Researchers have created open-source algorithms that improve the precision of robotic devices’ recognition of mental commands based on neural networks.
Robotic systems using neural networks do not have adequate command recognition accuracy because the command classifier frequently makes mistakes when identifying commands that are part of the device dictionary. To improve the accuracy of recognizing brain electrical patterns in robotic devices, mathematician and physician Alexandra Bernadotte, an associate professor in the department of information technologies and computer sciences at MISIS University, has developed an open-source algorithm. This algorithm optimizes the selection of a dictionary for the classification of mental or other commands from a specific dictionary. This can be used to send data accurately even in overcrowded communication channels.
“Existing algorithms, usually help increase the accuracy of the classification of the already created dictionary. I aimed to optimize the process of selecting dictionary commands. The Maximin algorithm is effective when the dictionary is large enough and you want words to be recognized equally well. The Maximal algorithm is used if we need to increase the accuracy of recognition, and there are significantly more resources for selecting a dictionary.” explained Alexandra Bernadotte, Ph.D., mathematician and doctor, associate professor of the Department of Information Technologies and Computer Sciences at MISIS University, an employee of the Faculty of Mechanics and Mathematics at Lomonosov Moscow State University.
In contrast to the brute-force algorithm, the Maximin and Maximal algorithms put forth by Alexandra Bernadotte in her work enable the selection of a set of dictionary objects to maximize classification accuracy while drastically reducing the time required to choose a dictionary of commands. This is achieved by implementing a k-partite graph.
“The idea of solving the problem is to represent a set of objects (encoded as a sequence of symbols or visual sequences) in the form of a k-partite graph, where each partite of the graph corresponds to a group of objects with a certain common feature (equivalence class). The edges of the graph have weights corresponding to the value of the Kolmogorov-Smirnov criterion obtained on the distributions of n-dimensional signals of mental commands belonging to different classes of equivalence—to different parties,” said Alexandra Bernadotte
“Finding a clique with the Maximal or Maximin total weight in this graph gives the desired dictionary of commands which are classified with an accuracy not lower than the specified one. This algorithm can be used for different tasks in the field of information transmission,” said Alexandra Bernadotte
Click here for the Published Research Paper
