Automated assessment and feedback systems
Summary of the project
The "Automated assessment and feedback system for AI-related programming challenges" project has been developed to improve the assessment of student assignments in the course Machine Learning for Signal Processing (5LSL0) at TU/e. In this project an AI-based system is created that automatically evaluates programming assignments and provides both quantitative grades and qualitative feedback. The method uses natural language processing techniques tailored for programming languages to assess the correctness of submitted codes and offer detailed, specific feedback. The development process involves iterative cycles of creating and refining the AI method, based on existing datasets and augmented with synthetic data to address potential data insufficiencies. The project team, consisting of a Postdoc and supported by the applicants, will collaborate closely with a student sounding board to ensure the feedback system meets the educational needs and expectations.
Aim of the project
The primary aim of this project is to enhance the learning experience for students by providing rapid, detailed, and specific feedback on their programming assignments, thus enabling them to identify and address knowledge gaps more effectively. By automating the assessment process, the project seeks to reduce the workload on instructors, allowing them to focus on more targeted educational activities and potentially adopt a flipped classroom approach. Furthermore, the project aspires to make the assignments more challenging and relevant to real-world problems by removing the constraints of manual assessment, ultimately better preparing students for industry demands in machine learning and signal processing. The developed system has also been intended to be generalizable to other courses, promoting broader educational innovation within and beyond TU/e.
Results and learnings
For more information, please contact:
