• Journal of information and communication convergence engineering
• /
• v.21 no.2
• /
• pp.139-144
• /
• 2023

Engineers prefer deep neural networks (DNNs) for solving computer vision problems. However, DNNs pose two major problems. First, neural networks require large amounts of well-labeled data for training. Second, the covariate shift problem is common in computer vision problems. Domain adaptation has been proposed to mitigate this problem. Recent work on adversarial-learning-based unsupervised domain adaptation (UDA) has explained transferability and enabled the model to learn robust features. Despite this advantage, current methods do not guarantee the distinguishability of the latent space unless they consider class-aware information of the target domain. Furthermore, source and target examples alone cannot efficiently extract domain-invariant features from the encoded spaces. To alleviate the problems of existing UDA methods, we propose the mixup regularization in adversarial discriminative domain adaptation (ADDA) method. We validated the effectiveness and generality of the proposed method by performing experiments under three adaptation scenarios: MNIST to USPS, SVHN to MNIST, and MNIST to MNIST-M.

• Journal of Communications and Networks
• /
• v.9 no.2
• /
• pp.118-127
• /
• 2007

We propose a link adaptation and selection method for the links constituting an orthogonal frequency division multiplexing (OFDM) based wireless relay network. The proposed link adaptation and selection method selects the forwarding, modulation, and channel coding schemes providing the highest end-to-end throughput and decides whether to use the relay or not. The link adaptation and selection is done for each sub-channel based on instantaneous signal to interference plus noise ratio (SINR) conditions in the source-to-destination, source-to-relay and relay-to-destination links. The considered forwarding schemes are amplify and forward (AF) and simple adaptive decode and forward (DF). Efficient adaptive modulation and coding decision rules are provided for various relaying schemes. The proposed end-to-end link adaptation and selection method ensures that the end-to-end throughput is always larger than or equal to that of transmissions without relay and non-adaptive relayed transmissions. Our evaluations show that over the region where relaying improves the end-to-end throughput, the DF scheme provides significant throughput gain over the AF scheme provided that the error propagation is avoided via error detection techniques. We provide a frame structure to enable the proposed link adaptation and selection method for orthogonal frequency division multiple access (OFDMA)-time division duplex relay networks based on the IEEE 802.16e standard.

• Proceedings of the KSPS conference
• /
• 2003.05a
• /
• pp.127-130
• /
• 2003

This paper describes an efficient method for unsupervised speaker adaptation. This method is based on selecting a subset of speakers who are acoustically close to a test speaker, and calculating adapted model parameters according to the previously stored sufficient HMM statistics of the selected speakers' data. In this method, only a few unsupervised test speaker's data are required for the adaptation. Also, by using the sufficient HMM statistics of the selected speakers' data, a quick adaptation can be done. Compared with a pre-clustering method, the proposed method can obtain a more optimal speaker cluster because the clustering result is determined according to test speaker's data on-line. Experiment results show that the proposed method attains better improvement than MLLR from the speaker independent model. Moreover the proposed method utilizes only one unsupervised sentence utterance, while MLLR usually utilizes more than ten supervised sentence utterances.

• Journal of dental hygiene science
• /
• v.21 no.1
• /
• pp.19-27
• /
• 2021

Background: The purpose of this study was to investigate the effects of dental hygiene students' communication ability and conflict management types on adaptation to college life improve adaptation to college life and to suggest strategies to play a role as a professional dental hygienist in clinical practice. Methods: The subjects of this study were selected by random sample extraction of dental hygiene students from two colleges in Gyeonggi and one Chungcheong area. The survey was conducted online for dental hygiene students from July 20 to October 31, 2020. In order to comply with research ethics, this study was investigated with the consent of the subjects. The sample was analyzed for the final 351. Results: Dental hygiene students' communication ability was 3.74 points, conflict management types was 3.01 points, and adaptation to college life was 3.30 points. In the type of conflict management according to the general characteristics, the problem-solving method, the avoidance method, the compromise method, and the concession method showed statistically significant differences in major satisfaction. Among the types of conflict management, the highest positive correlation (r=0.613) was shown in the relationship between the problem-solving method and the compromise method. Factors influencing the adaptation to college life were concentration, forced method, self-disclosure, conversational coherence, and social relaxation. Conclusion: This study is significant in that it derives sub-areas of communication ability and conflict management types that influence adaptation to college life. In the future, there is a need to improve the quality of dental hygienists by continuously conducting research on the types of conflict management targeting dental hygienists, and responding more proactively and actively to conflict situations in the clinical field.

• ETRI Journal
• /
• v.30 no.6
• /
• pp.865-867
• /
• 2008

A conventional environment adaptation for robust speech recognition is usually conducted using transform-based techniques. Here, we present a discriminative adaptation strategy based on a multi-condition-trained model, and propose a new method to provide universal application to a new environment using the environment's specific conditions. Experimental results show that a speech recognition system adapted using the proposed method works successfully for other conditions as well as for those of the new environment.

