• KNOM Review
• /
• v.23 no.2
• /
• pp.29-41
• /
• 2020

Networks are growing faster than ever before causing a multi-domain complexity. The diversity, variety and dynamic nature of network traffic and services require enhanced orchestration and management approaches. While many standard orchestrators and network operators are resulting in an increase of complexity for handling E2E slice orchestration. Besides, there are multiple domains involved in E2E slice orchestration including access, edge, transport and core network each having their specific challenges. Hence, handling of multi-domain, multi-platform and multi-operator based networking environments manually requires specified experts and using this approach it is impossible to handle the dynamic changes in the network at runtime. Also, the manual approaches towards handling such complexity is always error-prone and tedious. Hence, this work proposes an automated and abstracted solution for handling E2E slice orchestration using an intent-based approach. It abstracts the domains from the operators and enable them to provide their orchestration intention in the form of high-level intents. Besides, it actively monitors the orchestrated resources and based on current monitoring stats using the machine learning it predicts future utilization of resources for updating the system states. Resulting in a closed-loop automated E2E network orchestration and management system.

• Informatization Policy
• /
• v.31 no.2
• /
• pp.3-38
• /
• 2024

Hallucination is a significant barrier to the utilization of large-scale language models or multimodal models. In this study, we collected 654 computer science papers with "hallucination" in the abstract from arXiv from December 2022 to January 2024 following the advent of Chat GPT and conducted frequency analysis, knowledge network analysis, and literature review to explore the latest trends in hallucination research. The results showed that research in the fields of "Computation and Language," "Artificial Intelligence," "Computer Vision and Pattern Recognition," and "Machine Learning" were active. We then analyzed the research trends in the four major fields by focusing on the main authors and dividing them into data, hallucination detection, and hallucination mitigation. The main research trends included hallucination mitigation through supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF), inference enhancement via "chain of thought" (CoT), and growing interest in hallucination mitigation within the domain of multimodal AI. This study provides insights into the latest developments in hallucination research through a technology-oriented literature review. This study is expected to help subsequent research in both engineering and humanities and social sciences fields by understanding the latest trends in hallucination research.

• Journal of the Korean Society for information Management
• /
• v.41 no.1
• /
• pp.261-282
• /
• 2024

This study proposed and evaluated electroencephalography (EEG)-based and eye-tracking-based methods to determine relevance by utilizing users' implicit relevance feedback while navigating content in a digital library. For this, EEG/eye-tracking experiments were conducted on 32 participants using video, image, and text data. To assess the usefulness of the proposed methods, deep learning-based artificial intelligence (AI) techniques were used as a competitive benchmark. The evaluation results showed that EEG component-based methods (av_P600 and f_P3b components) demonstrated high classification accuracy in selecting relevant videos and images (faces/emotions). In contrast, AI-based methods, specifically object recognition and natural language processing, showed high classification accuracy for selecting images (objects) and texts (newspaper articles). Finally, guidelines for implementing a digital library interface based on EEG, eye-tracking, and artificial intelligence technologies have been proposed. Specifically, a system model based on implicit relevance feedback has been presented. Moreover, to enhance classification accuracy, methods suitable for each media type have been suggested, including EEG-based, eye-tracking-based, and AI-based approaches.

• Journal of Animal Reproduction and Biotechnology
• /
• v.34 no.2
• /
• pp.106-110
• /
• 2019

Sexed sperm can contribute to increase the profitability of the cow industry through the production of offspring of the craved sex, such as males for meat or females for dairy production. Therefore, the utilization of sexed sperms plays a very important role in the production of offspring of superior cattle. In this study, we examined the pregnancy rates and calves sexing proportion of male and female calves produced using AI, both performed using sexed and conventional sperm. In the result, the conception rates after ET were 73.3% (33/45) sexed semen and 52% (55/104) conventional semen. Thus, the sex ratio for sexed-semen inseminations was 70% (21/30) females for singleton births within a 272 to 292 day gestation interval. The sex ratio for conventional semen was 61% (34/56) females for births. As a result, it is suggested that the use of sex classification sperm will play a very important role in the offspring production of Korean bovine.

• International journal of advanced smart convergence
• /
• v.9 no.3
• /
• pp.17-27
• /
• 2020

With the continuous development and application of Internet technology, in recent years, new technologies such as cloud computing, big data, the Internet of Things, and AI are becoming more and more familiar to the general public. The development of a digital society has entered a new period of development. In this paper, we used on the 2018 annual data of 50 listed companies with blockchain concept stocks in China. Using data envelopment analysis (DEA) to study and analyze the input-output efficiency, it can be concluded that the input-output efficiency of 50 listed companies is very different. Inefficient companies are as high as 62%. Most companies have a large room for improvement in input-output efficiency due to uneconomical scale or inefficient technology. In order to better improve the company's input-output efficiency, one must improve the efficiency of resource utilization, optimize the company's research and development costs and the input and management of technical personnel; the second is to increase technological innovation and business innovation.

