diagnose-student-school-work

To generate a marking report from scanned student work without an answer key, provide the scanned images to infer correct responses and scores directly from on-paper teacher annotations, red marks, and printed rubrics. The system outputs JSON-first artifacts and derived Markdown learning reports.

Yes, you can reconstruct grading scores from teacher-annotated quizzes in different languages. The marking reconstruction process is subject-agnostic and works across any language for weighted assessments, class quizzes, or term tests returned as scans.

Grounded diagnosis for learning reports is the process of inferring correct answers and scores from teacher red marks, green corrections, and printed rubrics on the same page. It builds a canonical marking report by reconstructing grading logic directly from on-paper evidence.

Yes, the marking package schema outputs both JSON-first artifacts and derived Markdown reports. This canonical artifact compatibility ensures downstream tools can render learning reports consistently without needing a separate answer key.

To extract correct answers from a scanned test with only teacher corrections, the system applies grounded marking reconstruction to infer the correct responses from the visible green corrections and rubric marks. It uses this on-paper evidence to build a consistent learning report.

Reconstructing grading reports from teacher annotations relies heavily on the clarity of on-paper evidence. If scanned work lacks visible red marks, green corrections, or printed rubrics, the system cannot accurately infer correct answers or generate a canonical marking report.