Аннотация:Multidimensional scaling problems arise when it is necessary to reduce the dimension of vector representations of objects or to implement a matrix of proximity of a set of objects in vector space. From a practical point of view, this may be necessary for visualization and further analysis, for a more compact representation and less storage space, or for a faster calculation, for example, a dot product. The paper shows how the GNMDS problem can be effectively solved using a neural network of one fully connected layer.