• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.24 no.3
• /
• pp.87-94
• /
• 2024

We investigated the feasibility of utilizing Large Language Models (LLMs) to perform text-based games without training on game data in advance. We adopted ChatGPT-3.5 and its state-of-the-art, ChatGPT-4, as the systems that implemented LLM. In addition, we added the persistent memory feature proposed in this paper to ChatGPT-4 to create three game player agents. We used Zork, one of the most famous text-based games, to see if the agents could navigate through complex locations, gather information, and solve puzzles. The results showed that the agent with persistent memory had the widest range of exploration and the best score among the three agents. However, all three agents were limited in solving puzzles, indicating that LLM is vulnerable to problems that require multi-level reasoning. Nevertheless, the proposed agent was still able to visit 37.3% of the total locations and collect all the items in the locations it visited, demonstrating the potential of LLM.

• Progress in Medical Physics
• /
• v.14 no.4
• /
• pp.225-233
• /
• 2003

The goal of a radiation treatment plan is to deliver a homogeneous dose to a target with minimal irradiation of the adjacent normal tissues. Dose uniformity is especially important for stereotactic radiosurgery using a linear accelerator. The dose uniformity and high dose delivery of a single spherical dose distribution exceed 70％. This also results with a similar stereotactic radiosurgical plan using a Gamma Knife. The dose distribution produced in a stereotactic radiosurgical plan using a Gamma Knife and Linear accelerator is spherical, and the application of the sphere packing arrangement in a real radiosurgical plan requires much time and skill. In this study, we found a characteristic of dose distribution with transformation of beam parameters that must be considered in a radiosurgical plan for effective radiosurgery. First, we assumed a cylinder type tumor model and a cube type tumor model. Secondly, the results of the tumor models were compared and analyzed with dose profiles and DVH_(Dose Volume Histogram) representative dose distribution. We found the optimal composition of beam parameters_(i.e. collimator size, number of isocenter, gap of isocenters etc.), which allowed the tumor models to be involved in the isodose curve at a high level. In conclusion, the characteristics found in this study are helpful for improving the effectiveness and speed of a radiosurgical plan for stereotactic radiosurgery.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.38 no.2
• /
• pp.261-268
• /
• 2018

The study used the simplified flooding analysis model, SIMOD, to distribute the total flood discharge by time, so research on flooding in urban areas can be conducted. The conventional flooding analysis models have limitations in constructing input data and take a long time for analysis. However, SIMOD is useful because it supports rapid decision-making process using quick modeling based on simple hydrological data, such as topography and inflow flood of the study area, to analyze submerged routes formed by flooding. Therefore, the study used the SIMOD model to analyze flooding in urban areas before conducting a comparative study with the outputs from FLO-2D, which is one of the conventional flooding analysis models, to identify the model's applicability. Seongseoje was selected as the study area, as it is located downstream the Geumho river where streams flow in the adjacent areas, and dikes are high enough to apply the "Overflow and Break" scenario for urban areas. With regard to topography, the study applied DEM data for the conventional flooding analysis and DSM data to represent urban building communities, distribution of roads, etc. Input flood discharge was calculated by applying the rectangular weir equation under the bank and break scenario through a 200-year return period of a design flood level. Comparative analysis was conducted in a flooded area with a simulation time of 1-24 hours. The time for the 24-hour simulation in SIMOD was less than 7 minutes. Compared with FLO-2D, the difference in flooded areas was less than 20%. Furthermore, the study identified the need for topography data using DSM for urban areas, as the analysis result that applies DSM showed the influence of roads and buildings.

• Journal of the HCI Society of Korea
• /
• v.9 no.2
• /
• pp.61-68
• /
• 2014

Tabletop games have been applied to improve the ability of social collaboration based on the characteristics that more than two people simultaneously interact on the tabletop. Especially, the tabletop games can be used as an educational tool for children with autism when it is implemented by considering educational and psychological models for children's behavioral characteristics. However, the previous collaborative games were designed for only disabled children so that it is hard to reflect cognitive and humanistic effects in inclusive education, where disabled children and non-disabled children interact in a same spatiotemporal environment. In this paper, therefore, we design a collaborative game on a multi-touch tabletop to enable spontaneous communication between disabled children and non-disabled children. Through user study, we evaluate the improvement in terms of the positive interaction and the degree of attention by comparing with a conventional collaborative game(e.g., a board game). We found that negative interaction including disabled children's abnormal behavior decreased and positive interaction such as body gestures and verbal communications increased. In addition, the tabletop game supported high immersiveness to all children by deriving equal level of attention time including individual and joint attention. We anticipate that the proposed game design can be utilized to develop collaborative contents for people with differences on sociality and cognitive ability.