• Journal of Korea Game Society
• /
• v.18 no.4
• /
• pp.75-82
• /
• 2018

With the development of information and communication technology, various methods have been tried to provide learners with a fun educational environment through fun and interest. It is a good example to utilize technologies such as games and robots in education for edutainment and game-based learning. In this study, we propose an intelligent rhythm education system using user data collection and analysis for humanoid robot rhythm generation. To do this, the user selects music and inputs rhythm information according to the selected music. The robot utilization data of this user extracts patterns through collection and analysis. Patterns are based on frequency, and FFT similarity comparison method is applied when past data is insufficient. The proposed method is validated through experiments of kindergarten children.

• Journal of the Korea Society of Computer and Information
• /
• v.24 no.3
• /
• pp.159-166
• /
• 2019

The collection and manipulation of Internet of Things (IoT) data is increasing at a fast pace and its importance is recognized in every sector of our society. For efficient utilization of IoT data, the vast and varied IoT data needs to be reliable and meaningful. In this paper, we propose an IoT framework to realize this need. The IoT framework is based on a four layer IoT architecture onto which context aware computing technology is applied. If the collected IoT data is unreliable it cannot be used for its intended purpose and the whole service using the data must be abandoned. In this paper, we include techniques to remove uncertainty in the early stage of IoT data capture and collection resulting in reliable data. Since the data coming out of the various IoT devices have different formats, it is important to convert them into a standard format before further processing, We propose the RDF format to be the standard format for all IoT data. In addition, it is not feasible to process all captured Iot data from the sensor devices. In order to decide which data to process and understand, we propose to use contexts and reasoning based on these contexts. For reasoning, we propose to use standard AI and statistical techniques. We also propose an experiment environment that can be used to develop an IoT application to realize the IoT framework.

• KEPCO Journal on Electric Power and Energy
• /
• v.7 no.1
• /
• pp.115-123
• /
• 2021

Through machine learning-based load prediction, it is possible to prevent excessive power generation or unnecessary economic investment by estimating the appropriate amount of facility investment in consideration of the load that will increase in the future or providing basic data for policy establishment to distribute the maximum load. However, in order to secure the reliability of the developed load prediction model in the field, the performance comparison verification between the distribution line load prediction models must be preceded, but a comparative performance verification system between the distribution line load prediction models has not yet been established. As a result, it is not possible to accurately determine the performance excellence of the load prediction model because it is not possible to easily determine the likelihood between the load prediction models. In this paper, we developed a reliability verification system for load prediction models including a method of comparing and verifying the performance reliability between machine learning-based load prediction models that were not previously considered, verification process, and verification result visualization methods. Through the developed load prediction model reliability verification system, the objectivity of the load prediction model performance verification can be improved, and the field application utilization of an excellent load prediction model can be increased.

• Ecology and Resilient Infrastructure
• /
• v.9 no.2
• /
• pp.124-129
• /
• 2022

Urban river have been a key pillar in citizens' lives through the healthy urban environment of ecological nature along with the function of this dimension. On the other hand, conflicts are occurring in terms of conservation and use of rivers in the region along with the expansion of infrastructure due to urban densification. Appropriate agreement and coordination are needed for sustainable streams. This study aims to analyze various opinions through public discussion of residents' proposals through the resident participation cooperation budget. User awareness surveys and expert interviews were conducted on six rivers in Eunpyeong-gu. Through this, stable and sustainable preservation of local and small rivers and appropriate and balanced use measures were proposed.

• Journal of information and communication convergence engineering
• /
• v.20 no.4
• /
• pp.288-294
• /
• 2022

The media content market has been growing, as various types of content are being mass-produced owing to the recent proliferation of the Internet and digital media. In addition, platforms that provide personalized services for content consumption are emerging and competing with each other to recommend personalized content. Existing platforms use a method in which a user directly inputs video metadata. Consequently, significant amounts of time and cost are consumed in processing large amounts of data. In this study, keyframes and audio spectra based on the YCbCr color model of a movie trailer were extracted for the automatic generation of metadata. The extracted audio spectra and image keyframes were used as learning data for genre recognition in deep learning. Deep learning was implemented to determine genres among the video metadata, and suggestions for utilization were proposed. A system that can automatically generate metadata established through the results of this study will be helpful for studying recommendation systems for media super-personalization.