3
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.

Student's Scientific Initiatives View all
03 Nov 2017
June 22 - 26, 2017 in Barnaul, Altai State University, took place the Summer Academy of the BRICS Youth Assembly, an international event that brought together representatives of different countries
1051
Scientific Conferences View all
03 Nov 2017
RUDN University organized the first 5G Summit R&D Russia on June 19 - 20, 2017
1394
Similar newsletter View all
31 Mar
RUDN University awards for specific areas of science and technology based on the results of 2021

Every year, RUDN University selects the best of the best in the field of science and innovation and encourages with a special reward. Since 2009, the Academic Council of the University has been awarding one reward in natural and technical sciences and the other one in social and humanitarian sciences. Both individual researchers and groups of authors can become laureates.

67
31 Mar
International Day of Women and Girls in Science: women scientists of the RUDN talk about their path to science

“Science is the basis of all progress that facilitates the life of mankind and reduces its suffering,” — Marie Sklodowska—Curie. A symbol of a woman’s success in science. The first scientist in the world — twice winner of the Nobel Prize.

309
31 Mar
RUDN University Mathematicians Create a Model for Queue Organizing with Self-Sustained Servers

RUDN University mathematicians proposed a model for optimizing the operation of queuing systems (from computer networks to stores). Unlike analogues, the servers in it are self-sustained. They can determine when to start and stop working themselves. Such a model can be useful, for example, for online taxi services and other systems where workers choose their own operating hours.

76
Similar newsletter View all