• MALSORI
• /
• no.56
• /
• pp.207-223
• /
• 2005

The goal of language model adaptation is to improve the background language model with a relatively small adaptation corpus. This study presents a language model adaptation technique where additional text data for the adaptation do not exist. We propose the information retrieval (IR) technique with N-gram language modeling to collect the adaptation corpus from baseline text data. We also propose to use a dynamic language model interpolation coefficient to combine the background language model and the adapted language model. The interpolation coefficient is estimated from the word hypotheses obtained by segmenting the input speech data reserved for held-out validation data. This allows the final adapted model to improve the performance of the background model consistently The proposed approach reduces the word error rate by $13.6\%$ relative to baseline 4-gram for two-hour broadcast news speech recognition.

• The Journal of the Acoustical Society of Korea
• /
• v.23 no.2E
• /
• pp.51-55
• /
• 2004

In this paper, we propose a style-specific language model adaptation scheme using n-gram based tf*idf similarity for Korean spontaneous speech recognition. Korean spontaneous speech shows especially different style-specific characteristics such as filled pauses, word omission, and contraction, which are related to function words and depend on preceding or following words. To reflect these style-specific characteristics and overcome insufficient data for training language model, we estimate in-domain dependent n-gram model by relevance weighting of out-of-domain text data according to their n-. gram based tf*idf similarity, in which in-domain language model include disfluency model. Recognition results show that n-gram based tf*idf similarity weighting effectively reflects style difference.

• MALSORI
• /
• no.46
• /
• pp.117-126
• /
• 2003

This paper presents our language model adaptation for Korean spontaneous speech recognition. Korean spontaneous speech is observed various characteristics of content and style such as filled pauses, word omission, and contraction as compared with the written text corpus. Our approaches focus on improving the estimation of domain-dependent n-gram models by relevance weighting out-of-domain text data, where style is represented by n-gram based tf/sup */idf similarity. In addition to relevance weighting, we use disfluencies as Predictor to the neighboring words. The best result reduces 9.7% word error rate relatively and shows that n-gram based relevance weighting reflects style difference greatly and disfluencies are good predictor also.

• Proceedings of the KSPS conference
• /
• 2003.05a
• /
• pp.83-86
• /
• 2003

This paper presents our style-based language model adaptation for Korean conversational speech recognition. Korean conversational speech is observed various characteristics of content and style such as filled pauses, word omission, and contraction as compared with the written text corpora. For style-based language model adaptation, we report two approaches. Our approaches focus on improving the estimation of domain-dependent n-gram models by relevance weighting out-of-domain text data, where style is represented by n-gram based tf*idf similarity. In addition to relevance weighting, we use disfluencies as predictor to the neighboring words. The best result reduces 6.5％ word error rate absolutely and shows that n-gram based relevance weighting reflects style difference greatly and disfluencies are good predictor.

• 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.

• Annual Conference on Human and Language Technology
• /
• 2010.10a
• /
• pp.143-146
• /
• 2010

뇌-컴퓨터 인터페이스는 뇌에서 발생하는 생체신호를 통하여 컴퓨터나 외부기기를 직접 제어할 수 있는 기술이다. 자발적으로 언어를 생성하지 못하는 환자들을 위하여 뇌-컴퓨터 인터페이스를 이용하여 한국어를 자유롭게 입력할 수 있는 인터페이스에 대한 연구가 필요하다. 본 연구는 의사소통을 위한 뇌-컴퓨터 인터페이스에서 낮은 정보전달률을 개선하기 위해서 음절 n-gram과 어절 n-gram 모델을 이용하여 언어 예측 모델을 구현하였다. 또한 실제 이를 이용한 뇌 컴퓨터 한국어 입력기를 설계하였다, 이는 기존의 뇌-컴퓨터 인터페이스 연구에서 특징 추출이나 기계학습 방법의 성능향상을 위한 연구와는 차별적인 방법이다.

