Mathematicians from RUDN University and the Free University of Berlin proposed a new way of using neural networks for working with noisy high-dimensional data

Mathematicians from RUDN University and the Free University of Berlin proposed a new way of using neural networks for working with noisy high-dimensional data

Mathematicians from RUDN University and the Free University of Berlin have proposed a new approach to studying the probability distributions of observed data using artificial neural networks. The new approach works better with so-called outliers, i.e. input data objects that deviate significantly from the overall sample.

The restoration of the probability distribution of observed data by artificial neural networks is the most important part of machine learning. The probability distribution not only allows us to predict the behaviour of the system under study, but also to quantify the uncertainty with which forecasts are made. The main difficulty is that, as a rule, only the data are observed, but their exact probability distributions are not available. To solve this problem, Bayesian and other similar approximate methods are used. But their use increases the complexity of a neural network and therefore makes its training more complicated.

RUDN University and the Free University of Berlin mathematicians used deterministic weights in neural networks, which would help overcome the limitations of Bayesian methods. They developed a formula that allows one to correctly estimate the variance of the distribution of observed data. The proposed model was tested on different data: synthetic and real; on data containing outliers and on data from which the outliers were removed. The new method allows restoration of probability distributions with accuracy previously unachievable.

The mathematicians of RUDN University and the Free University of Berlin used deterministic weights for neural networks and used the networks outputs to encode the distribution of latent variables for the desired marginal distribution. An analysis of the training dynamics of such networks allowed them to obtain a formula that correctly estimates the variance of observed data, despite the presence of outliers in the data. The proposed model was tested on different data: synthetic and real. The new method allows restoring probability distributions with higher accuracy compared with other modern methods. Accuracy was assessed using the AUC method (area under the curve is the area under the graph that allows making assessment of the mean square error of the predictions depending on the sample size estimated by the network as “reliable”; the higher the AUC score, the better the predictions).

The article was published in the journal Artificial Intelligence.

Scientific Conferences View all
03 Nov 2017
RUDN University organized the first 5G Summit R&D Russia on June 19 - 20, 2017
Main Publications View all
15 Nov 2017
RUDN University scientists publish results of their scientific researches in highly-recognized in whole world and indexed in international databases journals (Web of Science, Scopus ect.). That, of course, corresponds to the high status of the University and its international recognition. Publications of June-September 2017 ( In Journals of categories Q1-Q3)
Similar newsletter View all
30 Dec
Biologists from RUDN University discovered the secret of flaxseed oil with long shelf life

Biologists from RUDN University working together with their colleagues from the Institute of Molecular Biology of the Russian Academy of Sciences and the Institute of Flax studied the genes that determine the fatty acid composition in flaxseed oil and identified polymorphisms in six of them. The team also found out what gene variations could extend the shelf life of flaxseed oil. This data can be used to improve the genetic selection of new flax breeds. The results were published in the BMC Plant Biology journal.

19 Apr
Dentists from RUDN University Presented a New Classification of Root Canal Shape Changes

Individual characteristics of the shape and cross-section of the root canal are one of the main issues for dentists. When treating a root canal, a doctor needs to properly clean it, fill it, and carry out a rebuilding procedure so that a canal is sealed. The first stage of endodontic treatment requires detailed knowledge of root canal anatomy. A team of dentists from RUDN University studied and classified various changes in root canal shapes. The new classification will help doctors avoid diagnostic errors, better select their tools, and treat patients more efficiently.

19 Apr
A chemist from RUDN developed a green catalyst for pharmaceutical and industrial chemistr

Many production facilities (e.g. plastic manufacturers, pharma companies, and others) use nanocatalysts that contain palladium—an expensive component that is not sustainably produced. A chemist from RUDN University found a way to reduce palladium consumption and to make its manufacture more eco-friendly. He developed a catalyst based on a substance that comes from plant waste. Using his invention, manufacturers could cut palladium consumption in half. Moreover, new catalysts can be reused multiple times without any decrease in efficiency.

Similar newsletter View all