Automatic Correction of Multimodal Experimental Reports Driven by LLM
(School of Electrical and Control Engineering, North China University of Technology, Beijing 100144, China)
Abstract
Automatic correction of experimental reports is an important task in the field of intelligent education. Following the OBE concept, the paper reproduces the teacher's correction idea, transforms the scoring items into questions, and coordinates the multimodal response information such as text, tables, and pictures to score, so as to be close to the actual teaching and curriculum construction needs. In the understanding and scoring stage of multimodal information, on the basis of Deep Learning, LLM is introduced to realize the content extraction and transformation of table questions, and solve the difficulties of positioning and logical discrimination. For text content, BERT is used to understand. For the image content, the self-training model constructed by the combination of BERT and ResNet-18 is used to scale the image matching weights for the image feature evaluation in graphic questions. The scheme uses small sample data for training, adapts to different subject experiments, and overcomes the pain points such as insufficient generalization and migration caused by relying on a large amount of data training. Through the correction test of two courses, the average accuracy of the report score reaches 92.20%, bridging the gap of automatic correction of non-customized experimental reports.
Keywords
automatic correction of experimental report; Deep Learning; LLM
References
- [1] Wu Lan'an, Yan Hanbing, Huang Faliang, et al. Application analysis and practical challenges of large-scale language models in higher education [J]. Modern Educational Technology, 2023, 33 (8): 29-37
- [2] Wang Shaoqing Design and Implementation of Automatic Correction System for Electronic Experiment Reports [J]. Journal of Shandong University of Technology: Natural Science Edition, 2010, 24 (3): 60-63
- [3] Jiang Jinwen, Zhang Weifeng Research and Implementation of Automatic Homework Correction Based on Code Repository [D]. Nanjing: Nanjing University of Posts and Telecommunications, 2021
- [4] LINO A,ROCHA A,SIZO A. Virtual Teaching and Learning Environments: Automatic Evaluation with Artificial Neural Networks [J].Cluster Computing,2019,22:7217-7227.
- [5] ABD EL-HALEEM A M,EID M M,ELMESALAWY M M,et al. A Generic AI-Based Technique for Assessing Student Performance in Conducting Online Virtual and Remote Controlled Laboratories [J].IEEE Access,2022,10:128046-128065.
- [6] Wang Shenghai, Liu Yumin The Application of Deep Learning in Virtual Experiment Guidance and Correction [D]. Beijing: Beijing University of Posts and Telecommunications, 2018
- [7] Zhang Jinghui, Wang Peijin Design and Implementation of Automatic Scoring System for Course Design [J]. Journal of Electrical and Electronic Teaching, 2019, 41 (4): 149-152
- [8] Zhang Lin, Chen Xisong, Wang Xiaojun Development of Intelligent Comprehensive Evaluation System Based on Remote Open Measurement and Control Teaching Experimental Platform [D]. Nanjing: Southeast University, 2020
- [9] Sun Yalin, Wen Fu'an Application of Artificial Intelligence in Language Training Guidance and Correction [D]. Beijing: Beijing University of Posts and Telecommunications, 2018
- [10] CHEN Y F,LIU X M,HUO P P,et al. The Design and Implementation for Automatic Evaluation System of Virtual Experiment Report [C]//2017 12th International Conference on Computer Science and Education(ICCSE).Houston:IEEE,2017:717-721.
- [11] Jiang Jianhua, Wang Yang, Wang Min, et al. Internet based experimental teaching and automatic correction of experimental reports [J]. Laboratory Research and Exploration, 2017, 36 (5): 157-161
- [12] Li Dongwei, Hu Hao, Long Yun Automatic Scoring System for Metalworking Internship Based on Halcon Platform [J]. Mechanical and Electrical Engineering Technology, 2025, 54 (2): 111-115
- [13] ZENG W,GUO J,HAO L,et al. Evaluation of Physical Electrical Experiment Operation Process Based on YOLOv5 and ResNeXt Cascade Networks [J].Neural Processing Letters,2022,55(2):1583-1603.
- [14] Cao Shanshan Design and Development of Intelligent Evaluation System for Engineering Online Experimental Reports [D]. Beijing: Beijing Forestry University, 2022
Funding
Beijing Municipal Education Commission Key Research Project on Digital Education in Beijing - Support for the Construction of Multidimensional Digital Learning Spaces for Excellent Talent Cultivation (BDEC2022619001)
About the authors
Xu Jining (1970-), female, Han ethnicity, from Xingping, Shaanxi Province, associate professor, PhD, research interests: control theory and pattern recognition, smart education; Corresponding author: Huang Nan (1997-), male, Han ethnicity, Beijing native, currently pursuing a master's degree, with research interests in deep learning, natural language processing, and smart education; Song Hao (1977-), male, Han ethnicity, from Langfang, Hebei, is an engineer with a bachelor's degree. His research interests include edge intelligence and intelligent manufacturing control systems, as well as industrial control networks.