• Asia Marketing Journal
• /
• v.15 no.3
• /
• pp.79-95
• /
• 2013

New product innovation is a process of embodying new knowledge in a product and technology licensing is getting popular as a means to innovations and introduction of new product to the market in today's competitive global market environment. Incumbents often rely on technology licensing to access new product opportunities created by other firms. Prior research has examined various aspects of technology licensing agreements such as specific contract terms of licensing agreements, e.g., distribution of control rights, exclusivity of licensing agreements, cross-licensing, and the scope of licensing agreements. This study aims to provide answers to an important, but under-researched question: why do some incumbents initiate more licensing agreement for exploratory learning while others do it for exploitative learning along the innovation process? We attempt to extend our knowledge of licensing agreements from an organizational learning perspective. Technology licensing as a specific form of interfirm linkages can be initiated with different learning objectives along the process of new product innovation. The exploratory stages of the innovation process such as discovery or research stages involve extensive searches to create new knowledge or capabilities, whereas the exploitative stages of the innovation process such as application or test stages near the commercialization are more focused on developing specific applications or improving their efficiency or reliability. Thus, different stages of the innovation process generate different types of learning and the resulting technological resources. We examine when incumbents as licensees initiate more licensing agreements for exploratory learning objectives and when more for exploitative learning objectives, focusing on two factors that may influence a firm's formation of exploratory and exploitative licensing agreements: 1) its past radical and incremental innovation experience and 2) its internal investments in R&D and marketing. We develop and test our hypotheses regarding the relationship between a firm's radical and incremental new product experience, R&D investment intensity and marketing investment intensity, and the likelihood of engaging in exploratory and exploitive licensing agreements. Using data collected from various secondary sources (Recap database, Compustat database, and FDA website), we analyzed technology licensing agreements initiated in the biotechnology and pharmaceutical industries from 1988 to 2011. The results of this study show that incumbents initiate exploratory rather than exploitative licensing agreements when they have more radical innovation experience and when they invest in R&D activities more intensively; in contrast, they initiate exploitative rather than exploratory licensing agreements when they have more incremental innovation experience and when they invest in marketing activities more intensively. The findings of this study contribute to the licensing and interfirm cooperation studies. First, this study lays a foundation to understand the organizational learning aspect of technology licensing agreements. Second, this study sheds lights on how a firm's internal investments in R&D and marketing are linked to its tendency to initiate licensing agreements along the innovation process. Finally, the findings of this study provide important insight to managers regarding which technologies to gain via licensing agreements. This study suggests that firms need to consider their internal investments in R&D and marketing as well as their past innovation experiences when they initiate licensing agreements along the process of new product innovation.

• Journal of Internet Computing and Services
• /
• v.24 no.5
• /
• pp.1-7
• /
• 2023

This paper proposes a small cell resource partitioning allocation method to solve interference to machine type communication devices (MTCD) and improve performance in 5G heterogeneous networks (HetNet) where macro base station (MBS) and many small cell base stations (SBS) are overlaid. In the 5G HetNet, since various types of MTCDs generate data traffic, the load on the MBS increases. Therefore, in order to reduce the MBS load, a cell range expansion (CRE) method is applied in which a bias value is added to the received signal strength from the SBS and MTCDs satisfying the condition is connected to the SBS. More MTCDs connecting to the SBS through the CRE will reduce the load on the MBS, but performance of MTCDs will degrade due to interference, so a method to solve this problem is needed. The proposed small cell resource partitioning allocation method allocates resources with less interference from the MBS to mitigate interference of MTCDs newly added in the SBS with CRE, and improve the overall MTCD performace using separating resources according to the performance of existing MTCDs in the SBS. Through simulation results, the proposed small cell resource partitioning allocation method shows performance improvement of 21% and 126% in MTCDs capacity connected to MBS and SBS respectively, compared to the existing resource allocation methods.

• The Journal of the Acoustical Society of Korea
• /
• v.42 no.6
• /
• pp.536-543
• /
• 2023

Speaker diarization, which labels for "who spoken when?" in speech with multiple speakers, has been studied on a deep neural network-based end-to-end method for labeling on speech overlap and optimization of speaker diarization models. Most deep neural network-based end-to-end speaker diarization systems perform multi-label classification problem that predicts the labels of all speakers spoken in each frame of speech. However, the performance of the multi-label-based model varies greatly depending on what the threshold is set to. In this paper, it is studied a speaker diarization system using single-label classification so that speaker diarization can be performed without thresholds. The proposed model estimate labels from the output of the model by converting speaker labels into a single label. To consider speaker label permutations in the training, the proposed model is used a combination of Permutation Invariant Training (PIT) loss and cross-entropy loss. In addition, how to add the residual connection structures to model is studied for effective learning of speaker diarization models with deep structures. The experiment used the Librispech database to generate and use simulated noise data for two speakers. When compared with the proposed method and baseline model using the Diarization Error Rate (DER) performance the proposed method can be labeling without threshold, and it has improved performance by about 20.7 %.

