Within the framework of the Up Great Pro//READING technology competition, the first cycle of tests was completed and the winners of the “Grammar” and “Grammar.Eng” nominations were announced.
The Up Great Pro//READING contest is aimed at creating an AI assistant capable of helping teachers identify logical, factual, stylistic and other semantic errors in the writings of schoolchildren in the Russian language, literature, history, social science and English. The aim of the competition is to stimulate the development of strong artificial intelligence technologies and support the development community in the field of machine learning.
More than 180 development teams took part in the first round of the competition. According to the results of the qualification, 11 teams were admitted to the tests and showed effective solutions. As part of the tests, AI assistants checked 500 essays and essays of the unified state exam in Russian and the same number in English. The automated platform compared the texts verified by AI assistants and real teachers and identified the best quality solutions.
The winners of the special categories of “Grammar” and “Grammar. Eng”, whose AI systems correctly identified the largest number of grammatical and speech errors, divided the prize Fund of 20 million rubles.
Results of the nomination “Grammar” for texts in Russian:
* In accordance with the rules of the competition, one team can claim the prize Fund only in one of the categories. The Antiplagiat team took the 2nd place in both the “Grammar” and “Grammar.ENG” categories. According to the decision of the team, the cash prize is awarded in the category “Grammar.ENG”.
Results of the nomination “Grammar.ENG” for English texts:
As part of the main task of the competition, no team has yet achieved a result corresponding to the level of teachers. Participants will be able to compete again for the prize Fund of the competition in the amount of 200 million rubles at the second cycle trials planned for the spring of 2021. In the course of the analysis of texts in Russian, the best intermediate result was shown by the team “Antiplagiat” (accuracy level — 57.8%). For the English language, the best quality solution was demonstrated by the NeuroChtenie team (accuracy level — 84.5%). Among the leaders was the team from St. Petersburg Seedh AI3, as well as the teams “Rocket” and “Nanosemantika”. At this stage, the participants ' AI assistants most successfully coped with identifying lexical, grammatical and speech errors. The greatest difficulties arise when determining logical errors.
Participants ' solutions are planned for pilot implementation and operation in 2021-22 in a number of schools in different regions of Russia.
The Up Great Pro//READING technology competition was organized by the SKOLKOVO Foundation and ASI WITHIN the framework of the National technology initiative, with the support of the ROSNANO School League and FEFU. The technical partner of the competition is the NTI competence Center in the area of “Artificial intelligence” on the basis of MIPT.
Up Great technology competitions were initiated by RVC, SKOLKOVO Foundation and ASI for Russian innovative companies as part of the National technology initiative. Participants are invited to find solutions to the most complex technological problems that the world has no solutions yet. The winning team, which has overcome the global technological barrier, receives a large cash prize for demonstrating a clear and repeatable solution to the competition task before experts, the jury and the General public. The mechanics of the competitions are developed in accordance with the best practices of the world technological competitions: XPrize, Darpa Grand Challenge, etc. The operator of up Great Technology competitions is RVC.