딥러닝 자연어처리(NLP)와 일반수사학 (General Rhetoric)과의 융합적 접점 분석 - 그룹 뮤(Groupe μ)의 『A General Rhetoric』을 중심으로 = Convergence Analysis of Deep Learning Natural Language Processing(NLP) and General Rhetoric: Focusing on Groupe μ ’s 『A General Rhetoric』

국문 초록 (Abstract)

본 논문에서는 그룹 뮤(Groupe μ)의 『일반수사학』에서 주목하고 있는 시학과 수사 학적 언어의 형태와 RNN 언어모델의 구조적 유사성을 분석하고자 한다. 여기에서 주로 동원되는 것은 시학과 수사학적 언어의 기능을 분류하는 ‘메타볼’(Metaboles)의 유 형이다. 특히 메타볼은 본 논문에서 다루게 될 ‘수사학적 공간’의 아이디어로 발전하여 딥러닝 기술을 통한 자연어처리와 실질적으로 연결될 수 있다는 점에서 중요한 개념이라고 할 수 있다. 또한 수사학적 공간 개념을 RNN 언어모델에 적용하기 위해서는 무엇보다도 정밀한 언어 데이터 셋 마련과 적확한 ‘의미의 영점’(Degree-Zero) 마련 역시 중요하다. 이런 사실은 결국 학문적 융합이 현재 A.I. 기술에 얼마나 필수적인지를 보여준다. 마지막으로 『일반수사학』의 내용은 현재의 딥러닝 언어모델에서도 여전히 유효하다는 점에서 날카로운 혜안을 증언하고 있으며 우리에게 여전히 인문학과 공학의 초학제적 태도는 중요할 수밖에 없음을 보여준다.

다국어 초록 (Multilingual Abstract)

In this article, I attempt to analyze the structural similarity between the semantic structure of poetry and rhetorical language on which Groupe μ is focusing and the RNN language model. The main focus of this article was concentrated on the type of ‘metaboles’ that categorizes the functions of poetry and rhetorical language. This type and its insight are considerably useful not only in connection with natural language processing technology but also a application field. In particular, the approach for the idea of ‘rhetorical space’ can be substantially an important concept in terms of connection with deep learning technology of NLP. In addition, this approach shows the fact that it is the paramount importance to prepare a precise language data set which has ‘Degree Zero’ and ‘Subunits’ of the linguistic code. Ultimately, through this convergence analysis, we realize that the trans-disciplinary attitude of the humanities and engineering is inevitably important.

참고문헌 (Reference)

1 토도르프. T., "상징이론" 한국문화사 1995

2 Logan, Robert, "What is Information?: Propagating Organization in the Biosphere, Symbolosphere, Technosphere and Econosphere" Demo Publishing 2014

3 Ciaburro, Giuseppe, "Venkateswaran, Balaji, Neural Networks with R: Smart models using CNN, RNN, deep learning, and artificial intelligence principles" Packt Publishing Ltd 2017

4 Ramos, Juan, "Using tf-idf to determine word relevance in document queries" 242 : 2003

5 Hinton, Geoffrey, "Unsupervised Learning: Foundations of Neural Computation A Bradford Book Computational neuroscience" MIT Press 1999

6 Ling, Wang, "Two/too simple adaptations of word2vec for syntax problems" 2015

7 Leung, H., "The complex backpropagation algorithm" 39 (39): 2102-2104, 1991

8 McTear, Michael, "The Conversational Interface:Talking to Smart Devices" Springer 2016

9 Liu, Donghai, "Some cosine similarity measures and distance measures between q‐rung orthopair fuzzy sets" 34 (34): 1572-1587, 2019

10 Benveniste, E., "Problems in General Linguistics" University of Miami Press 1971

11 Prakash, M. Nadkarni, "Natural laguage porcessing: an introduction" 18 (18): 544-551, 2011

12 Cavnar, B. William, "N-gram-based text categorization" 161175 : 1994

13 Mitchell, T. M., "Machine Learning" McGraw-Hill 1997

14 Brown, B. Tom, "Language Models are Few-Shot Learners" Johns Hopkins University 2020

15 Douglas, C. Montgomery, "Introduction to Linear Regression Analysis" John Wiley & Sons 2012

16 Pennington, Jeffrey, "Glove: Global vectors for word representation" 2014

17 Dubois, J., "General Rhetoric" The Johns Hopkins Press 1981

18 Devlin, Jacob, "Bert: Pre-training of deep bidirectional transformers for language understanding"

19 Deshpande, Anand, "Artificial Intelligence for Big Data: Complete guide to automating Big Data solutions using Artificial Intelligence techniques" Packt Publishing Ltd 2018

20 Trott, M. Adriel, "Aristotle on the Nature of Community" Cambridge University Press 2014