• The Journal of the Acoustical Society of Korea
• /
• v.39 no.5
• /
• pp.476-482
• /
• 2020

Sequence-to-sequence deep neural networks with attention mechanisms have shown superior performance across various domains, where the sizes of the input and the output sequences may differ. However, if the input sequences are much longer than the output sequences, and the characteristic of the input sequence changes within a single output token, the conventional attention mechanisms are inappropriate, because only a single context vector is used for each output token. In this paper, we propose a double-attention mechanism to handle this problem by using two context vectors that cover the left and the right parts of the input focus separately. The effectiveness of the proposed method is evaluated using speech recognition experiments on the TIMIT corpus.

• Journal of Communications and Networks
• /
• v.6 no.1
• /
• pp.78-87
• /
• 2004

This paper presents deterministic, worst-case analysis of a queueing system whose multiple homogeneous input streams are regulated by the associated leaky buckets and the queueing system that has a single stream regulated by the jumping-window. Queueing delay averaged over all items is used for performance measure, and the worst-case input traffic and the worst-case performance are identified for both queueing systems. For the former queueing system, the analysis explores different phase relations among leaky-bucket token generations. This paper observes how the phase differences among the leaky buckets affect the worst-case queueing performance. Then, this paper relates the worst-case performance of the former queueing system with that of the latter (the single stream case, as in the aggregate streams from many users, whose item arrivals are regulated by one jumping-window). It is shown that the worst-case performance of the latter is identical to that of the former in which all leaky buckets have the same phase and have particular leaky bucket parameters.

• Journal of information and communication convergence engineering
• /
• v.1 no.3
• /
• pp.123-128
• /
• 2003

This paper has addressed the implementation of the single rate Rate Adaptive Shaper(srRAS) described in RFC2963. This shaper has been proposed to use at the ingress of differentiated services networks providing Assured Forwarding Per Hop Behavior (AFPHB). srRAS is typically used in conjunction with single rate Three Color Marker(srTCM) described in RFC2697. srRAS itself is the tail-drop FIFO that is drained at a variable rate, and srTCM is the marker with metering function. G-srRAS is the same as srRAS except that RAS receives the green token state information from the downstream srTCM to avoid delaying a packet in RAS although there are sufficient tokens available to color the packet green. In this paper, we have addressed the algorithm and the architecture of srRAS, and the scheme to implement srRAS using VHDL(Very high-speed integrated circuit Hardware Description Language) and its related tools.

• Journal of Korean Academy of Nursing
• /
• v.26 no.1
• /
• pp.5-14
• /
• 1996

The purpose of this study is to find the Korean women's view to pregnancy within the Korean cultural context, to help nurses understand psychosocioemotional schema of the pregnant women, and to contribute to theory-building on the meaning of pregnancy. The interviewees were 21 women in their twenties or thirties and they were selected by theoretical sampling technique. Thirteen women of the interviewees were single and others married. Data were collected by in-depth interviews from September November in 1994. Interviews were tape-recorded and transcribed. Data were analyzed by Van Kaam's phenomenological method. Results were as follows. Eighty two descriptive expression were found and they were grouped under fourteen common factors. These are 「to maintain family」, 「instinct」, 「obligation」, 「what one should do」, 「to make home」, 「to tie couple firmly」, 「means to overcome a period of lassitude」, 「token of love」, 「hope」, 「to avoid loneliness」, 「wanted to have a baby」, 「to provide for old ages」, 「to be a mother」, 「to achieve her desire through baby」. Finally, fourteen common factors were grouped under four higher categories. Five common factors, 「to maintain family」, 「instinct」, 「obligation」, 「what one should do」, 「to make home」 were grouped under〈custom〉. Three factors, 「to tie couple firmly」, 「means to overcome a period of lassitude」, 「token of love」 were grouped under 〈love between couple〉. Four factors, 「hope」, 「to avoid loneliness」, 「wanted to have a baby」, 「to provide for old ages」 were grouped under 〈pursuiting safety〉. Two factors, 「to be a mother」, 「to achieve her desire through baby」 were grouped under 〈self-fulfillment〉. At present, the most distinguishing motivation of childbearing was rooted in the cultural custom. Korean women thought that it was their duties to have children and children were necessary in life. But they were beginning to recognize that baby could tie couple more firmly and get rid of their loneliness. Korean women no longer considered their children as the one who would take care of them when they were old or who would satisfy their unmet needs. As pregnancy is not women's duty only, authors recommend further studies on men's view to pregnancy.

• Journal of the Korea Society of Computer and Information
• /
• v.26 no.8
• /
• pp.47-54
• /
• 2021

Given a document, keyphrase extraction is to automatically extract words or phrases which topically represent the content of the document. In unsupervised keyphrase extraction approaches, candidate words or phrases are first extracted from the input document, and scores are calculated for keyphrase candidates, and final keyphrases are selected based on the scores. Regarding the computation of the scores of candidates in unsupervised keyphrase extraction, this study proposes a method of adjusting the scores of keyphrase candidates according to the types of keyphrase candidates: word-type or phrase-type. For this, type-token ratios of word-type and phrase-type candidates as well as information content of high-frequency word-type and phrase-type candidates are collected from the input document, and those values are employed in adjusting the scores of keyphrase candidates. In experiments using four keyphrase extraction evaluation datasets which were constructed for full-text articles in English, the proposed method performed better than a baseline method and comparison methods in three datasets.

