• Korean Journal of Cognitive Science
• /
• v.18 no.1
• /
• pp.1-14
• /
• 2007

In a goal-oriented dialogue, speakers' intentions can be represented by domain actions that consist of pairs of a speech act and a concept sequence. Therefore, if we plan to implement an intelligent dialogue system, it is very important to correctly infer the domain actions from surface utterances. In this paper, we propose a statistical model to determine speech acts and concept sequences using conditional random fields at the same time. To avoid biased learning problems, the proposed model uses low-level linguistic features such as lexicals and parts-of-speech. Then, it filters out uninformative features using the chi-square statistic. In the experiments in a schedule arrangement domain, the proposed system showed good performances (the precision of 93.0% on speech act classification and the precision of 90.2% on concept sequence classification).

• Journal of Korea Multimedia Society
• /
• v.22 no.12
• /
• pp.1457-1465
• /
• 2019

3D convolution is to stack multiple consecutive frames to form a cube, and then apply the 3D convolution kernel in the cube. In this structure, each feature map of the convolutional layer is connected to multiple adjacent sequential frames in the previous layer, thus capturing the motion information. However, due to the changes of pedestrian posture, motion and position, the convolution at the same place is inappropriate, and when the 3D convolution kernel is convoluted in the time domain, only time domain features of three consecutive frames can be extracted, which is not a good enough to get action information. This paper proposes an action recognition method based on feature fusion of 3D convolutional neural network. Based on the VGG16 network model, sending a pre-acquired optical flow image for learning, then get the time domain features, and then the feature of the time domain is extracted from the features extracted by the 3D convolutional neural network. Finally, the behavior classification is done by the SVM classifier.

• ETRI Journal
• /
• v.39 no.4
• /
• pp.502-513
• /
• 2017

For video understanding, namely analyzing who did what in a video, actions along with objects are primary elements. Most studies on actions have handled recognition problems for a well-trimmed video and focused on enhancing their classification performance. However, action detection, including localization as well as recognition, is required because, in general, actions intersect in time and space. In addition, most studies have not considered extensibility for a newly added action that has been previously trained. Therefore, proposed in this paper is an extensible hierarchical method for detecting generic actions, which combine object movements and spatial relations between two objects, and inherited actions, which are determined by the related objects through an ontology and rule based methodology. The hierarchical design of the method enables it to detect any interactive actions based on the spatial relations between two objects. The method using object information achieves an F-measure of 90.27%. Moreover, this paper describes the extensibility of the method for a new action contained in a video from a video domain that is different from the dataset used.

• Journal of Korean Academic Society of Home Health Care Nursing
• /
• v.6
• /
• pp.19-32
• /
• 1999

The purpose of this study was to develop of home nursing care classification and home health care costs of the free-standing home nursing care agency. This study was done through 3 steps The First stage, home nursing care classification was identified and classified by literature, review-committee and expert meeting. The second stage, cost elements for home nursing care visit were identified and accounted. That were divided into direct nursing care cost, indirect nursing care cost, management cost and transportation cost. Third stage, total cost of per visit was produced. Data were collected from 810 visits of 120 patients received home dare and from January. 1999 to November, 1999, and analysed with EXCEL program. The obtained results are as follows : 1. Home nursing care classification was consisted of 6 high level classification domain and 10 low level classification domain and 163 home nursing care behavior. 2. The cost of home nursing care per visit was 30,638 won which were direct and indirect nursing care cost(16.305won), management cost(5,255won) and transportation cost (9,098won). In conclusion. Home nursing behavior care classification developed in this study would be used as home health care standard. And the home nursing care costs can be used as a fundamental data for the further development of home health care costs in Korea.

• Journal of Convergence for Information Technology
• /
• v.9 no.9
• /
• pp.1-12
• /
• 2019

A dialog system becomes a new way of communication between human and computer. The dialog system takes human voice as an input, and gives a proper response in voice or perform an action. Although there are several well-known products of dialog system (e.g., Amazon Echo, Naver Wave), they commonly suffer from a problem of out-of-domain utterances. If it poorly detects out-of-domain utterances, then it will significantly harm the user satisfactory. There have been some studies aimed at solving this problem, but it is still necessary to study about this intensively. In this paper, we give an overview of the previous studies of out-of-domain detection in terms of three point of view: dataset, feature, and method. As there were relatively smaller studies of this topic due to the lack of datasets, we believe that the most important next research step is to construct and share a large dataset for dialog system, and thereafter try state-of-the-art techniques upon the dataset.

• Journal of Internet Computing and Services
• /
• v.22 no.6
• /
• pp.99-113
• /
• 2021

In these days, as cyberspace has been recognized as the fifth battlefield area following the land, sea, air, and space, attention has been focused on activities that view cyberspace as an operational and mission domain in earnest. Also, in the 21st century, cyber operations based on cyberspace are being developed as a 4th generation warfare method. In such an environment, the success of the operation is determined by the commander's decision. Therefore, in order to increase the rationality and objectivity of such decision-making, it is necessary to systematically establish and select a course of action (COA). In this study, COA is established by using the method of classifying operational elements necessary for cyber operation, and it is intended to suggest a direction for quantitative evaluation of COA. To this end, we propose a method of composing the COES (Cyber Operational Elements Set), which becomes the COA of operation, and classifying the cyber operational elements identified in the target development process based on the 5W1H Method. In addition, by applying the proposed classification method to the cyber operation elements used in the STUXNET attack case, the COES is formed to establish the attack COAs. Finally, after prioritizing the established COA, quantitative evaluation of the policy was performed to select the optimal COA.

• Journal of KIISE:Software and Applications
• /
• v.35 no.3
• /
• pp.189-196
• /
• 2008

To build autonomous agents who can make a decision on behalf of humans in time-critical complex environments, the formulation of operational knowledge base could be essential. This paper proposes the methodology of how to formulate the knowledge base and evaluates it in a practical application domain. We analyze threat data received from the multiple sensors of Aircraft Survivability Equipment(ASE) for Korean helicopters, and integrate the threat data into the inductive model through compilation technique which extracts features of the threat data and relations among them. The compiled protocols of state-action rules can be implemented as the brain of the ASE. They can reduce the amounts of reasoning, and endow the autonomous agents with reactivity and flexibility. We report experimental results that demonstrate the distinctive and predictive patterns of threats in simulated battlefield settings, and show the potential of compilation methods for the successful detection of threat systems.

• Journal of IKEEE
• /
• v.23 no.2
• /
• pp.727-734
• /
• 2019

This paper confirms whether the movement or specific operation of the muscles in the process of transferring a person from the brain can find a signal showing an essential feature of a certain part of the brain. As a rule, the occurrence of EEG(Electroencephalogram) changes when a signal is received from a specific action or from an induced action. These signals are very vague and difficult to distinguish from the naked eye. Therefore, it is necessary to define a signal for analysis before classification. The EEG form can be divided into the alpha, beta, delta, theta and gamma regions in the frequency ranges. The specific size of these signals does not reflect the exact behavior or intention, since the band or energy difference of the activated frequencies varies depending on the EEG measurement domain. However, if different actions are performed in a specific method, it is possible to classify the movement based on EEG activity and to determine the EEG tendency affecting the movement. Therefore, in this article, we first study the EEG expression pattern based on the activation of the left and right biceps EMG, and then we determine whether there is a significant difference between the EEG due to the activation of the left and right muscles through EEG. If we can find the EEG classification criteria in accordance with the EMG activation, it can help to understand the form of the transmitted signal in the process of transmitting signals from the brain to each muscle. In addition, we can use a lot of unknown EEG information through more complex types of brain signal generation in the future.