Multi-layered knowledge validation system for academic integrity

Abstract

The present invention relates to an AI-supported assessment system for instructors that verifies students' understanding beyond purely text-based submissions. By integrating with learning management systems (LMS), the system extracts student assignments, generates personalized, cognitive-level-related questions, and enables an oral validation process via a real-time assessment interface. The system ensures an adaptive knowledge validation process by applying hierarchical cognitive models, structuring the distribution of questions according to the difficulty level of the respective concepts, and dynamically adjusting the assessment methods based on parameters defined by the instructor. The oral examination module operates via a digital interface through which students must answer examination questions orally.The answers are transcribed, processed, and evaluated against an expected answer generated by a language model. A semantic similarity assessment procedure determines the conceptual correspondence between the student's answer and the reference answer, generating a confidence-based comprehension score that is weighted according to the levels of cognitive complexity defined by the instructor. A final comprehension report is provided to the instructors, offering insight into the student's level of knowledge without prescribing any grading decisions.