並行AIモデルによる手書き英作文の文字認識自動採点システムの開発

由于无法获得预期的入学考试数据，大分大学从零开始创建了定制的写作提示，并对一年级的学生进行了管理。到目前为止，已经收集了大约250份学生作文回复并手动评分。已经实现了对这种类型的写作回答进行自动评分的模型。到目前为止，自动评分的结果好坏参半，但测试仍在进行中。手写识别的初步模型已经运行，但需要更合适的数据。具体地说，需要更接近实际入学考试提交的笔迹样本。

Lacking access to the expected entrance exam data, customized writing prompts were created from scratch and administered to groups of 1st year students at Oita University. Approximately 250 student writing responses have been collected and graded manually to date. Models for automatically scoring this type of writing response have been implemented. The auto-grading results are so far mixed, but testing is ongoing. A preliminary model for handwriting recognition has been run but requires more appropriate data. Specifically, handwriting samples closer to those which would be submitted for actual entrance exams are needed.