Subhakanta Panda,
Harendra Kumar,
Dhawaleswar Rao,
Gitanjali Mishra,
- Teaching Associate, Department of CSE, Centurion University of Technology and Management, R. Sitapur, Odisha, India
- Project Associate, Department of CSE, Gandhi Institute of Engineering and Technology, Gunupur, Odisha, India
- Associate Professor, Department of CSE, Centurion University of Technology and Management, R. Sitapur, Odisha, India
- Assistant Professor,, Department of CSE, Gandhi Institute of Engineering and Technology, Gunupur, Odisha, India
Abstract
This paper offers a smarter system of automatic evaluation of descriptive responses in educational tests. The time-consuming aspect of testing with traditional manual marking, the subjectivity of that process, and the impossibility of scaling it makes it inapplicable, particularly to large academic environments. To resolve the issues, the proposed solution combines the methods of Natural Language Processing (NLP) and hybrid image-text processing on the responses of handwriting and typed answers. The system follows five key steps, namely, text preprocessing, feature extraction/embedding generation, similarity computation, post-processing, and evaluation. It uses the lexical, syntactic, semantic, and discourse-level characteristics to quantify the quality of answers. The common techniques that are applied to compare student responses to gold-standard answers created by experts include Bag of Words (BoW), TF-IDF, N-grams, cosine similarity, Jaccard similarity, LSA, and LDA. Additionally, deep learning-based language models and contextual embeddings are incorporated to improve semantic understanding, enabling the system to recognize paraphrased responses and conceptually equivalent answers more accurately. The framework also supports automated feedback generation, helping students identify areas for improvement while assisting educators in reducing evaluation workload and maintaining consistency across assessments. The resulting score is established by the proximity of prediction of the system compared to the assessment of the experts. The system should serve to offer gradable, objective, and uniform grading while still being in line with academic rubrics and assessment criteria.
Keywords: Natural language processing (NLP), machine learning (ML), text similarity analysis, semantic analysis, feature extraction, automated scoring system, artificial intelligence (AI).
[This article belongs to Emerging Trends in Languages ]
References
- Tan H, Wang C, Duan Q, Lu Y, Zhang H, Li R. Automatic short answer grading by encoding student responses via a graph convolutional network. Interactive Learning Environments. 2023 Apr 3;31(3):1636-50.
- Singla YK, Gupta A, Bagga S, Chen C, Krishnamurthy B, Shah RR. Speaker-conditioned hierarchical modeling for automated speech scoring. In Proceedings of the 30th ACM international conference on information & knowledge management 2021 Oct 26 (pp. 1681-1691).
- Kaushal A, Mahowald K. What do tokens know about their characters and how do they know it?. In Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies 2022 Jul (pp. 2487-2507).
- Xu X, Wang Y, Xu C, Ding Z, Jiang J, Ding Z, Karlsson BF. A survey on game playing agents and large models: Methods, applications, and challenges. arXiv preprint arXiv:2403.10249. 2024 Mar 15.
- Zhu X, Wu H, Zhang L. Automatic short-answer grading via BERT-based deep neural networks. IEEE Transactions on Learning Technologies. 2022 May 19;15(3):364-75.
- Song Y, Zhu Q, Wang H, Zheng Q. Automated essay scoring and revising based on open-source large language models. IEEE Transactions on Learning Technologies. 2024 May 6; 17:1880-90.
- Sahu A, Bhowmick PK. Feature engineering and ensemble-based approach for improving automatic short-answer grading performance. IEEE Transactions on Learning Technologies. 2019 Feb 7;13(1):77-90.
- Oswald AJ, Proto E, Sgroi D. Happiness and productivity. Journal of labor economics. 2015 Oct 1;33(4):789-822.
- Alqurashi A, Alharbi B, Sabbeh S. An Automatic Grading System for Arabic Language Short-Answer Questions Using Deep Learning. Engineering, Technology & Applied Science Research. 2025 Oct 6;15(5):26665-75.
- Kale S. The Future of IP Protection: Harnessing the Power of AI Language and Vision Models. Technology (JARET). 2024;3(1):268-76.
- Chavva RK, Muthyam SR, Seelam MS, Nalliboina N. A Transformer-Based Approach for Enhancing Automated Essay Scoring. In2024 1st International Conference on Advanced Computing and Emerging Technologies (ACET) 2024 Aug 23 (pp. 1-6). IEEE.
- Nuthalapati SB. Transforming agriculture with deep learning approaches to plant health monitoring. Remittances Review. 2022 Aug;7(1):227-38.
- Bauer E, Sailer M, Niklas F, Greiff S, Sarbu‐Rothsching S, Zottmann JM, Kiesewetter J, Stadler M, Fischer MR, Seidel T, Urhahne D. AI‐based adaptive feedback in simulations for teacher education: an experimental replication in the field. Journal of Computer Assisted Learning. 2025 Feb;41(1): e13123.
- Yu X, Lyu X, Peng R, Shen J. Solving arithmetic word problems by synergizing syntax-semantics extractor for explicit relations and neural network miner for implicit relations. Complex & Intelligent Systems. 2023 Feb;9(1):697-717.
- Ch DR, Saha SK. Generation of multiple-choice questions from textbook contents of school-level subjects. IEEE Transactions on Learning Technologies. 2022 Nov 25;16(1):40-52.
- Ren P, Yang L, Luo F. Automatic scoring of student feedback for teaching evaluation based on aspect-level sentiment analysis. Education and information technologies. 2023 Jan;28(1):797-814.
- Nguyen HT, Nguyen CT, Oka H, Ishioka T, Nakagawa M. Fully automatic scoring of handwritten descriptive answers in Japanese language tests. IEICE Technical Report; IEICE Tech. Rep.. 2021 Dec 9;121(304):45-50.
Emerging Trends in Languages
| Volume | 03 | |
| Issue | 02 | |
| Received | 29/06/2026 | |
| Accepted | 14/07/2026 | |
| Published | 15/07/2026 | |
| Publication Time | 16 Days |