ACADEMIC ASSESSMENT BASED ON FINE TUNED LLMS AND NEURAL FEATURE BASED DIAGRAM EVALUATION

Sachin Jain; Rupa Rani; Mukulit Goel; Anuj Kumar

doi:10.29121/shodhkosh.v7.i12s.2026.8237

Authors

Sachin Jain Associate Professor, Department of Computer Science and Engineering, Ajay Kumar Garg Engineering College, Ghaziabad, Uttar Pradesh, India
Rupa Rani Assistant Professor, Department of Computer Science and Engineering, Ajay Kumar Garg Engineering College, Ghaziabad, Uttar Pradesh, India
Mukulit Goel Assistant Professor, Department of MCA, Ajay Kumar Garg Engineering College, Ghaziabad, Uttar Pradesh, India
Anuj Kumar Associate Professor, Department of Computer Science and Engineering, JB Institute of Technology, Dehradun, Uttarakhand, India

DOI:

https://doi.org/10.29121/shodhkosh.v7.i12s.2026.8237

Keywords:

Automated Academic Assessment, Large Language Models (LLMs), Neural Network, Finetuning, Optical Character Recognition, Diagram Detection and Matching, Prompt Tuning

Abstract [English]

The process of evaluation traditionally carried out on the answer sheets is laborious, inconsistent, and error prone. Automation of this process will significantly improve both speed and accuracy. Compared to the existing automated answer sheet evaluation solutions; while exploring subjective and essay-type answers, they did not capture the deep semantics of the language. Effective analysis of Diagrams remains neglected. This paper proposes a multimodal approach for answer sheet evaluation, which integrates both textual and visual elements. The contribution of the proposed approach is twofold: firstly, the evaluation of textual responses is improved by state-of-the-art natural language processing and fine-tuned large language model Llama 3.2; secondly, diagram evaluation has been enhanced with a neural network-based feature matcher, LightGlue, further complemented by a custom image preprocessing pipeline, integration of OCR, and NLP metrics to improve diagram feature evaluation accuracy and thus allow for the precise extraction and analysis of diagram labels. Experimental results reveal that our system achieves very good accuracy and consistency comparable to those of human evaluators. However, system performance may degrade due to digitization quality, such as poor handwriting or an unclear image. In conclusion, the proposed system overcomes the existing gaps in automated evaluation methods and hence provides a holistic solution to assess answer sheets.

References

Al Mahmud, T., Hussain, M. G., Kabir, S., Ahmad, H., & Sobhan, M. (2020). A Keyword Based Technique to Evaluate Broad Question Answer Script. Proceedings of the 2020 9th International Conference on Software and Computer Applications, 167–171. https://doi.org/10.1145/3384544.3384604 DOI: https://doi.org/10.1145/3384544.3384604

Azubogu, K., Asogwa, E. C., Ezeugbor, I. C., Okwuchukwu Ejike, C., & Onyeizu, M. N. (2024). Development of Natural Language Processing-Based Descriptive Answer Evaluation Platform (Gradescriptive). Engineering and Technology Journal, 09(08). https://doi.org/10.47191/etj/v9i08.47 DOI: https://doi.org/10.47191/etj/v9i08.47

Balat, H. F., El-dosuky, M. A., El-Razek, E.-S. M. A., & Rashed, M. Z. (2020). Automatic Exam Evaluation based on Brain Computer Interface. International Journal of Computer Applications, 175(25), 15–21. https://doi.org/10.5120/ijca2020920792 DOI: https://doi.org/10.5120/ijca2020920792

Bashir, M. F., Arshad, H., Javed, A. R., Kryvinska, N., & Band, S. S. (2021). Subjective Answers Evaluation Using Machine Learning and Natural Language Processing. IEEE Access, 9, 158972–158983. https://doi.org/10.1109/ACCESS.2021.3130902 DOI: https://doi.org/10.1109/ACCESS.2021.3130902

Bhonsle, V., Sapkal, P., Mukadam, D., & Raut, V. (2019). An Adaptive Approach for Subjective Answer Evaluation. International Journal for Research and Innovation, 1(2). www.viva-technology.org/New/IJRI

Brown, M., & Program, E. (2017). Automated Grading of Handwritten Numerical Answers.