• Journal of the Korean Institute of Telematics and Electronics S
• /
• v.35S no.8
• /
• pp.96-105
• /
• 1998

The aim of this paper is to propose the method of WDHMM(Weighted DHMM), using the MLP-VQ for the improvement of speaker-independent connect-digit recognition system. MLP neural-network output distribution shows a probability distribution that presents the degree of similarity between each pattern by the non-linear mapping among the input patterns and learning patterns. MLP-VQ is proposed in this paper. It generates codewords by using the output node index which can reach the highest level within MLP neural-network output distribution. Different from the old VQ, the true characteristics of this new MLP-VQ lie in that the degree of similarity between present input patterns and each learned class pattern could be reflected for the recognition model. WDHMM is also proposed. It can use the MLP neural-network output distribution as the way of weighing the symbol generation probability of DHMMs. This newly-suggested method could shorten the time of HMM parameter estimation and recognition. The reason is that it is not necessary to regard symbol generation probability as multi-dimensional normal distribution, as opposed to the old SCHMM. This could also improve the recognition ability by 14.7% higher than DHMM, owing to the increase of small caculation amount. Because it can reflect phone class relations to the recognition model. The result of my research shows that speaker-independent connected-digit recognition, using MLP-VQ and WDHMM, is 84.22%.

• Journal of the Korean Society of Environmental Restoration Technology
• /
• v.7 no.2
• /
• pp.75-86
• /
• 2004

This study aims to figure out ecological characteristic of natural forests focusing on vegetation as a way of ecological renovation for the restoration of naturalness for golf courses that were constructed in the Country and to present vegetation models and appropriate tree species for the purpose. The study site is P golf club, which is located in Gapyong-gun, Kyeonggi Province. The site is within a forest where the grade from the natural ecology map is the first one and the level from the green index accounts for the eight, thus showing a typical environment for a golf course in terms of location. The location of the site explains a reason for restoration. The major substance of the study is to conduct ecological evaluation of vegetation structures around and inside the golf course and to present a vegetation model. In order to evaluate the ecological characteristics of the vegetation structures, the analysis of the study covered succession stages, multi-layer vegetation structures and species diversity. Plant communities that have high species diversity were selected for the vegetation model and proper density and species were proposed considering the number of species and individuals and distances between trees. The vegetation restoration model targets succession into an oak forest. Within a unit of 100 $m^2$, the recommended model include a tall-tree layer with 11 trees such as Quercus serrata and Quercus mongolica, a sub-tall-tree layer with 12 trees including Quercus mongolica, Quercus serrata, Prunus sargentii, Fraxinus rhynchophylla and Acer pseudo sieboldianum, a shrub layer with 32 trees from 16 species, and a grass layer with a cover rate of 45 %. The proposed vegetation restoration model needs to apply to : 1) damaged natural forests by the construction of golf courses; 2) boundaries between golf courses and surrounding forests; 3) buffer zones; 4) open spaces in between courses; and 5) areas between greens and tees where open spaces are available in a mass. In conclusion, one of the most important factors in presenting a vegetation model for the restoration of naturalness in the golf course and other damaged forests is to provide multi-layer vegetation structures that are composed of native species. As for the specific application for the site, it is recommended to manage the vegetation in such a way that the environment of the site can have a similar environment to the surrounding forest which is expected to succeed into an oaks-dominant one.

• Journal of the Korea Society of Computer and Information
• /
• v.25 no.6
• /
• pp.109-117
• /
• 2020

An edutainment system aims to help learners to recognize problems effectively, grasp and classify important information needed to solve the problems and convey the contents of what they have learned. Edutainment contents can be usefully applied to education and training in the both scientific and industrial areas. Our present work proposes an edutainment system that can be applied to a drug discovery process including virtual screening by using intuitive multi-modal interfaces. In this system, a stereoscopic monitor is used to make three-dimensional (3D) macro-molecular images, with supporting multi-modal interfaces to manipulate 3D models of molecular structures effectively. In this paper, our system can easily solve a docking simulation function, which is one of important virtual drug screening methods, by applying gaming factors. The level-up concept is implemented to realize a bio-game approach, in which the gaming factor depends on number of objects and users. The quality of the proposed system is evaluated with performance comparison in terms of a finishing time of a drug docking process to screen new inhibitors against target proteins of human immunodeficiency virus (HIV) in an e-drug discovery process.

• The Journal of the Convergence on Culture Technology
• /
• v.10 no.3
• /
• pp.421-426
• /
• 2024

The Analytic Hierarchy Process (AHP) converts people's judgment criteria into objective numerical values using pairwise comparisons. However, the need for an excessive number of pairwise comparisons poses a problem. To mitigate this issue, most existing studies have utilized the process separation approach. The method of process separation devised in this study is a "separation and integration approach," where 1) the standard AHP process is used for evaluating judgment criteria, and 2) the Multi-Attributive Utility Technique (MAUT) is applied for comparing alternatives. This AHP-MAUT Hybrid model was applied to a real analysis case, specifically analyzing the transportation choices of commuters between Bundang and Gangnam Station in Gyeonggi Province. The results showed that the computational process was reduced by 42.03% when applying the Hybrid model compared to using the AHP model alone. Furthermore, the choice results of residents using the Hybrid model were compared with those using the standard AHP. The consistency between the two models' choices was 82.1%, indicating a significant level of consistency. In conclusion, this study contributes by presenting a simpler, more convenient, yet equally effective Hybrid model as a new decision-support system alternative to AHP.

