EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

A Computational Analysis of AI Models in Recognizing Formal Languages

Saadat Hussain, Arapna Bai, Pooja Kumari, Karan Kumar, Aliza Salman, Khalid Rasheed, Syed Samar Yazdani

International Journal of Innovations in Science and Technology · 11.02.2026 · 343

DOI: 10.33411/ijist/ojs1775 ↗

Kurz gefasst

Fachartikel in der Zeitschrift International Journal of Innovations in Science and Technology, online erschienen am 11.02.2026. Verfasst von Saadat Hussain und 6 weiteren Autor:innen. Die Arbeit nennt 30 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Recent progress in Large Language Models (LLMs) has exhibited exceptional efficacy in natural language generation and downstream tasks. Nonetheless, research has consistently demonstrated that these models encounter difficulties with formal grammatical systems, even at the basic levels of the Chomsky Hierarchy. This work empirically investigates the constraints of modern Large Language Models (LLMs) by focusing on specific operations in regular languages, such as NFA-DFA conversion, DFA minimization, and Mealy-Moore state machine transformations. We extensively tested Google Gemini 2.5 Flash by running 50 controlled tests and comparing its outputs to validated results from a deterministic automata theory tool (control baseline: 100% accuracy across all transition tables). Our findings indicate persistent inaccuracies in transition table formulation, incomplete minimization, and recurrent erroneous state assignments. The model had a 0% success rate on complex Mealy–Moore conversions and showed significant problems when recursive ε-closure computation was required. These patterns substantiate theoretical assertions concerning the computational constraints of attention-based architectures in managing deterministic state transitions.

Zitieren

APA

Hussain, S., Bai, A., Kumari, P., Kumar, K., Salman, A., Rasheed, K., & Yazdani, S. S. (2026). A Computational Analysis of AI Models in Recognizing Formal Languages. International Journal of Innovations in Science and Technology, 343. https://doi.org/10.33411/ijist/ojs1775

Vancouver

Hussain S, Bai A, Kumari P, Kumar K, Salman A, Rasheed K, et al. A Computational Analysis of AI Models in Recognizing Formal Languages. International Journal of Innovations in Science and Technology. 2026:343. doi:10.33411/ijist/ojs1775

BibTeX
@article{Hussain2026,
  author = {Hussain, Saadat and Bai, Arapna and Kumari, Pooja and Kumar, Karan and Salman, Aliza and Rasheed, Khalid and Yazdani, Syed Samar},
  title = {A Computational Analysis of AI Models in Recognizing Formal Languages},
  journal = {International Journal of Innovations in Science and Technology},
  year = {2026},
  pages = {343},
  doi = {10.33411/ijist/ojs1775}
}
Kostenlos starten →