem of highdimensional image recognition, but their training has exponential computational complexity. It is technically impossible to train and retrain a "deep" neural network rapidly. For mobile "artificial nose" systems we proposed to employ a number of "wide" neural networks trained in accordance with (GOST R 52633.52011). This standardized learning algorithm has a linear computational complexity, i.e. for each new smell image a time of about 0.3 seconds is sufficient for creating and training a new neural network with 2024 inputs and 256 outputs. This leads to the possibility of the rapid training of the artificial intelligence "artificial nose" and a gradual expansion of its database consisting of 10 000 or more trained artificial neural networks.