🚀💳 Trustee Plus - більше ніж криптогаманець з європейською платіжною карткою. Спробуй 👉
Наталя ХандусенкоStartup
7 February 2025, 11:45
2025-02-07
A free AI application for learning English has been launched in Ukraine: GrammarTrack will help you master grammar at any level
Senior Software Engineer Artem Petrusenko from Lviv developed the web application GrammarTrack. dev.ua told about the history of its creation earlier. The application successfully passed beta testing and was recently released in full. Now Ukrainians have the opportunity to study English grammar of all levels of complexity for free, namely: assess their current level of knowledge, set short-term goals and track progress, as well as instantly check tasks with detailed explanations. In addition, AI will help identify weaknesses in the user’s learning and will provide daily recommendations on selected topics.
Senior Software Engineer Artem Petrusenko from Lviv developed the web application GrammarTrack. dev.ua told about the history of its creation earlier. The application successfully passed beta testing and was recently released in full. Now Ukrainians have the opportunity to study English grammar of all levels of complexity for free, namely: assess their current level of knowledge, set short-term goals and track progress, as well as instantly check tasks with detailed explanations. In addition, AI will help identify weaknesses in the user’s learning and will provide daily recommendations on selected topics.
GrammarTrack is a free web application with different levels of difficulty for learning English from beginner to advanced and covers various grammar topics.
On the portal, users will find: 50 different grammar topics, 6,500 practice tasks, 4 levels of difficulty for gradual progress, and 6 thematic courses for different needs.
GrammarTrack was beta-tested throughout October 2024, with over 500 participants joining. After receiving feedback, the team began working on improving the app, which became publicly available on February 1.
«Beta testing turned out to be extremely productive for the development of the platform. Based on the feedback we received, we significantly transformed GrammarTrack, updating about 80% of the initial functionality,» Artem Petrusenko told dev.ua.
Key improvements include:
introduction of a system of short-term learning goals, which allows users to plan their progress more effectively;
development of comprehensive courses for all levels of language proficiency;
creating a knowledge assessment module to accurately determine the user’s current level;
expansion of response formats: in addition to traditional text entry, the ability to construct responses from ready-made blocks has been added.
A particularly important innovation was the expanded feedback system: now, when making a mistake, the user receives not only an indication of the incorrect answer, but also a detailed explanation of the error with a reference to the corresponding grammatical rule.
GrammarTrack has implemented a machine learning system to create a personalized learning experience.
The AI functionality offers three unique approaches to learning grammar:
Learn New Topics — focuses on mastering new material. The user can choose up to 5 new grammar topics and study them at a comfortable pace within the framework of the set goal.
Review Your Weak Areas — uses intelligent analysis to identify problem areas. The system analyzes the user’s error history and the time elapsed since the last repetition of each topic. Based on this data, a personalized rating of topics that require repetition is formed, which allows you to effectively eliminate knowledge gaps.
Learn & Review Mix — combines learning new material with systematic review. The user selects 2 new topics to study, and the system automatically adds 3 topics for review every day. Thanks to a machine learning algorithm, the set of topics for review is updated daily, ensuring optimal assimilation of the material.
«In our application, we use TensorFlow.js to implement a neural network that recommends topics for repetition. The model has three layers: input (5 nodes, ReLU), hidden (5 nodes, ReLU), and output (1 node, sigmoid). At the input, the model receives 5 features: error rate, time since last practice, average response time, etc. All data is normalized to [0,1]. The model is trained on historical user data with 100 epochs and gives a recommendation score from 0 to 1 for each topic,» Artem told about the technical side of introducing AI into the platform, which, by the way, was not in the first version of the platform.
We remind you that GrammarTrack is being created by a team at its own expense, which includes developer Artem Petrusenko, English teachers Stanislava Jung and Yaroslav Gerasimov, as well as project manager Darina Zheltikova.
«We continue to work with the core team that was formed at the beginning of the project. Such a compact team structure allows us to maintain flexibility in decision-making and effectively coordinate work on the project. We also continue to develop GrammarTrack using our own resources, which allows us to maintain complete independence in decision-making and focus on product quality. At this stage, this approach fully corresponds to our vision of project development,» explains Artem.
The plans include launching a mobile version, which is almost ready, adding other foreign languages, and creating a toolkit for teachers.
The team plans to launch a mobile app, which is already 75% complete. This extension will allow users to access effective grammar learning anywhere, anytime.
If we talk in terms of 2025, this year they plan to create a specialized toolkit for English teachers.
«We aim to make GrammarTrack not only a platform for self-study, but also a powerful assistant for educators, helping them teach grammar more effectively,» the developer shares.
In addition, they are going to expand the language coverage of the platform.
«We plan to adapt our successful teaching methodology for other languages, developing the same high-quality and structured training courses for each of them as we did for English. This will include creating specialized grammar materials, exercises, and an assessment system for each new language,» Artem added.
GrammarTrack is a completely free web application.
At this stage, GrammarTrack remains completely free for all users. The team’s priority is to create the most useful and effective language learning tool possible, so they are currently focused on two key tasks: attracting an active user base and constantly improving the platform based on real user experience.
«In the future, we plan to introduce a paid subscription system, but this will happen only after we are convinced that our product fully meets the needs of users and provides real value. At the same time, we intend to keep access to the basic functionality free so that the platform remains accessible to a wide range of users,» says the developer.
Як нейромережі бачать вільну та незалежну Україну? Тест dev.ua
Нейронні мережі для генерації зображень бачать світ по-своєму, їхню логіку зрозуміти часом зовсім неможливо. Але таки хочеться. На честь Дня Незалежності України редакція dev.ua вирішила провести невеликий експеримент.
Ми задали чотирьом різним нейронним мережам п’ять однакових запитів: «прапор України», «День Незалежності України», «український Крим», «перемога України» та «українці». Отриманими результатами ми ділимося з вами нижче.
У TikTok тепер можна генерувати фон за допомогою нейромережі. Ми протестували її та ділимося результатами
У TikTok з’явилася нова функція «Розумний фон». З її допомогою як фон для тіктоків можна підставляти згенеровані нейромережею зображення. Редакція dev.ua протестувала цю технологію і ділиться своїми враженнями.
