• 제목/요약/키워드: Task adaptation

검색결과 104건 처리시간 0.022초

Performance of Vocabulary-Independent Speech Recognizers with Speaker Adaptation

  • Kwon, Oh Wook;Un, Chong Kwan;Kim, Hoi Rin
    • The Journal of the Acoustical Society of Korea
    • /
    • 제16권1E호
    • /
    • pp.57-63
    • /
    • 1997
  • In this paper, we investigated performance of a vocabulary-independent speech recognizer with speaker adaptation. The vocabulary-independent speech recognizer does not require task-oriented speech databases to estimate HMM parameters, but adapts the parameters recursively by using input speech and recognition results. The recognizer has the advantage that it relieves efforts to record the speech databases and can be easily adapted to a new task and a new speaker with different recognition vocabulary without losing recognition accuracies. Experimental results showed that the vocabulary-independent speech recognizer with supervised offline speaker adaptation reduced 40% of recognition errors when 80 words from the same vocabulary as test data were used as adaptation data. The recognizer with unsupervised online speaker adaptation reduced abut 43% of recognition errors. This performance is comparable to that of a speaker-independent speech recognizer trained by a task-oriented speech database.

  • PDF

A New Speaker Adaptation Technique using Maximum Model Distance

  • Lee, Man-Hyung;Hong, Suh-Il
    • 제어로봇시스템학회:학술대회논문집
    • /
    • 제어로봇시스템학회 2001년도 ICCAS
    • /
    • pp.99.1-99
    • /
    • 2001
  • This paper presented an adaptation approach based on maximum model distance (MMD) method. This method shares the same framework as they are used for training speech recognizers with abundant training data. The MMD method could adapt to all the models with or without adaptation data. If large amount of adaptation data is available, these methods could gradually approximate the speaker-dependent ones. The approach is evaluated through the phoneme recognition task on the TIMIT corpus. On the speaker adaptation experiments, up to 65.55% phoneme error reduction is achieved. The MMD could reduce phoneme error by 16.91% even when only one adaptation utterance is used.

  • PDF

치위생과 학생들의 성취목표지향성과 자기효능감이 전공적응 및 만족에 미치는 영향 (The effects of dental-hygiene students' goal achievement orientation on their self-efficacy and major adaptation and satisfaction)

  • 정기옥
    • 한국치위생학회지
    • /
    • 제12권1호
    • /
    • pp.167-178
    • /
    • 2012
  • Objectives : The purpose of this research is to examine the effects of the goal achievement orientation of the students in dental-hygiene practice classes on their self-efficacy and on their adaptation to and satisfaction with their major. Methods : A survey was conducted among the dental-hygiene students in schools located in the Daegu region on October 24-31, 2011. A total of 212 accomplished questionnaires were gathered as sources of final analytical data. Statistical analysis was conducted on the collected data, using the SPSS 19.0 program. Results : 1. The students who had good personal relationships(p=.000) and who chose their major and school based on their aptitude(p=.029) were found to have a higher degree of satisfaction than those who did not have good personal relationships and who did not choose their major and school based on their aptitude. 2. Self-confidence(p=.004), self-control efficacy(p=.000), mastery approach(p=.000), and performance approach(p=.003) were found to have positive correlations with major adaptation while task difficulty preference(p=.000) was found to have a negative correlation with the same. 3. Self-confidence(p=.004), self-control efficacy(p=.000), mastery approach(p=.000), and performance approach(p=.045) were also found to have a positive correlation with major satisfaction while task difficulty preference(p=.000) was found to have a negative correlation with the same. 4. Mastery approach(p=.000) in goal achievement orientation was found to have a positive effect on self-efficacy while mastery avoidance(p=.001) was found to have a negative effect on the same. Mastery avoidance showed the greatest effect on self-confidence(p=.000) while mastery approach showed the greatest effect on self-control efficacy(p=.000) and task difficulty preference(p=.000). 5. In goal achievement orientation, mastery approach(p=.000) was found to have a positive effect on major adaptation while mastery avoidance was found to have a negative effect on the same. Meanwhile, with regard to self-efficacy, self-control efficacy(p=.037) and task difficulty preference(p=.008) were found to have positive effects on major adaptation. 6. In the case of major satisfaction, mastery approach(p=.000) in goal achievement orientation was found to have a positive effect on it, and in self-efficacy, self-control efficacy(p=.038) and task difficulty preference(p=.002) were found to have positive effects on it. Conclusions : Accordingly, to raise the level of the dental-hygiene students' adaptation to and satisfaction with their major, the students should be given appropriate tasks so that they can raise their mastery goal, and they should be encouraged to raise their approach goal rather than their avoidance goal. In addition, the students should be provided assistance in setting self-expectations and in developing confidence in their capabilities.

