Qizeng Sun
Moyi Tech, Iselin, NJ 08830, USA.
*Corresponding author: Qizeng Sun
Abstract
A study was conducted on improving the accuracy of text generation algorithms in intelligent transcription systems. The research aimed to explore how algorithmic structure optimization, multimodal information fusion, and post-processing correction mechanisms can enhance the accuracy of generated texts. The methodology combined a review of technical literature with analyses of international application cases, focusing on corpus adaptation, semantic modeling, and domain-specific practices. The results indicate that optimized algorithms demonstrate higher accuracy and stability in real-time educational meeting transcription and medical terminology recognition. The study concludes that the future development of intelligent transcription systems should emphasize algorithm generalization and cross-scenario adaptability to support broader international applications.
References
[1] Matos A, Almeida P, Correia LP, et al. iForal: Automated Handwritten Text Transcription for Historical Medieval Manuscripts. J Imaging. 2025;11(2):36.
[2] Amo IR, Aparicio BAJ, Labrador GLJ, et al. Automatic Generation of Moodle Cloze Questions for the Assessment of Knowledge About Lexical Analysis Algorithms. Educ Sci. 2025;15(1):75.
[3] Haoran F, Chundong W, Jiaqi S, et al. WordIllusion: An adversarial text generation algorithm based on human cognitive system. Cogn Syst Res. 2024;83.
[4] Bushwick S, Twombly M, Hobbs A. How AI Generates Images from Text: Algorithms use probability to predict what they should show. Sci Am. 2023;329(3):66.
[5] Jianjie W, Yilin G, Chao G, et al. An automatic text generation algorithm of technical disclosure for catenary construction based on knowledge element model. Adv Eng Inform. 2023;56.
[6] GGLOT Profers Professional Solution for Smooth and Automatic Transcription of Audio/Video Files to Text. M2 Presswire. 2022.
[7] Song B, Wu P, Zhang Q, et al. Intelligent Assessment of 95598 Speech Transcription Text Quality Based on Topic Model. IOP Conf Ser: Mater Sci Eng. 2019;563(4).
[8] Algorithms; New Algorithms Study Results from Tsinghua University Described (Lattice Based Transcription Loss for End-to-End Speech Recognition). Electron Newsw. 2018;254.
How to cite this paper
Research on Accuracy Improvement of Text Generation Algorithms in Intelligent Transcription Systems
How to cite this paper: Qizeng Sun. (2025) Research on Accuracy Improvement of Text Generation Algorithms in Intelligent Transcription Systems. Advances in Computer and Communication, 6(4), 206-211.
DOI: http://dx.doi.org/10.26855/acc.2025.10.009