• Phonetics and Speech Sciences
• /
• v.6 no.3
• /
• pp.155-164
• /
• 2014

This paper proposes a new method to determine the recognition units for large vocabulary continuous speech recognition (LVCSR) in Korean by applying unsupervised segmentation and merging. In the proposed method, a text sentence is segmented into morphemes and position information is added to morphemes. Then submorpheme units are obtained by splitting the morpheme units through the maximization of posterior probability terms. The posterior probability terms are computed from the morpheme frequency distribution, the morpheme length distribution, and the morpheme frequency-of-frequency distribution. Finally, the recognition units are obtained by sequentially merging the submorpheme pair with the highest frequency. Computer experiments are conducted using a Korean LVCSR with a 100k word vocabulary and a trigram language model obtained by a 300 million eojeol (word phrase) corpus. The proposed method is shown to reduce the out-of-vocabulary rate to 1.8% and reduce the syllable error rate relatively by 14.0%.

• Korean Management Science Review
• /
• v.32 no.2
• /
• pp.125-143
• /
• 2015

When studying the film industry, researchers have seldom addressed the dynamic interaction between marketing information and word of mouth in the motion picture industry mainly because of the limitation of traditional research methodologies. This study explores integration and competition among important variables influencing on audience's choice on movie selection, particularly by using a new method of agent-based modeling including competitive environment. Decision process of moviegoer composed of transition probability based on multinomial logit model, considering marketing and box-office information, critique, and word of mouth from other moviegoers. After validating the fitness of market share among released movies, this study conducted a set of simulation experiments considering several variables such as market size, change of weight between variables, and movie performance under competition. Propositions are derived from the simulation results is also suggested for future research.

• The Journal of the Acoustical Society of Korea
• /
• v.13 no.4
• /
• pp.40-49
• /
• 1994

In this paper, we propose methods for improving the performance of word recognition system. The ray stratey of the first method is to apply the inertia to the feature vector sequences of speech signal to stabilize the transitions between VQ cdoes. The second method is generating the new observation probabilities using the transition matrix of VQ codes as weights at the observation probability of the output symbol, so as to take into account the time relation between neighboring frames in DHMM. By applying the inertia to the feature vector sequences, we can reduce the overlapping of probability distribution of the response paths for each word and stabilize state transitions in the HMM. By using the transition matrix of VQ codes as weights in conventional DHMM. we can divide the probability distribution of feature vectors more and more, and restrict the feature distribution to a suitable region so that the performance of recognition system can improve. To evaluate the performance of the proposed methods, we carried out experiments for 50 DDD area names. As a result, the proposed methods improved the recognition rate by $4.2\%$ in the speaker-dependent test and $12.45\%$ in the speaker-independent test, respectively, compared with the conventional DHMM.

• The Journal of the Acoustical Society of Korea
• /
• v.20 no.3
• /
• pp.56-62
• /
• 2001

In this paper, we adopt a missing data theory to speech recognition. It can be used in order to maintain high performance of speech recognizer when the missing data occurs. In general, hidden Markov model (HMM) is used as a stochastic classifier for speech recognition task. Acoustic events are represented by continuous probability density function in continuous density HMM(CDHMM). The missing data theory has an advantage that can be easily applicable to this CDHMM. A marginalization method is used for processing missing data because it has small complexity and is easy to apply to automatic speech recognition (ASR). Also, a spectral subtraction is used for detecting missing data. If the difference between the energy of speech and that of background noise is below given threshold value, we determine that missing has occurred. We propose a new method that examines the reliability of detected missing data using voicing probability. The voicing probability is used to find voiced frames. It is used to process the missing data in voiced region that has more redundant information than consonants. The experimental results showed that our method improves performance than baseline system that uses spectral subtraction method only. In 452 words isolated word recognition experiment, the proposed method using the voicing probability reduced the average word error rate by 12％ in a typical noise situation.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.11 no.3
• /
• pp.200-207
• /
• 2001

In this paper, we propose an HMM(Hidden Markov modeJ)-MLP(Multi-layer perceptron) hybrid model for recognizing legal words on the English bank check. We adopt an explicit segmentation-based word level architecture to implement an HMM engine with nonscaled and non-normalized symbol vectors. We also introduce an MLP for implicit segmentation-based word recognition. The final recognition model consists of a hybrid combination of the HMM and MLP with a new hybrid probability measure. The main contributions of this model are a novel design of the segmentation-based variable length HMMs and an efficient method of combining two heterogeneous recognition engines. ExperimenLs have been conducted using the legal word database of CENPARMI with encouraging results.