A New Speaker Adaptation Technique using Maximum Model Distance

  • Tahk, Min-Jea
    • 제어로봇시스템학회:학술대회논문집
    • /
    • 제어로봇시스템학회 2001년도 ICCAS
    • /
    • pp.154.2-154
    • /
    • 2001
  • This paper presented a adaptation approach based on maximum model distance (MMD) method. This method shares the same framework as they are used for training speech recognizers with abundant training data. The MMD method could adapt to all the models with or without adaptation data. If large amount of adaptation data is available, these methods could gradually approximate the speaker-dependent ones. The approach is evaluated through the phoneme recognition task on the TIMIT corpus. On the speaker adaptation experiments, up to 65.55% phoneme error reduction is achieved. The MMD could reduce phoneme error by 16.91% even when ...

  • PDF

콜센터 상담사의 직무성과 및 이직의도에 영향을 미치는 요인 : 개인특성과 조직특성을 중심으로 (Factors Affecting Job Performance and Turnover Intention of Call Center Representatives : Focusing on Individual Characteristics and Organizational Characteristics)

  • 정경숙;곡민
    • 한국IT서비스학회지
    • /
    • 제19권6호
    • /
    • pp.55-82
    • /
    • 2020
  • This study examined the factors that influence the turnover intention, job performance of call center representatives based on the adaptive structuration theory (AST). This study intended to empirically examine how individual characteristics of representative affect the technological and task adaptation, how they affect job performance and turnover intention. On the other hand, this study also explains how rational culture and organization a reputation which are considered as dimensions of organizational characteristics affects organizational commitment, and verifies the relationship between organizational commitment and job performance and turnover intention. Finally this paper aim to provide academic and practical implications. In order to solve the above research problems, this research proposed a model based on the adaptive structuration theory. In order to identify the relationship between the proposed variables and the AST for individual, we conducted an empirical test on the call center representatives. The structural equation model was used to verify the research model and hypotheses. The results of the empirical analysis show that the personal characteristics of counselors, such as communication skills, multitasking abilities, and innovativeness have a positive effect on skill adaptation, and skill adaptation has a positive effect on task adaptation, furthermore, it influences on job performance and turnover intention Respectively. In addition, among the factors of organizational environmental dimensions of the call center, it was found that organizational reputation not only increase continuance commitment but also increase normative commitment. Contrary to our expectations, perceived rational culture didn't have a positive effect on organizational commitment. Also, continuance commitment and normative commitment are valid predictors of job performance, but they have nothing to do with turnover intention. On the contrary, emotional commitment is the only one variable among three dimensions of organizational commitment have a positive effect on turnover intention, but is not a valid predictor of job performance.

Cross-layer Design of Rate and Quality Adaptation Schemes for Wireless Video Streaming

  • Lee, Sun-Hun;Chung, Kwang-Sue
    • KSII Transactions on Internet and Information Systems (TIIS)
    • /
    • 제4권3호
    • /
    • pp.324-340
    • /
    • 2010
  • Video streaming service over wireless networks is a challenging task because of the changes in the wireless channel conditions that can occur due to interference, fading, and station mobility. To provide an efficient wireless video streaming service, the rate adaptation scheme should improve wireless node performance and channel utilization. Moreover, the quality adaptation scheme should be considered at the streaming application. To meet these requirements, we propose a new cross-layer design for video streaming over wireless networks. This design includes the rate and quality adaptation schemes. The rate adaptation scheme selects the optimal transmission mode and resolves the performance anomaly problem. Based on performance improvement by the proposed rate adaptation scheme, our quality adaptation scheme improves the quality of video streaming. Through performance evaluations, we prove that our cross-layer design improves the wireless channel utilization and the quality of video streaming.

가중치 분포 특성을 이용한 Eigenvoice 기반 고속화자적응 (Rapid Speaker Adaptation Based on Eigenvoice Using Weight Distribution Characteristics)

  • 박종세;김형순;송화전
    • 한국음향학회지
    • /
    • 제22권5호
    • /
    • pp.403-407
    • /
    • 2003
  • 최근 고속화자적응 기법으로 eigenvoice 방식이 많이 사용되고 있다. Eigenvoice 적응방식에서도 적응화자의 적응 데이터가 매우 적은 경우에는 적절한 가중치의 추정이 어렵기 때문에 적응 데이터가 어느 정도 많은 경우에 비해 인식성능 향상이 크지 않다. 본 논문에서는 적응 데이터가 적을 때의 성능향상을 위하여 eigenvoice의 가중치 분포 특성을 이용한 eigenvoice 기반 고속화자적응을 제안한다. PBW 452 데이터베이스를 사용한 어휘독립 단어인식 실험 결과에서 가중치 문턱치(threshold) 적용 방식을 사용하여 적응 데이터가 매우 적은 경우의 상대적인 성능 저조 문제를 완화시켰다. 적응단어를 단 1개만 사용한 경우 가중치 문턱치 적용 방식을 사용하여 단어 오인식률을 9-18% 정도 감소시켰다.