• Korean Journal of Environment and Ecology
• /
• v.34 no.4
• /
• pp.282-293
• /
• 2020

This study investigated a long-term variation trend of water quality, fish compositions, and ecological health conditions in the Gwanpyeong stream located in the nearby Daejeon metropolitan city to understand the impact of urban development projects on the aquatic ecosystem. The sampling was made in four surveys (2009, 2010, 2016, 2019) before and after urbanization. The urban development was conducted in 2008, resulting in the stream's ecological disturbance, and the stream restoration was conducted in 2012. Thus, stream monitoring was conducted to analyze the ecological trends before and after the restoration. The multi-metric models for Fish Assessment Index(FAI) and necropsy-based Health Assessment Index(HAI) were applied in the fish community and organ-level, respectively, to assess the ecological health of the stream. Minimum turbidity and chlorophyll-a(Chl-a) occurred in the mid-stream(St. 2), and this was probably due to rapid current velocity in the riffle zone. We collected 18 fish species, and the dominant species was Zacco platypus (40.6%). In 2016 immediately after the stream restoration, the relative proportions of sensitive species and insectivore species were the highest along with highest values in the species diversity and species richness index, resulting in the best condition in the ecological health, based on FAI model values. However, the ecological health, based on the FAI, became worse in the latest survey conducted in 2019. The analysis of the HAI model based on the organ-level approach showed skin erosion in the fish of upper stream, kidney defects in downstream, and the liver and gill defects observed in all sites, indicating that the anatomical health was also affected.

• Journal of Intelligence and Information Systems
• /
• v.19 no.2
• /
• pp.39-54
• /
• 2013

Recent explosive increase of electronic commerce provides many advantageous purchase opportunities to customers. In this situation, customers who do not have enough knowledge about their purchases, may accept product recommendations. Product recommender systems automatically reflect user's preference and provide recommendation list to the users. Thus, product recommender system in online shopping store has been known as one of the most popular tools for one-to-one marketing. However, recommender systems which do not properly reflect user's preference cause user's disappointment and waste of time. In this study, we propose a novel recommender system which uses data mining and multi-model ensemble techniques to enhance the recommendation performance through reflecting the precise user's preference. The research data is collected from the real-world online shopping store, which deals products from famous art galleries and museums in Korea. The data initially contain 5759 transaction data, but finally remain 3167 transaction data after deletion of null data. In this study, we transform the categorical variables into dummy variables and exclude outlier data. The proposed model consists of two steps. The first step predicts customers who have high likelihood to purchase products in the online shopping store. In this step, we first use logistic regression, decision trees, and artificial neural networks to predict customers who have high likelihood to purchase products in each product group. We perform above data mining techniques using SAS E-Miner software. In this study, we partition datasets into two sets as modeling and validation sets for the logistic regression and decision trees. We also partition datasets into three sets as training, test, and validation sets for the artificial neural network model. The validation dataset is equal for the all experiments. Then we composite the results of each predictor using the multi-model ensemble techniques such as bagging and bumping. Bagging is the abbreviation of "Bootstrap Aggregation" and it composite outputs from several machine learning techniques for raising the performance and stability of prediction or classification. This technique is special form of the averaging method. Bumping is the abbreviation of "Bootstrap Umbrella of Model Parameter," and it only considers the model which has the lowest error value. The results show that bumping outperforms bagging and the other predictors except for "Poster" product group. For the "Poster" product group, artificial neural network model performs better than the other models. In the second step, we use the market basket analysis to extract association rules for co-purchased products. We can extract thirty one association rules according to values of Lift, Support, and Confidence measure. We set the minimum transaction frequency to support associations as 5%, maximum number of items in an association as 4, and minimum confidence for rule generation as 10%. This study also excludes the extracted association rules below 1 of lift value. We finally get fifteen association rules by excluding duplicate rules. Among the fifteen association rules, eleven rules contain association between products in "Office Supplies" product group, one rules include the association between "Office Supplies" and "Fashion" product groups, and other three rules contain association between "Office Supplies" and "Home Decoration" product groups. Finally, the proposed product recommender systems provides list of recommendations to the proper customers. We test the usability of the proposed system by using prototype and real-world transaction and profile data. For this end, we construct the prototype system by using the ASP, Java Script and Microsoft Access. In addition, we survey about user satisfaction for the recommended product list from the proposed system and the randomly selected product lists. The participants for the survey are 173 persons who use MSN Messenger, Daum Caf$\acute{e}$, and P2P services. We evaluate the user satisfaction using five-scale Likert measure. This study also performs "Paired Sample T-test" for the results of the survey. The results show that the proposed model outperforms the random selection model with 1% statistical significance level. It means that the users satisfied the recommended product list significantly. The results also show that the proposed system may be useful in real-world online shopping store.