• Journal of Korea Society of Industrial Information Systems
• /
• v.28 no.6
• /
• pp.91-98
• /
• 2023

The amount of power generation of a solar plant has a high correlation with weather conditions, geographical conditions, and the installation conditions of solar panels. Previous studies have found the elements which impacts the amount of power generation. Some of them found the optimal conditions for solar panels to generate the maximum amount of power. Considering the realistic constraints when installing a solar power plant, it is very difficult to satisfy the conditions for the maximum power generation. Therefore, it is necessary to know how sensitive the solar power generation amount is to factors affecting the power generation amount, so that plant owners can predict the amount of solar power generation when examining the installation of a solar power plant. In this study, we propose a polynomial regression analysis method to analyze the relationship between solar power plant's power generation and related factors such as weather, location, and installation conditions. Analysis data were collected from 10 solar power plants installed and operated in Daegu and Gyeongbuk. As a result of the analysis, it was found that the amount of power generation was affected by panel type, amount of insolation and shade. In addition, the power generation was affected by interaction of the installation angle and direction of the panel.

• KIPS Transactions on Software and Data Engineering
• /
• v.12 no.10
• /
• pp.461-470
• /
• 2023

While speech animation generation employing deep learning has been actively researched for English, there has been no prior work for Korean. Given the fact, this paper for the very first time employs supervised deep learning to generate Korean speech animation. By doing so, we find out the significant effect of deep learning being able to make speech animation research come down to speech recognition research which is the predominating technique. Also, we study the way to make best use of the effect for Korean speech animation generation. The effect can contribute to efficiently and efficaciously revitalizing the recently inactive Korean speech animation research, by clarifying the top priority research target. This paper performs this process: (i) it chooses blendshape animation technique, (ii) implements the deep-learning model in the master-servant pipeline of the automatic speech recognition (ASR) module and the facial action coding (FAC) module, (iii) makes Korean speech facial motion capture dataset, (iv) prepares two comparison deep learning models (one model adopts the English ASR module, the other model adopts the Korean ASR module, however both models adopt the same basic structure for their FAC modules), and (v) train the FAC modules of both models dependently on their ASR modules. The user study demonstrates that the model which adopts the Korean ASR module and dependently trains its FAC module (getting 4.2/5.0 points) generates decisively much more natural Korean speech animations than the model which adopts the English ASR module and dependently trains its FAC module (getting 2.7/5.0 points). The result confirms the aforementioned effect showing that the quality of the Korean speech animation comes down to the accuracy of Korean ASR.

• KIPS Transactions on Software and Data Engineering
• /
• v.12 no.12
• /
• pp.519-524
• /
• 2023

In hashing-based image retrieval, the hash code of a manipulated image is different from the original image, making it difficult to search for the same image. This paper proposes and evaluates a self-supervised deephashing model that generates perceptual hash codes from feature information such as texture, shape, and color of images. The comparison models are autoencoder-based variational inference models, but the encoder is designed with a fully connected layer, convolutional neural network, and transformer modules. The proposed model is a variational inference model that includes a SimAM module of extracting geometric patterns and positional relationships within images. The SimAM module can learn latent vectors highlighting objects or local regions through an energy function using the activation values of neurons and surrounding neurons. The proposed method is a representation learning model that can generate low-dimensional latent vectors from high-dimensional input images, and the latent vectors are binarized into distinguishable hash code. From the experimental results on public datasets such as CIFAR-10, ImageNet, and NUS-WIDE, the proposed model is superior to the comparative model and analyzed to have equivalent performance to the supervised learning-based deephashing model. The proposed model can be used in application systems that require low-dimensional representation of images, such as image search or copyright image determination.

