• Phonetics and Speech Sciences
• /
• v.6 no.2
• /
• pp.29-34
• /
• 2014

In the Vector Taylor Series (VTS)-based noisy speech recognition methods, Hidden Markov Models (HMM) are usually trained with clean speech. However, better performance is expected by training the HMM with noisy speech. In a previous study, we could find that Minimum Mean Square Error (MMSE) estimation of the training noisy speech in the log-spectrum domain produce improved recognition results, but since the proposed algorithm was done in the log-spectrum domain, it could not be used for the HMM adaptation. In this paper, we modify the previous algorithm to derive a novel mathematical relation between test and training noisy speech in the cepstrum domain and the mean and covariance of the Multi-condition TRaining (MTR) trained noisy speech HMM are adapted. In the noisy speech recognition experiments on the Aurora 2 database, the proposed method produced 10.6% of relative improvement in Word Error Rates (WERs) over the MTR method while the previous MMSE estimation of the training noisy speech produced 4.3% of relative improvement, which shows the superiority of the proposed method.

• International Journal of Computer Science & Network Security
• /
• v.22 no.2
• /
• pp.115-122
• /
• 2022

Component adaptation is widely recognized as one of the main problems of the components, used in component based software engineering (CBSE). We developed methods to adjust the components classified by the keywords. Three main methods are discussed in this article those methods are combined with several domain component interfaces, high level simple notation for the adapter design patterns. The automated process for classifying high-level components are using adaptation is novel to software engineering domain. All Specifications and many technologies for re-using software, CBD and further developments have been emerged in recent years. The effects of these technologies on program quality or software costs must be analyzed. The risk concerns a single technology and must identify its combinations. In this paper, we are going to discuss the methods to adapt components of different technologies

• The Journal of the Acoustical Society of Korea
• /
• v.22 no.8
• /
• pp.727-739
• /
• 2003

In this paper, we propose a generation method of regression classes for adaptation in the HM-Net (Hidden Markov Network) system. The MLLR (Maximum Likelihood Linear Regression) adaptation approach is applied to the HM-Net speech recognition system for expressing the characteristics of speaker effectively and the use of HM-Net in various tasks. For the state level sharing, the context domain state splitting of PDT-SSS (Phonetic Decision Tree-based Successive State Splitting) algorithm, which has the contextual and time domain clustering, is adopted. In each state of contextual domain, the desired phoneme classes are determined by splitting the context information (classes) including target speaker's speech data. The number of adaptation parameters, such as means and variances, is autonomously controlled by contextual domain state splitting of PDT-SSS, depending on the context information and the amount of adaptation utterances from a new speaker. The experiments are performed to verify the effectiveness of the proposed method on the KLE (The center for Korean Language Engineering) 452 data and YNU (Yeungnam Dniv) 200 data. The experimental results show that the accuracies of phone, word, and sentence recognition system increased by 34∼37%, 9%, and 20%, respectively, Compared with performance according to the length of adaptation utterances, the performance are also significantly improved even in short adaptation utterances. Therefore, we can argue that the proposed regression class method is well applied to HM-Net speech recognition system employing MLLR speaker adaptation.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2023.01a
• /
• pp.17-18
• /
• 2023

비지도형 도메인 적응(UDA)에 대한 최근 연구는 도메인 적응에 대한 설명 및 전이 가능한 특징을 풀어 내기 위해 적대적 학습에 의존한다. 그러나 기존 방법에는 대상 도메인의 클래스 인식(class-aware) 정보를 고려하지 않고는 잠재 공간의 구별 가능성을 완전히 보장할 수 없다는 것과 소스 및 대상 도메인의 샘플만으로는 잠재 공간에서 도메인 불변(domain- invariant) 특성을 추출하기에 부족하다는 두 가지 문제가 있다고 알려져 있다. 본 논문에서는 기존 알려진 UDA의 도메인 적응시 발생되는 문제를 해결하기 위해 Adversarial Discriminative Domain Adaptation(ADDA)에서 mixup을 활용해 신경망의 로버스트네스를 향상시키는 것을 확인하였다.

