Using AI for Personalized English Writing Feedback on Educational Platforms

Over the last few years, AI has had appreciable advances in enhancing various fields, education is also on this list. The skill of writing is one of the most significant ones that are not limited to a certain discipline and used in every academic year and all kinds of professions. But the feedback on the writing process is one of the most complex tasks to accomplish, especially in large-scale educational processes where the teacher can have a lack of time and tools. This paper aims at investigating the use of AI to give real-time feedback on English compositions on an online learning environment using deep learning techniques, specifically BERT-LSTM models. Writing feedback provided by traditional methods cannot be efficient and effective because the ability and time of the instructors differ, causing fluctuating quality feedback from tutor to tutor. These limitations are resolved in this research by developing a system using BERT integrated with LSTM networks for the analysis of the refined aspects of English writing. This is accompanied by LSTM networks that are widely used when it comes to modeling sequential information and error trapping of context over different sentences. The proposed system is aimed at presenting the users with individual feedbacks based on the learning styles and writing behaviors of the users and so the educational value would be improved. The proposed framework is carried through Python software. This model is highly accurate with 96.2% accuracy when compared to the other deep learning models such as DT, RF, GBC AND SVM. This approach is set to ensure that the gap between writing assessments where the end product is an automated report and standardized feedback from a teacher is closed and therefore make high quality feedback on writing more standardized and scalable in current and future learning platforms.