• KIISE Transactions on Computing Practices
• /
• v.22 no.7
• /
• pp.315-319
• /
• 2016

Recurrent neural network based language models (RNN LM) have shown improved results in language model researches. The RNN LMs are limited to post processing sessions, such as the N-best rescoring step of the wFST based speech recognition. However, it has considerable vocabulary problems that require large computing powers for the LM training. In this paper, we try to find the 1st pass N-gram model using word embedding, which is the simplified deep neural network. The class based language model (LM) can be a way to approach to this issue. We have built class based vocabulary through word embedding, by combining the class LM with word N-gram LM to evaluate the performance of LMs. In addition, we propose that part-of-speech (POS) tagging based LM shows an improvement of perplexity in all types of the LM tests.

• Annual Conference on Human and Language Technology
• /
• 2006.10e
• /
• pp.125-132
• /
• 2006

한국어의 자연어처리 및 정보검색분야에서 자동 띄어쓰기는 매우 중요한 문제이다. 신문기사에서조차 잘못된 띄어쓰기를 발견할 수 있을 정도로 띄어쓰기가 어려운 경우가 많다. 본 논문에서는 자기 조직화 n-gram모델을 이용해 자동 띄어쓰기의 정확도를 높이는 방법을 제안한다. 본 논문에서 제안하는 방법은 문맥의 길이를 바꿀 수 있는 가변길이 n-gram모델을 기본으로 하여 모델이 자동으로 문맥의 길이를 결정하도록 한 것으로, 일반적인 n-gram모델에 비해 더욱 높은 성능을 얻을 수 있다. 자기조직화 n-gram모델은 최적의 문맥의 길이를 찾기 위해 문맥의 길이를 늘였을 때 나타나는 확률분포와 문맥의 길이를 늘이지 않았을 태의 확률분포를 비교하여 그 차이가 크다면 문맥의 길이를 늘이고, 그렇지 않다면 문맥의 길이를 자동으로 줄인다. 즉, 더 많은 정보가 필요한 경우는 데이터의 차원을 높여 정확도를 올리며, 이로 인해 증가된 계산량은 필요 없는 데이터의 양을 줄임으로써 줄일 수 있다. 본 논문에서는 실험을 통해 n-gram모델의 자기 조직화 구조가 기본적인 모델보다 성능이 뛰어나다는 것을 확인하였다.

• Smart Media Journal
• /
• v.7 no.2
• /
• pp.54-59
• /
• 2018

With the increasing use of touch-enabled mobile devices such as smartphones and tablet PCs, the works are done on desktop computers and smartphones, and tablet PCs perform laptops. However, due to the nature of smart devices that require portability, QWERTY keyboard is densely arranged in a small screen. This is the cause of different typographical errors when using the mechanical QWERTY keyboard. Unlike the mechanical QWERTY keyboard, which has enough space for each button, QWERTY keyboard on the touch screen often has a small area assigned to each button, so that it is often the case that the surrounding buttons are input rather than the button the user intends to press. In this paper, we propose a method to automatically correct the input errors of the QWERTY keyboard in the touch screen environment by using the n-gram language model using the word unigram and the bigram probability.

• Annual Conference on Human and Language Technology
• /
• 2020.10a
• /
• pp.554-558
• /
• 2020

최근 사전학습 모델의 발달로 기계독해 시스템 성능이 크게 향상되었다. 하지만 기계독해 시스템은 주어진 단락에서 질문에 대한 정답을 찾기 때문에 단락을 직접 검색해야하는 실제 환경에서의 성능 하락은 불가피하다. 즉, 기계독해 시스템이 오픈 도메인 환경에서 높은 성능을 보이기 위해서는 높은 성능의 검색 모델이 필수적이다. 따라서 본 논문에서는 검색 모델의 성능을 보완해 줄 수 있는 오픈 도메인 기계독해를 위한 단락 재순위화 모델을 제안한다. 제안 모델은 합성곱 신경망을 이용하여 질문과 단락을 구절 단위로 표현했으며, N-gram 구절 사이의 상호 주의 집중을 통해 질문과 단락 사이의 관계를 효과적으로 표현했다. KorQuAD를 기반으로한 실험에서 제안모델은 MRR@10 기준 93.0%, Top@1 Precision 기준 89.4%의 높은 성능을 보였다.