• Journal of the Korean Institute of Intelligent Systems
• /
• v.14 no.3
• /
• pp.357-362
• /
• 2004

Language models are essential in predicting the next word in a spoken sentence, thereby enhancing the speech recognition accuracy, among other things. However, spoken language domains are too numerous, and therefore developers suffer from the lack of corpora with sufficient sizes. This paper proposes a method of combining two n-gram language models, one constructed from a very small corpus of the right domain of interest, the other constructed from a large but less adequate corpus, resulting in a significantly enhanced language model. This method is based on the observation that a small corpus from the right domain has high quality n-grams but has serious sparseness problem, while a large corpus from a different domain has more n-gram statistics but incorrectly biased. With our approach, two n-gram statistics are combined by extending the idea of Katz's backoff and therefore is called a dual-source backoff. We ran experiments with 3-gram language models constructed from newspaper corpora of several million to tens of million words together with models from smaller broadcast news corpora. The target domain was broadcast news. We obtained significant improvement (30%) by incorporating a small corpus around one thirtieth size of the newspaper corpus.

• Journal of the Korea Society of Computer and Information
• /
• v.26 no.10
• /
• pp.157-165
• /
• 2021

Recently, studies using deep learning to analyze a large amount of text are being actively conducted. In particular, a pre-trained language model that applies the learning results of a large amount of text to the analysis of a specific domain text is attracting attention. Among various pre-trained language models, BERT(Bidirectional Encoder Representations from Transformers)-based model is the most widely used. Recently, research to improve the performance of analysis is being conducted through further pre-training using BERT's MLM(Masked Language Model). However, the traditional MLM has difficulties in clearly understands the meaning of sentences containing new words such as newly coined words. Therefore, in this study, we newly propose NTM(Newly coined words Target Masking), which performs masking only on new words. As a result of analyzing about 700,000 movie reviews of portal 'N' by applying the proposed methodology, it was confirmed that the proposed NTM showed superior performance in terms of accuracy of sensitivity analysis compared to the existing random masking.

• Journal of Intelligence and Information Systems
• /
• v.28 no.3
• /
• pp.45-55
• /
• 2022

Recently, large-scale language models (LPLM) have been shown state-of-the-art performances in various tasks of natural language processing including intent classification. However, fine-tuning LPLM requires much computational cost for training and inference which is not appropriate for dialog system. In this paper, we propose compressed intent classification model for multi-agent in low-resource like CPU. Our method consists of two stages. First, we trained sentence encoder from LPLM then compressed it through knowledge distillation. Second, we trained agent-specific adapter for intent classification. The results of three intent classification datasets show that our method achieved 98% of the accuracy of LPLM with only 21% size of it.

• MALSORI
• /
• no.51
• /
• pp.151-165
• /
• 2004

For educational and research purposes, a Korean speech recognition platform is designed. It is based on an object-oriented architecture and can be easily modified so that researchers can readily evaluate the performance of a recognition algorithm of interest. This platform will save development time for many who are interested in speech recognition. The platform includes the following modules: Noise reduction, end-point detection, met-frequency cepstral coefficient (MFCC) and perceptually linear prediction (PLP)-based feature extraction, hidden Markov model (HMM)-based acoustic modeling, n-gram language modeling, n-best search, and Korean language processing. The decoder of the platform can handle both lexical search trees for large vocabulary speech recognition and finite-state networks for small-to-medium vocabulary speech recognition. It performs word-dependent n-best search algorithm with a bigram language model in the first forward search stage and then extracts a word lattice and restores each lattice path with a trigram language model in the second stage.

• Annual Conference on Human and Language Technology
• /
• 2020.10a
• /
• pp.124-127
• /
• 2020

관계 추출은 문장에서 두 개의 엔티티가 주어졌을 때 두 개의 엔티티에 대한 의미적 이해를 통해 관계를 분류하는 작업이다. 이와 같이 관계 추출에서 관계를 분류하기 위해서는 두 개의 엔티티에 대한 정보가 필요하다. 본 연구에서는 관계 추출을 하기 위해 문장에서 엔티티들의 표현을 다르게하여 관계 추출의 성능을 비교 실험하였다. 첫번째로는 문장에서 [CLS] 토큰(Token)으로 관계를 분류하는 Standard 엔티티 정보 표현과 두번째로는 엔티티의 앞과 뒤에 Special Token을 추가하여 관계를 분류하는 Entity-Markers 엔티티 정보 표현했다. 이를 기반으로 문장의 문맥 정보를 학습한 사전 학습(Pre-trained)모델인 BERT-Large와 ALBERT-Large를 적용하여 실험을 진행하였다. 실험 결과 Special Token을 추가한 Entity-Markers의 성능이 높았으며, BERT-Large에서 더 높은 성능 결과를 확인하였다.