• IEMEK Journal of Embedded Systems and Applications
• /
• v.17 no.4
• /
• pp.199-208
• /
• 2022

Korean is an agglutinative language, and one or more morphemes are combined to form a single word. Part-of-speech tagging method separates each morpheme from a word and attaches a part-of-speech tag. In this study, we propose a new Korean part-of-speech tagging method based on the Head-Tail tokenization technique that divides a word into a lexical morpheme part and a grammatical morpheme part without decomposing compound words. In this method, the Head-Tail is divided by the syllable boundary without restoring irregular deformation or abbreviated syllables. Korean part-of-speech tagger was implemented using the Head-Tail tokenization and deep learning technique. In order to solve the problem that a large number of complex tags are generated due to the segmented tags and the tagging accuracy is low, we reduced the number of tags to a complex tag composed of large classification tags, and as a result, we improved the tagging accuracy. The performance of the Head-Tail part-of-speech tagger was experimented by using BERT, syllable bigram, and subword bigram embedding, and both syllable bigram and subword bigram embedding showed improvement in performance compared to general BERT. Part-of-speech tagging was performed by integrating the Head-Tail tokenization model and the simplified part-of-speech tagging model, achieving 98.99% word unit accuracy and 99.08% token unit accuracy. As a result of the experiment, it was found that the performance of part-of-speech tagging improved when the maximum token length was limited to twice the number of words.

• Journal of the Korea Society of Computer and Information
• /
• v.14 no.12
• /
• pp.1-8
• /
• 2009

Grid services offer SSO(single sign-on) mechanism using GSI(grid security infrastructure) based on X.509. However. portal applications in web environment use ID and password model for single sign-on. Grid portals means a system which provides grid services by integrating portlet contents on single web interface. In existing research such as GAMA and PURSE, SSO for a whole grid portal is figured out in the way that user is authenticated by ID and password in front and call grid service via GSI at back-end. Other types of web applications outside of portlet framework cannot unfortunately access grid service in SSO way in the existing researches, because the SSO mechanism is developed for portlet framework only. In this paper, we suggest a SSO mechanism based on ID and password model, which forwards authentication information and a GSI token for grid access among portlets and grid-enabled web applications. This mechanism is applied to MGrid for SSO, which consists of applications of java web start, applet, servlet, and etc. as also as portlets.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.6 no.9
• /
• pp.2405-2423
• /
• 2012

In a traditional Single Sign-On (SSO) scheme, the user and the Service Providers (SPs) have given their trust to the Identity Provider (IdP) or Authentication Service Provider (ASP) for the authentication and correct assertion. However, we still need a better solution for the local/native true SSO to gain user confidence, whereby the trusted entity must play the role of the ASP between distinct SPs. This technical gap has been filled by Trusted Computing (TC), where the remote attestation approach introduced by the Trusted Computing Group (TCG) is to attest whether the remote platform integrity is indeed trusted or not. In this paper, we demonstrate a Trustworthy Mutual Attestation (TMutualA) protocol as a proof of concept implementation for a local true SSO using the Integrity Measurement Architecture (IMA) with the Trusted Platform Module (TPM). In our proposed protocol, firstly, the user and SP platform integrity are checked (i.e., hardware and software integrity state verification) before allowing access to a protected resource sited at the SP and releasing a user authentication token to the SP. We evaluated the performance of the proposed TMutualA protocol, in particular, the client and server attestation time and the round trip of the mutual attestation time.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.21 no.10
• /
• pp.2620-2633
• /
• 1996

In this paper, a new high-speed local area network using a dual mode switching access (DMSA) protocol implemented on a dual unidirectional bus is described. By utilizing the implicit positionalordering of stations on a unidirectional bus, the proposed system switches between random access mode and the token access model withoug unnecessary delay. Therefore, unlike other hybrid systems such as Buzz-net and Z-net, DMSA does not show a rapid degradation in performance as the load increases. We obtain the average channel utilization and the average access delay by using a simplified analytic model. The numerical results obtained via analysis are compared to the simulation resuls for a partial validation of the approximate model. The performance characteristics of DMSA are superior delay-throughput characteristics at light and medium loads, compared to compared to other LAN systems, and the capability of providing a single active station with full capabity of the channel.

• Speech Sciences
• /
• v.11 no.1
• /
• pp.105-116
• /
• 2004

The fundamental frequency of a vowel is known to be affected by the identity of the preceding consonant. The general agreement is that strong consonants trigger higher F0 than weak consonants. However, there has been a disagreement on the shape of this segmentally affected F0 contours. Some studies report that shapes of contours are differentiated based on the consonant type, but others regard this observation as misleading. This research attempts to resolve this controversy by investigating shapes and slopes of F0 contours of Korean word level speech data produced by four male speakers. Instead of entirely relying on traditional human intuition and judgment, I employed an automatic F0 contour analysis technique known as tilt parameterisation (Taylor 2000). After necessary manipulation of an F0 contour of each data token, various parameters are collapsed into a single tilt value which directly indicates the shape of the contour. The result, in terms of statistical inference, shows that it is not viable to conclude that the type of consonant is significantly related to the shape of F0 contour. A supplementary measurement is also made to see if the slope of each contour bears meaningful information. Unlike shapes themselves, slopes are suspected to be practically more practical for consonantal differentiation, although confirmation is required through further refined experiments.