• Journal of Ship and Ocean Technology
• /
• v.5 no.3
• /
• pp.27-35
• /
• 2001

In this research, hybrid method with case-based reasoning and rule-based reasoning is applied. Using case-based reasoning, design experts'experience and know-how are effectively represented in order to obtain a proper configuration of midship section in the initial ship design stage. Since there is not sufficient domain knowledge available to us, traditional case-adaptation algorithms cannot be applied to our problem, i.e., creating the configuration of midship section. Thus, new case-adaptation algorithms not requiring any domain knowledge are developed antral applied to our problem. Using the knowledge representation of DnV rules, rule-based reasoning can perform deductive inference in order to obtain the scantling of midship section efficiently. The results from the case-based reasoning and the rule-based reasoning are examined by comparing the results with various conventional methods. And the reasonability of our results is verified by comparing the results wish actual values from parent ship.

• 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.

• Proceedings of the KSPS conference
• /
• 2005.04a
• /
• pp.53-56
• /
• 2005

In this paper, we propose a transformation based robust adaptation technique that uses the maximum mutual information(MMI) estimation for the objective function and the linear spectral transformation(LST) for adaptation. LST is an adaptation method that deals with environmental noises in the linear spectral domain, so that a small number of parameters can be used for fast adaptation. The proposed technique is called MMI-LST, and evaluated on TIMIT and FFMTIMIT corpora to show that it is advantageous when only a small amount of adaptation speech is used.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.38A no.8
• /
• pp.697-703
• /
• 2013

We propose an efficient frequency-time domain resource allocation scheme in multi-carrier (MC) direct-sequence code-division multiple-access (DS/CDMA) communications. We consider, as a power allocation strategy in the frequency domain, transmitting each user's DS waveforms over the user's sub-band with the largest channel gain. We then consider rate adaptation in the time domain, where the data rate is adapted such that a desired transmission quality is maintained. We analyze the achievable average data rate of the proposed scheme with fixed average transmission power, and compare the performance to single carrier DS/CDMA systems with power and rate adaptations.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2023.05a
• /
• pp.678-680
• /
• 2023

본 논문은 정확하면서도 효율적인 한국어 문장 분류 기법에 대해서 논의한다. 최근 자연어처리 분야에서 사전 학습된 언어 모델(Pre-trained Language Models, PLM)은 미세조정(fine-tuning)을 통해 문장 분류 하위 작업(downstream task)에서 성공적인 결과를 보여주고 있다. 하지만, 이러한 미세조정은 하위 작업이 바뀔 때마다 사전 학습된 언어 모델의 전체 매개변수(model parameters)를 학습해야 한다는 단점을 갖고 있다. 본 논문에서는 이러한 문제를 해결할 수 있도록 도메인 적응기(domain adapter)를 활용한 한국어 문장 분류 프레임워크인 DAKS(Domain Adaptation-based Korean Sentence classification framework)를 제안한다. 해당 프레임워크는 학습되는 매개변수의 규모를 크게 줄임으로써 효율적인 성능을 보였다. 또한 문장 분류를 위한 특징(feature)으로써 한국어 사전학습 모델(KLUE-RoBERTa)의 다양한 은닉 계층 별 은닉 상태(hidden states)를 활용하였을 때 결과를 비교 분석하고 가장 적합한 은닉 계층을 제시한다.

• The Journal of Korea Robotics Society
• /
• v.18 no.3
• /
• pp.346-351
• /
• 2023

With the release of numerous open driving datasets, the demand for domain adaptation in perception tasks has increased, particularly when transferring knowledge from rich datasets to novel domains. However, it is difficult to solve the change 1) in the sensor domain caused by heterogeneous LiDAR sensors and 2) in the environmental domain caused by different environmental factors. We overcome domain differences in the semi-supervised setting with 3-stage model parameter training. First, we pre-train the model with the source dataset with object scaling based on statistics of the object size. Then we fine-tine the partially frozen model weights with copy-and-paste augmentation. The 3D points in the box labels are copied from one scene and pasted to the other scenes. Finally, we use the knowledge distillation method to update the student network with a moving average from the teacher network along with a self-training method with pseudo labels. Test-Time Augmentation with varying z values is employed to predict the final results. Our method achieved 3rd place in ECCV 2022 workshop on the 3D Perception for Autonomous Driving challenge.