AI-Based Mobile Paper Grading: Trends and Challenges

Many universities, schools, and other educational institutions embraced distance learning due to the Covid-19 Pandemic in previous years. Because some instructors are unfamiliar with it, creating exams for students using online platforms is not always simple. Therefore, it takes a lot of time and effort to create online assessments and manually correct them. Additionally, grading many tests can burden instructors, especially if they are unfamiliar with online technology. Exams can be made as multiple-choice questions, which can help instructors overcome human error in judging exam papers, as a temporary fix for this rather than creating lengthy questions. This, however, is not a long-term fix because instructors still have to spend time grading the multiple-choice questions. Artificial intelligence can therefore be quite useful in this case. This paper offers an overview of notable implementations that autonomously grade and compute exam paper outcomes using vision-based and Artificial Intelligence-based mobile approaches. Additionally, different techniques offer a variety of methods, each with its benefits and limitations. In this paper, each strategy's methodologies, as well as its results performance, are discussed. Also, it discusses which approach is more appropriate for an automated Grading system. Based on the results reviewed from different sources, it was found that Artificial Intelligence-based approaches and vision-based approaches are accurate in detecting the objectives of this project. However, the AI-based approach has more efficiency due to the combination of various models, resulting in higher accuracy. Finally, both approaches' advantages and limitations are described throughout the report.