• MALSORI
• /
• v.55
• /
• pp.93-102
• /
• 2005

In this paper, sub-stream based eigenvoice method is proposed to overcome the weak points of conventional eigenvoice and dimensional eigenvoice. In the proposed method, sub-streams are automatically constructed by the statistical clustering analysis that uses the correlation information between dimensions. To obtain the reliable distance matrix from covariance matrix for dividing into optimal sub-streams, MAP adaptation technique is employed to the covariance matrix of training data and the sample covariance of adaptation data. According to our experiments, the proposed method shows $41\%$ error rate reduction when the number of adaptation data is 50.

• Proceedings of the KSPS conference
• /
• 2007.05a
• /
• pp.33-36
• /
• 2007

In this paper, we investigate a fast speaker adaptation method based on eigenvoice in several noisy environments. In order to overcome its weakness against noise, we propose a noisy environment clustering method which divides the noisy adaptation utterances into utterance groups with similar environments by the vector quantization based clustering using a cepstral mean as a feature vector. Then each utterance group is used for adaptation to make an environment dependent model. According to our experiment, we obtained 19-37 % relative improvement in error rate compared with the simultaneous speaker adaptation and environmental compensation method

• Journal of the Korean Institute of Telematics and Electronics B
• /
• v.32B no.5
• /
• pp.765-776
• /
• 1995

This paper describes a speaker adaptation method to improve the recognition performance of MLP(multiLayer Perceptron) based HMM(Hidden Markov Model) speech recognizer. In this method, we use lst-order linear transformation network to fit data of a new speaker to the MLP. Transformation parameters are adjusted by back-propagating classification error to the transformation network while leaving the MLP classifier fixed. The recognition system is based on semicontinuous HMM's which use the MLP as a fuzzy vector quantizer. The experimental results show that rapid speaker adaptation resulting in high recognition performance can be accomplished by this method. Namely, for supervised adaptation, the error rate is signifecantly reduced from 9.2% for the baseline system to 5.6% after speaker adaptation. And for unsupervised adaptation, the error rate is reduced to 5.1%, without any information from new speakers.

• Korean Journal of Health Education and Promotion
• /
• v.3 no.1
• /
• pp.43-55
• /
• 1985

The recent problems of adolescents are leading us to amphasize school mental health program. Therefore, for the first step of that program, this study was attempted to find out the students' daily life stress, their adaptation beharior to resolve that stress, and their self-concept which has an effect on that beharior, to analyze the relationship among those three variables, and to provide basic data which help maintain and promote students' adaptability and mental health. The subjects were 4 classes each grade, 743 students who were chosen as a sample subject among the total of 3,250 students of a junior girls' school in Seoul by stratified sampling method. The Oquestionnair surrey was done from April 1, 1985 to April 6, 1985. The instruments were (1) a part of Junghoon chois' perceptual Orientation Scale for measuring self-perception of students, (2) Rosenberg's Questionnair for measuring students' evaluation of self-esteem, (3) modified Bell's Adoptive Behavior Questionnair, and (4) the Measuring scale for stress developed by the investigator. The collected data were analyzed by using descriptive statistics, x²-test, t-test, ANOVA, and Pearsons' Correlation. The results are summarized as follows: 1. The range of mean score of self-concept, stress, and adaptation beharior is 1-5 respectively. The mean score of self concept was 3.45. The mean score of stress was 2.53, and the degree of stress was in order of the problem of study and grade marks, girl friend, appearance, and relaxation. The subjects preferred effective long-term adaptation method (Mean=2.81) to ineffective short-term adaptation method. (Mean=2.47) 2. The self-concept had the highest degree when students were 1st grade, in good health condition, taking high grade marks, both parents existing, fathers' educational level being above college, and being in high economic status. The degree of stress was significantly high when students were 3rd grade, in poor health condition, having low grade marks, being in poor economic status, and fathers' educational level being under elimentary school. In relation with adaptation method, 1st grade students used more sbort-term and long-term method, and when students had high grade marks, being in high economic status. they used more long-term adaptation behavior. 3. Two Hypotheses of this research were tested: Hypothesis 1. "The higher the students' self-concept is, the more they use long-term adaptation method." was accepted. (r=.2482, p<.01) (t=-4.99, p<.001) Hypothesis 2. "The higher the students' stress is, the more they use short-term adaptation method." was accepted. (r=.2897, p<.01) (t= -7.51, p<.001) In conclusion, a person in charge of school health can help students to adapt effectively by planning and implementing methods which highten their self-concept and lower their stress. But because the correlation among self-concept, stress and adaptation behavior was at a low level, I think that the study, which confirm the relationship of those three variables, is needed.

• The Journal of the Acoustical Society of Korea
• /
• v.20 no.2
• /
• pp.21-26
• /
• 2001

An expectation-maximization (EM) based Bayesian adaptation method for the mean of noise is proposed for noise-robust speech recognition. In the algorithm, the on-line testing utterances are used for the unsupervised Bayesian adaptation and the prior distribution of the noise mean is estimated using the off-line training data. For the noisy speech modeling, the parallel model combination (PMC) method is employed. The proposed method has shown to be effective compared with the conventional PMC method for the speech recognition experiments in a car-noise condition.