Treffer: Application of the Fibonacci Series in Natural Language Processing.

The article investigates some possibilities of using the new Javascript programming language for generating parsing trees and histograms for natural language processing (NLP). At the same time, we are trying to break away from the AI-type paradigm ("stochastic parrot") applied in language theory considering it too limited and focusing more on the rational approach, exploring the functional characteristics of Javascript language in the key of predicate logic. Two of Chomsky's ideas were crucial for the development of our research: first, there is a basal grammar innate to the child that provides the very structures the language is built upon, and second, the recursive nature of the human language Chomsky had noticed that allows us to build an indefinite amount of statements from a finite set of grammar rules, plus the composable character of grammar that begets infinite long verbal structures. The last observation gave us the idea to establish isomorphic relations between natural language and formal systems to prove our theorems (future work). The agglutinative mechanism of making both sensical or not-sensical verbal content drew our attention to the Fibonacci series of numbers that is fundamental for developing living structures both at the molecular and macro level (e.g. breeding of rabbits, use of the phi constant in architecture, and so on). This way the bigrams (from n-grams) can result from concatenating strings one after another creating entities in a coherent, linked-list style, rational manner. These collocations may be interpreted either as new concepts (e.g. military-doctor) or camouflaged predicates built upon identity principles (e.g. is or exists). The practical way we decided to test our hypothesis was to employ the functional capabilities of Javascript programming language that brings us even closer to the logical nature of the human language. The new ES6 streaming process of transforming a text was another aspect similar to the pipeline style of the human brain in processing data. The two-way parsing on texts calculating the frequency of pairs' appearance proved to be of significant importance in dead language studies or searching for anagrams. The use of the Wink software package language model allowed us to create predicates like verbs (subject, object) based on the SVO structure of IE languages, the future knowledge base for our next proofing language system. [ABSTRACT FROM AUTHOR]

Copyright of Didactica Danubiensis is the property of Danubius University of Galati and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)