Enhancing Education Efficiency with Automated Grading

Client Background

Kyoikudojinsha Co., Ltd is a Japanese publisher specializing in teaching materials for elementary and middle school students. Guided by its vision of creating a future where every child can maximize their diverse potential, collaboratively solve challenges, and contribute to a happy society, Kyoikudojinsha is committed to developing high-quality, tailored educational materials that meet diverse needs.

Challenge

Paper-based homework and tests remain predominant in Japan’s education system, particularly in elementary schools. Grading these materials manually requires significant time and effort, adding to the already heavy workloads of teachers. Digitalizing and automating these tasks is essential for improving efficiency and enhancing the quality of education.

To address this challenge and contribute to the digitalization of education, the client sought to develop a tool capable of automatically evaluating students’ handwritten math answers in their workbooks. Kyoikudojinsha selected Recursive as its development partner due to our expertise in custom AI development and our shared vision of creating a better future for generations to come.

Recursive's Solution

Recursive developed an advanced AI Grading Assistant that automatically evaluates elementary school students’ handwritten math answers. The assistant supports core mathematical functions, including addition, subtraction, multiplication, and division with both whole numbers and decimals, as well as linear (horizontal) and column (vertical) calculations.

Students can upload photographs of homework and tests directly from Kyoikudojinsha's math workbooks through a web app accessible on PCs, tablets, and smartphones. The assistant intelligently interprets the handwritten input, converting it into a machine-readable format for analysis, computes correct solutions, and provides instant score results.

High Accuracy and Compliance

The assistant was fine-tuned to adhere strictly to Japanese government guidelines for scoring output. In addition to accuracy, answers must follow specific task completion formats to be considered correct, ensuring compliance with standardized educational practices.

Recursive’s engineering team, in collaboration with the client, collected and labeled the vast amount of handwritten math answers to fine-tune the model. This extensive dataset significantly enhanced its performance, resulting in highly accurate recognition of answers and precise identification of key elements within the handwritten input.

Security and Future Development

Recursive always prioritizes security and data control. The grading assistant runs entirely on our own secure GPU infrastructure, ensuring that all data remains protected and fully controlled by the client.

In the future development stage, Recursive plans to expand the AI assistant’s capabilities to cover a wider range of elementary school math problems and implement a performance aggregation system to compile scores into individual student records. This will enable personalized learning while supporting the effective use of educational data in schools.