Comparative Analysis of Analytical and Synthetic Language Systems in LLM Training

Read full text Summarise with VIRAT
Authors: O. V. Bychkovska
Cited by: 0
References: 3

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Knowledge acquisition in large language models (LLMs) is fundamentally influenced by the structural properties of the training language, which shape the results of large-scale textual data processing.Analytical language systems, characterized by their syntactic regularity and relatively simple morphology, have served as the primary environments for model training.Their structural consistency facilitates formalized knowledge encoding, while synthetic systems, characterized by rich morphological paradigms and higher inflectional variability, present unique challenges for linguistic pattern recognition.These challenges are further compounded by the status of many synthetic languages as low-resource environments compared to the extensive digitized corpora available for widely used analytical languages.This paper presents a comparative analysis of analytical and synthetic language systems in the context of LLM training, exploring how structural typology and data availability impact knowledge representation within these models.

Fields of Study & Subjects

Natural Language Processing Techniques Topic Modeling Syntax Semantics Linguistic Variation

Publication details

Pages
63–65
Publication date
14 March 2026
Language
EN