• The Transactions of the Korea Information Processing Society
• /
• v.13 no.5
• /
• pp.209-216
• /
• 2024

Real estate market prices are determined by various factors, including macroeconomic variables, as well as the influence of a variety of unstructured text data such as news articles and social media. News articles are a crucial factor in predicting real estate transaction prices as they reflect the economic sentiment of the public. This study utilizes sentiment analysis on news articles to generate a News Sentiment Index score, which is then seamlessly integrated into a real estate price prediction model. To calculate the sentiment index, the content of the articles is first summarized. Then, using AI, the summaries are categorized into positive, negative, and neutral sentiments, and a total score is calculated. This score is then applied to the real estate price prediction model. The models used for real estate price prediction include the Multi-head attention LSTM model and the Vector Auto Regression model. The LSTM prediction model, without applying the News Sentiment Index (NSI), showed Root Mean Square Error (RMSE) values of 0.60, 0.872, and 1.117 for the 1-month, 2-month, and 3-month forecasts, respectively. With the NSI applied, the RMSE values were reduced to 0.40, 0.724, and 1.03 for the same forecast periods. Similarly, the VAR prediction model without the NSI showed RMSE values of 1.6484, 0.6254, and 0.9220 for the 1-month, 2-month, and 3-month forecasts, respectively, while applying the NSI led to RMSE values of 1.1315, 0.3413, and 1.6227 for these periods. These results demonstrate the effectiveness of the proposed model in predicting apartment transaction price index and its ability to forecast real estate market price fluctuations that reflect socio-economic trends.

• Journal of Korea Water Resources Association
• /
• v.57 no.5
• /
• pp.311-320
• /
• 2024

Climate change has been intensifying drought frequency and severity. Such prolonged droughts reduce reservoir levels, thereby exacerbating drought impacts. While previous studies have focused on optimizing reservoir operations using historical data to mitigate these impacts, their scope is limited to analyzing past events, highlighting the need for predictive methods for future droughts. This research introduces a novel approach for predicting minimum inflow at the Seomjingang dam which has experienced significant droughts. This study utilized the Stochastic Analysis Modeling and Simulation (SAMS) 2007 to generate inflow sequences for the same period of observed inflow. Then we simulate reservoir operations to assess firm yield and predict minimum inflow through synthetic inflow analysis. Minimum inflow is defined as the inflow where firm yield is less than 95% of the synthetic inflow in many sequences during periods matching observed inflow. The results for each case indicated the firm yield for the minimum inflow is on average 9.44 m³/s, approximately 1.07 m³/s lower than the observed inflow's firm yield of 10.51 m³/s. The minimum inflow estimation can inform reservoir operation standards, facilitate multi-reservoir system reviews, and assess supplementary capabilities. Estimating minimum inflow emerges as an effective strategy for enhancing water supply reliability and mitigating shortages.

• The Journal of the Convergence on Culture Technology
• /
• v.10 no.2
• /
• pp.517-524
• /
• 2024

This study aims to train and implement a deep learning model for the fusion of website creation and artificial intelligence, in the era known as the AI revolution following the launch of the ChatGPT service. The deep learning model was trained using 3,000 collected web page images, processed based on a system of component and layout classification. This process was divided into three stages. First, prior research on AI models was reviewed to select the most appropriate algorithm for the model we intended to implement. Second, suitable web page and paragraph images were collected, categorized, and processed. Third, the deep learning model was trained, and a serving interface was integrated to verify the actual outcomes of the model. This implemented model will be used to detect multiple paragraphs on a web page, analyzing the number of lines, elements, and features in each paragraph, and deriving meaningful data based on the classification system. This process is expected to evolve, enabling more precise analysis of web pages. Furthermore, it is anticipated that the development of precise analysis techniques will lay the groundwork for research into AI's capability to automatically generate perfect web pages.

• Information Systems Review
• /
• v.26 no.2
• /
• pp.23-43
• /
• 2024

Nowadays, online reviews have become a common word of mouth that many users produce and consume. Posting online reviews is a kind of job that consumers do themselves. Since posting online reviews is not mandatory, it entirely relies on the consumer's voluntary willingness. In this respect, this study aims to describe the motivation for posting online reviews and their behavior processes, such as why online reviewers generate reviews and what types of reviews they create. In this study, we have conducted an in-depth study with 18 participants who have experience in posting reviews. By analyzing interview manuscripts from the grounded theory method approach, we have ultimately presented motivating factors for review posting (mutual reciprocity, material rewards), determinants of review browsing (trust toward review contents, preference for review format), and shadow work (a job that must be done, voluntary data production, consumer's share). We have also proposed the dynamics between core dimensions for theorizing a cycle process of review production and consumption. Our findings could bridge the gap in the existing online review research and offer practical implications for platform companies that need review management.