- Translation using A.I. ^^;; 성능에 놀랐음..
1. Introduction
1.1 Research Purpose
The perspective of language research reflects the way language itself exists. ①From a structuralist perspective, language is regarded as an organic natural entity. It is seen as an observable phenomenon that exists independently of humans. ②From a functionalist perspective, language is considered a relational and dynamic phenomenon that can serve different functions depending on context or situation. It is viewed as part of socio‑cultural sign systems that exist for communicative purposes, interacting with situations and socio‑cultural codes to fulfill multiple functions. These two perspectives seem to correspond respectively to langue and parole as primary objects of study. ③From the perspective of an individual, language also constitutes a cognitive ability processed in the brain. As the totality of what a human brain acquires and processes, language can be approached from neurobiological or cognitive perspectives.
Human “language” can be studied from all three perspectives, which makes it a highly intriguing research subject. In particular, the third perspective has attracted increasing attention with the development of neuroscience, cognitive science, and computer science, and with the emergence of artificial intelligence. Theoretical exploration of how language is represented in the brain can itself be fascinating.
Artificial neural networks are constructed by connecting perceptrons, which model human neurons. Today, these networks are used to perform various language processing tasks. This raises the question: might the process by which the human brain acquires language resemble the process by which artificial intelligence learns language? If AI’s language learning is similar to that of the human brain, could perceptron models be used to represent linguistic phenomena? Might that even amount to grammatical description? Could we design a model that integrates neural information processing with brain structure and function to describe human linguistic ability? To what extent can traditional linguistics contribute to this?
To test this possibility, I attempt to design a perceptron model that explains the grammatical phenomenon of the Korean element “‑었‑.”
The full text is attached as a file