• Proceedings of the KIEE Conference
• /
• 1988.07a
• /
• pp.194-197
• /
• 1988

The unequal bit-error-probability of rate r=b/n binary convolutional code is analyzed. The error protection affored each digit of the b-tuple information word can be different from that afforded other digit. The property of the unequal protection can be applied to transmitting sampled data in PCM system.

• Journal of the military operations research society of Korea
• /
• v.22 no.2
• /
• pp.200-214
• /
• 1996

In this paper, the problem of determining the probability of kill(or survival) of a vulnerable target by one or more missiles is considered. The general formulas are obtained for the kill or survival probability the target is killed or survival. Several well-known concepts such as those of vulnerability, lethality, multi-component target, and a general combinatorial theorem of probability are introduced and used. For the convenience in this paper, the word missile is used in a very general sense and the target is generally taken to be a point target. And, this paper, is concentrated primarily with the probabilistic aspects of the problem, also a general numerical procedures are also described. Two examples are shown to illustrate the use of some of the formulas in this study, but also illustrate a few points which may not have been sufficiently emphasized. The extension study to complete a software package will be followed.

• Journal of Service Research and Studies
• /
• v.14 no.2
• /
• pp.1-17
• /
• 2024

One of the most important decisions for managers in the online banner advertising industry, is to choose the best banner alternative for exposure to customers. Since it is difficult to know the click probability of each banner alternative in advance, managers must experiment with multiple alternatives, estimate the click probability of each alternative based on customer clicks, and find the optimal alternative. In this reinforcement learning process, the main decision problem is to find the optimal balance between the level of exploitation strategy that utilizes the accumulated estimated click probability information and exploration strategy that tries new alternatives to find potentially better options. In this study we analyze the impact of word-of-mouth effects and the number of alternatives on the optimal exploration-exploitation strategies. More specifically, we focus on the word-of-mouth effect, where the click-through rate of the banner increases as customers promote the related product to those around them after clicking the exposed banner, and add it to the overall reinforcement learning process. We analyze our problem by employing the Multi-Armed Bandit model, and the analysis results show that the larger the word-of-mouth effect and the fewer the number of banner alternatives, the higher the optimal exploration level of advertising reinforcement learning. We find that as the probability of customers clicking on the banner increases due to the word-of-mouth effect, the value of the previously accumulated estimated click-through rate knowledge decreases, and therefore the value of exploring new alternatives increases. Additionally, when the number of advertising alternatives is small, a larger increase in the optimal exploration level was observed as the magnitude of the word-of-mouth effect increased. This study provides meaningful academic and managerial implications at a time when online word-of-mouth and its impact on society and business is becoming more important.

• Journal of Pharmacopuncture
• /
• v.11 no.4
• /
• pp.59-64
• /
• 2008

Objectives The aim of study was to evaluate the effect of Herbal medicine on post stroke cognitive deficit. Methods All groups were treated with acupunture treatment, moxa treatment, herbal medicines, physical and occupational therapy for 4 weeks, additionally cardiotonic pills(CP) were taken in the cardiotonic pills group. The effect of treatment was assessed using Verval fluency, MMSE-KC, Word List Immediate Recall test. Statistical significance was achived if the probability was less than 5%(p,0.05). Results Verval fluency, MMSE-KC, Word List Immediate Recall test scores increased in both group. MMSE-KC, Word List Immediate Recall test scores were significantly increased in the CP group. Verval fluency, MMSE-KC, Word List Immediate Recall test scores were significantly increased in the control group. In the Verval fluency, MMSE-KC, Word List Immediate Recall test of the CP group more increased compared to the control group. There were no significant differences between two groups. In the CP group, the scores of the infarction group more increased compared to the hemorrhage group. Conclusions According to the these results, herbal medicines are effective to improve post stroke cognitive-deficit. Futher studies are needed to know cardiotonic pills in the ischemic stroke.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.12 no.6
• /
• pp.2470-2491
• /
• 2018

Web-scale open information extraction (Open IE) plays an important role in NLP tasks like acquiring common-sense knowledge, learning selectional preferences and automatic text understanding. A large number of Open IE approaches have been proposed in the last decade, and the majority of these approaches are based on supervised learning or dependency parsing. In this paper, we present a novel method for web scale open information extraction, which employs cosine distance based on Google word vector as the confidence score of the extraction. The proposed method is a purely unsupervised learning algorithm without requiring any hand-labeled training data or dependency parse features. We also present the mathematically rigorous proof for the new method with Bayes Inference and Artificial Neural Network theory. It turns out that the proposed algorithm is equivalent to Maximum Likelihood Estimation of the joint probability distribution over the elements of the candidate extraction. The proof itself also theoretically suggests a typical usage of word vector for other NLP tasks. Experiments show that the distance-based method leads to further improvements over the newly presented Open IE systems on three benchmark datasets, in terms of effectiveness and efficiency.