Chandrapati, L. M., & Rao, Ch. K. (2024). Descriptive Answers Evaluation Using Natural Language Processing Approaches. IEEE Access, 12, 87333–87347. https://doi.org/10.1109/ACCESS.2024.3417706 DOI: https://doi.org/10.1109/ACCESS.2024.3417706

G, J., & G, C. S. (2020). Online Subjective answer verifying system Using Artificial Intelligence. 2020 Fourth International Conference on I-SMAC (IoT in Social, Mobile, Analytics and Cloud) (I-SMAC), 1023–1027. https://doi.org/10.1109/I-SMAC49090.2020.9243601 DOI: https://doi.org/10.1109/I-SMAC49090.2020.9243601

Kulkarni, M., Adhav, G., Wadile, K., Chavan, R., & Deshmukh, V. (2024). Digital Handwritten Answer Sheet Evaluation System. https://doi.org/10.21203/rs.3.rs-3978232/v1 DOI: https://doi.org/10.21203/rs.3.rs-3978232/v1

Lindenberger, P., Sarlin, P.-E., & Pollefeys, M. (2023). LightGlue: Local Feature Matching at Light Speed. http://arxiv.org/abs/2306.13643 DOI: https://doi.org/10.1109/ICCV51070.2023.01616

Meenakshi, A. T., Pradeep, B. M., & Vishaka, M. (2022). Web app for quick evaluation of subjective answers using natural language processing. Scientific and Technical Journal of Information Technologies, Mechanics and Optics, 22(3), 594–599. https://doi.org/10.17586/2226-1494-2022-22-3-594-599 DOI: https://doi.org/10.17586/2226-1494-2022-22-3-594-599

Nayudu, P. P., & Rani, I. R. (n.d.). Efficient Online Exam Grading with AI Powered Answer Verification. Retrieved www.joics.org

Nikhil, D. (2024). Question Paper Checking Using Generative Ai. International Journal for Research in Applied Science and Engineering Technology, 12(5), 4362–4368. https://doi.org/10.22214/ijraset.2024.62488 DOI: https://doi.org/10.22214/ijraset.2024.62488

Rahman, M. M., & Akter, F. (n.d.). An Automated Approach for Answer Script Evaluation Using Natural Language Processing. Retrieved www.ijcset.net

Raina, S., Amin, H., Sanghvi, S., Bharti, S. K., & Gupta, R. K. (2023). Automatic Subjective Answer Evaluator Using BERT Model (pp. 531–538). https://doi.org/10.1007/978-981-99-3315-0_40 DOI: https://doi.org/10.1007/978-981-99-3315-0_40

Rowtula, V., Oota, S. R., & C.V, J. (2019). Towards Automated Evaluation of Handwritten Assessments. 2019 International Conference on Document Analysis and Recognition (ICDAR), 426–433. https://doi.org/10.1109/ICDAR.2019.00075 DOI: https://doi.org/10.1109/ICDAR.2019.00075

S, P. M., Chavan, S. M., Bathula, R., Saikumar, S., & Dayalan, G. (n.d.). International Journal of INTELLIGENT SYSTEMS AND APPLICATIONS IN ENGINEERING Eval-Automatic Evaluation of Answer Scripts using Deep Learning and Natural Language Processing. In Original Research Paper International Journal of Intelligent Systems and Applications in Engineering IJISAE (Vol. 2023, Number 1). Retrieved www.ijisae.org

Saharan, R., Chauhan, R. K., Singh, S., & Sharma, P. (n.d.). AUTOMATED CONTENT GRADING USING MACHINE LEARNING Theoretical Content Grading from Exam Papers.

Salim, H. R., De, C., Pratamaputra, N. D., & Suhartono, D. (2022). Indonesian automatic short answer grading system. Bulletin of Electrical Engineering and Informatics, 11(3), 1586–1603. https://doi.org/10.11591/eei.v11i3.3531 DOI: https://doi.org/10.11591/eei.v11i3.3531

Sanuvala, G., & Fatima, S. S. (2021). A Study of Automated Evaluation of Student’s Examination Paper using Machine Learning Techniques. 2021 International Conference on Computing, Communication, and Intelligent Systems (ICCCIS), 1049–1054. https://doi.org/10.1109/ICCCIS51004.2021.9397227 DOI: https://doi.org/10.1109/ICCCIS51004.2021.9397227

Sridevi, V., Kumar S., S., Supraja, B., & Udhayakumar, S. (2019). Knowledge Representation and Answer Evaluation System using Language Processing Algorithm. 2019 International Conference on Vision Towards Emerging Trends in Communication and Networking (ViTECoN), 1–4. https://doi.org/10.1109/ViTECoN.2019.8899525 DOI: https://doi.org/10.1109/ViTECoN.2019.8899525

Säuberli, A., & Clematide, S. (2024). Automatic Generation and Evaluation of Reading Comprehension Test Items with Large Language Models. http://arxiv.org/abs/2404.07720 DOI: https://doi.org/10.63317/3d27vuenh2qr

Tan, L. Y., Hu, S., Yeo, D. J., & Cheong, K. H. (2025). A Comprehensive Review on Automated Grading Systems in STEM Using AI Techniques. Mathematics, 13(17), 2828. https://doi.org/10.3390/math13172828 DOI: https://doi.org/10.3390/math13172828

Vanathi, B. (2023). Automated Exam Paper Evaluation System. In International Journal of Current Science (Vol. 13, Number 2). www.ijcspub.org