• Journal of KIISE:Software and Applications
• /
• v.26 no.11
• /
• pp.1359-1371
• /
• 1999

대규모 네트워크 상에서 동작하는 분산 시스템의 구현을 위해 제시된 방법 중의 하나인 이동 코드 개념은 네트워크 공유 자원에 접근할 수 있는 효과적인 방법을 제시하였고 이 개념을 지원하는 많은 언어들의 개발을 가져왔다. 개발된 언어들이 가지고 있는 이동 코드를 지원하기 위한 언어 구문과 적용하려는 문제 영역의 특성을 반영한 언어 구문은 네트워크 프로그래밍을 하는데 있어서 효율과 문제 중심의 프로그램의 두 가지를 모두 가능하게 하고 있다. 본 논문에서는 현재 분산 컴퓨팅 환경에서 가장 많이 사용되고 있는 클라이언트-서버 모델을 확장하여 서버의 자원에 접근할 수 있는 또 다른 방법을 가진 모델을 제시하고, 이 모델을 표현할 수 있는 언어를 설계하였다. 설계된 언어는 이동 코드의 개념을 지원함으로써 대규모 네트워크에서 수행되는 프로그램의 작성을 가능하게 하고, 분산 범위 규칙을 채택함으로써 이동 코드의 기술을 일반 함수를 기술하듯 명확한 관점에서 할 수 있도록 하였다. 또한 네트워크 관련 자원들을 언어 구문으로 채택하여 네트워크 프로그래밍을 언어 수준에서 할 수 있도록 하였다. 언어의 이론적인 설계에 그치지 않고 설계된 언어를 수행할 수 있는 실행 시간 지원 시스템을 구현하였다. 실행 시간 지원 시스템은 언어를 해석하고 실행하는 코드 해석기와 이동 코드를 지원하는 네트워크 감독으로 구성되며 설계된 언어를 사용하여 실제로 네트워크 응용 프로그램을 작성하고 테스트 해 볼 수 있다.Abstract Some studies bring up a concept of code mobility as an innovative way to access network resources in order to develop distributed systems working on a large scale network. After that, many languages are suggested to support this concept. In these languages, language constructors for their particular application domains and mobile codes provide both problem-oriented views to the programmer and reasonable performance to the system. In this thesis, we extend the client-server model that is the most popular model in developing distributed systems these days. We propose a model to have another method to access server's resources and extend the C language to implement the proposed model for the large scale network. The new language has capability to build a software working on a large scale network by supporting mobile code and gives a consistent network programming view to the programmer by adapting distributed semantics. The language also makes network programming easy by providing network primitives at the language level. We implement a prototype of run-time system to support this language. The run-time system is composed of two major parts: code-interpreter that interprets and executes the language and network-daemon that supports mobile codes.

• Journal of the Korean Society for information Management
• /
• v.24 no.1 s.63
• /
• pp.209-226
• /
• 2007

XML documents format on the Web provides a mechanism to impose their content and logical structure information. Therefore, an XML processor provides access to their content and structure. The purpose of this study is to investigate the usefulness of structural hints in the leaf node language model-based XML document retrieval. In order to this purpose, this experiment tested the performances of the leaf node language model-based XML retrieval system to compare the queries for a topic containing only content-only constraints and both content constrains and structure constraints. A newly designed and implemented leaf node language model-based XML retrieval system was used. And we participated in the ad-hoc track of INEX 2005 and conducted an experiment using a large-scale XML test collection provided by INEX 2005.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.16 no.4
• /
• pp.193-202
• /
• 2023

In recent years, the outstanding performance of large language models (LLMs) trained on extensive datasets has become a hot topic. Since studies on LLMs are available on open-source approaches, the ecosystem is expanding rapidly. Models that are task-specific, lightweight, and high-performing are being actively disseminated using additional training techniques using pre-trained LLMs as foundation models. On the other hand, the performance of LLMs for Korean is subpar because English comprises a significant proportion of the training dataset of existing LLMs. Therefore, research is being carried out on Korean-specific LLMs that allow for further learning with Korean language data. This paper identifies trends of open source based LLMs and introduces research on Korean specific large language models; moreover, the applications and limitations of large language models are described.

• Annual Conference on Human and Language Technology
• /
• 2023.10a
• /
• pp.301-305
• /
• 2023

최근 등장한 Large Language Models (LLM)은 자연어 처리 분야에서 눈에 띄는 성과를 보여주었지만, 주로 영어 중심의 연구로 진행되어 그 한계를 가지고 있다. 본 연구는 사전 학습된 LLM의 언어별 지식 전이 가능성을 한국어를 중심으로 탐구하였다. 이를 위해 한국어와 영어로 구성된 코드 스위칭 코퍼스를 구축하였으며, 기본 모델인 LLAMA-2와 코드 스위칭 코퍼스를 추가 학습한 모델 간의 성능 비교를 수행하였다. 결과적으로, 제안하는 방법론으로 학습한 모델은 두 언어 간의 희미론적 정보가 효과적으로 전이됐으며, 두 언어 간의 지식 정보 연계가 가능했다. 이 연구는 다양한 언어와 문화를 반영하는 다국어 LLM 연구와, 소수 언어를 포함한 AI 기술의 확산 및 민주화에 기여할 수 있을 것으로 기대된다.

• Journal of the Korea Society of Computer and Information
• /
• v.27 no.2
• /
• pp.15-23
• /
• 2022

In the field of patents, as NLP(Natural Language Processing) is a challenging task due to the linguistic specificity of patent literature, there is an urgent need to research a language model optimized for Korean patent literature. Recently, in the field of NLP, there have been continuous attempts to establish a pre-trained language model for specific domains to improve performance in various tasks of related fields. Among them, ELECTRA is a pre-trained language model by Google using a new method called RTD(Replaced Token Detection), after BERT, for increasing training efficiency. The purpose of this paper is to propose KorPatELECTRA pre-trained on a large amount of Korean patent literature data. In addition, optimal pre-training was conducted by preprocessing the training corpus according to the characteristics of the patent literature and applying patent vocabulary and tokenizer. In order to confirm the performance, KorPatELECTRA was tested for NER(Named Entity Recognition), MRC(Machine Reading Comprehension), and patent classification tasks using actual patent data, and the most excellent performance was verified in all the three tasks compared to comparative general-purpose language models.