한국어 연결 숫자음 인식을 일한 최대 사후 Eigenvoice에 근거한 자기적응 기법 (Self-Adaptation Algorithm Based on Maximum A Posteriori Eigenvoice for Korean Connected Digit Recognition)

  • 김동국;전형배
    • 한국음향학회지
    • /
    • 제23권8호
    • /
    • pp.590-596
    • /
    • 2004
  • 본 논문에서는 한국어 연결 숫자음 인식을 위한 최대 사후 eigenvoice을 사용한 자기적응 기법을 제안한다. 제안된 최대 사후 eigenvoice 기법은 eigenvoice 계수에 대한 확률 밀도 함수를 가정함으로 구성된다. 제안된 알고리즘은 기존 eigenvoice 추정 과정에 선 분포 모델을 포함하는 일반적인 해를 제공하는 구조를 갖는다. 인식할 한 문장만을 사용하는 자기 적응 시스템을 위해 매우 강인한 특성을 갖는 최대 사후 eigenvoice 적응 기법을 사용하였다. 한국어 연결 숫자음에 대한 일련의 자기 적응 실험결과 제안된 알고리즘의 성능은 매우 적은 량의 적응 데이터에 대해 기존 eigenvoice 알고리즘에 비해 우수한 성능을 나타냈었다.

바이어스 보상과 차원별 Eigenvoice 모델 평균을 이용한 고속화자적응의 성능향상 (Performance Improvement of Rapid Speaker Adaptation Using Bias Compensation and Mean of Dimensional Eigenvoice Models)

  • 박종세;김형순;송화전
    • 한국음향학회지
    • /
    • 제23권5호
    • /
    • pp.383-389
    • /
    • 2004
  • 본 논문에서는 훈련 및 인식 환경이 다른 상황에서 eigenvoice 기반 고속화자적응의 성능향상을 위하여 바이어스 보상을 적용한 eigenvoice 적응방식과 차원별 eigenvoice 모델 평균 가중합 방식을 제안하였다. PBW 452 DB를 사용한 어휘독립 단어인식 실험 결과에서 적은 양의 적응데이터를 사용했을 때 제안된 방식이 기존의 eigenvoice 방식에 비하여 많은 성능향상을 얻을 수 있었다. 적응단어 수를 1개에서 50개로 변경시키면서 바이어스 보상을 적용한 eigenvoice 적응방식을 사용한 경우 기존 eigenvoice 방식보다 단어 오인식률이 약 22∼30% 감소하였다. 또한 차원별 eigenvoice 모델 평균을 이용한 eigenvoice 적응방식에서는 1개의 단어를 적응데이터로 사용했을 경우에 기존 eigenvoice 방식보다 단어 오인식률이 최고 41%까지 감소하였다.

Poly-encoder기반의 COVID-19 질의 응답 태스크 (Poly-encoder based COVID-19 Question and Answering with Task Adaptation)

  • 이설화;임희석
    • 한국정보과학회 언어공학연구회:학술대회논문집(한글 및 한국어 정보처리)
    • /
    • 한국정보과학회언어공학연구회 2020년도 제32회 한글 및 한국어 정보처리 학술대회
    • /
    • pp.188-191
    • /
    • 2020
  • 본 연구는 COVID-19 질의 응답 태스크를 위한 Poly-encoder 기반의 태스크를 제안하였다. COVID-19 질의 응답 시스템은 사람들에게 최신 정보에 대해 빠르고 신뢰성이 높은 정보를 전달하는 특성을 가져야한다. 검색 기반 질의 응답 시스템은 pairwise 연산을 기반으로 수행되는데, Poly-encoder는 사전 학습된 트랜스포머(transformer)기반의 pairwise 연산 방법론 중 기존 Cross-encoder와 Bi-encoder보다 실사용 및 성능이 뛰어남을 보였다 [1]. 특히, Poly-encoder는 정확도가 높으면서도 빠른 응답속도를 가지며 검색기반의 각종 태스크에서 좋은 성능을 보였다. 따라서 본 연구는 COVID-19를 위한 Poly-encoder기반의 질의 응답 태스크를 위하여 기존 질의 응답 태스크와 페르소나 기반의 질의 응답 태스크로 두 가지 유형의 태스크를 생성하여 모델을 학습하였다. 또한 신뢰성 있는 리소스정보로부터 모델에 최신 정보 반영을 위하여 자동 크롤러를 구축하여 데이터를 수집하였다. 마지막으로 전문가를 통한 데이터셋을 구축하여 질문-응답과 질의어-질문에 대한 모델 검증을 수행하였다.

  • PDF