• Electronics and Telecommunications Trends
• /
• v.36 no.2
• /
• pp.32-42
• /
• 2021

The rapid growth of deep-learning applications has invoked the R&D of artificial intelligence (AI) processors. A dedicated software framework such as a compiler and runtime APIs is required to achieve maximum processor performance. There are various compilers and frameworks for AI training and inference. In this study, we present the features and characteristics of AI compilers, training frameworks, and inference engines. In addition, we focus on the internals of compiler frameworks, which are based on either basic linear algebra subprograms or intermediate representation. For an in-depth insight, we present the compiler infrastructure, internal components, and operation flow of ETRI's "AI-Ware." The software framework's significant role is evidenced from the optimized neural processing unit code produced by the compiler after various optimization passes, such as scheduling, architecture-considering optimization, schedule selection, and power optimization. We conclude the study with thoughts about the future of state-of-the-art AI compilers.

• Advanced Industrial SCIence
• /
• v.1 no.1
• /
• pp.1-6
• /
• 2022

As different news is reported every day in our daily life, the amount of data generated around us is enormous. Some of this data is personal, but there are some things that are common to everyone, such as weather and traffic information. This paper is a study on information transmission using public data to use information effectively and quickly, and investigated open API based on public data and API technology. Based on this, we explain a method of transmitting information using an open API that can be easily used in daily life, and propose a method to transmit information using various open APIs by applying this. If public data is used using the suggested method, it is possible to transmit information more easily and accurately, so it is possible to apply it to various fields.

• Journal of Internet Computing and Services
• /
• v.24 no.1
• /
• pp.9-15
• /
• 2023

Single-Page Application(SPA) requires data transformation for communication with RESTful API. The Backend for Frontend(BFF) pattern handles this transformation in the server, but there is some problem that increases the number of communication and makes development and distribution difficult. In this study, we propose an architecture that maps the ViewModel of SPA and the model of RESTful API directly in SPA. The proposed architecture automatically generates a mapping model between the RESTful API model and the ViewModel using the OpenAPI specification, which is the document model of the RESTful API. The data transfer component of SPA automatically converts RESTful API data and ViewModel using the created model. As a result of comparison with the existing BFF method through case study, the proposed architecture showed higher development productivity than BFF, and as a result of load tests, it recorded about 6% lower server CPU occupancy compared to BFF.

• Smart Media Journal
• /
• v.11 no.10
• /
• pp.22-29
• /
• 2022

Voice interactions are particularly effective in applications targeting the digital underprivileged who are not proficient in the use of smart devices. However, applications based on open APIs are using voice signals only for short, fragmentary input and output due to the limitations of existing touchscreen-oriented UI and API provided. In this paper, we design a conversational voice interaction model for interactions between users and intelligent mobile/IoT applications and propose a keyword detection algorithm based on the edit distance. The proposed model and scheme were implemented in an Android environment, and the edit distance-based keyword detection algorithm showed a higher recognition rate than the existing algorithm for keywords that were incorrectly recognized through speech recognition.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.16 no.5
• /
• pp.1489-1515
• /
• 2022

The number of interconnected real-world devices such as sensors, actuators, and physical devices has increased with the advancement of technology. Due to this advancement, users face difficulties searching for the location of these devices, and the central issue is the findability of Things. In the WoT environment, keyword-based and geospatial searching approaches are used to locate these devices anywhere and on the web interface. A few static methods of indexing and ranking are discussed in the literature, but they are not suitable for finding devices dynamically. The authors have proposed a mechanism for dynamic and efficient searching of the devices in this paper. Indexing and ranking approaches can improve dynamic searching in different ways. The present paper has focused on indexing for improving dynamic searching and has indexed the Things Description in Solr. This paper presents the Things Description according to the model of W3C JSON-LD along with the open-access APIs. Search efficiency can be analyzed with query response timings, and the accuracy of response timings is critical for search results. Therefore, in this paper, the authors have evaluated their approach by analyzing the search query response timings and the accuracy of their search results. This study utilized different indexing approaches such as key-words-based, spatial, and hybrid. Results indicate that response time and accuracy are better with the hybrid approach than with keyword-based and spatial indexing approaches.

• Korean Journal of Artificial Intelligence
• /
• v.11 no.4
• /
• pp.29-34
• /
• 2023

The flow from developing a machine learning model to deploying it in a production environment suffers challenges. Efficient and reliable deployment is critical for realizing the true value of machine learning models. Bridging this gap between development and publication has become a pivotal concern in the machine learning community. FastAPI, a modern and fast web framework for building APIs with Python, has gained substantial popularity for its speed, ease of use, and asynchronous capabilities. This paper focused on leveraging FastAPI for deploying machine learning models, addressing the potentials associated with integration, scalability, and performance in a production setting. In this work, we explored the seamless integration of machine learning models into FastAPI applications, enabling real-time predictions and showing a possibility of scaling up for a more diverse range of use cases. We discussed the intricacies of integrating popular machine learning frameworks with FastAPI, ensuring smooth interactions between data processing, model inference, and API responses. This study focused on elucidating the integration of machine learning models into production environments using FastAPI, exploring its capabilities, features, and best practices. We delved into the potential of FastAPI in providing a robust and efficient solution for deploying machine learning systems, handling real-time predictions, managing input/output data, and ensuring optimal performance and reliability.

• The KIPS Transactions:PartC
• /
• v.11C no.4
• /
• pp.419-428
• /
• 2004

Due to the deployment problem of the IP Multicast service, the Application Layer Multicast (or Overlay Multicast) has appeared as an alter-native of the If Multicast. However, even though plenty of the Application Layer Multicast (ALM) Protocols were designed and their applications were developed according to the diverse requirements of each multicast service, researches on the ALM Protocols are focused on only a protocol design or an efficient multicast group management algorithm. And there is little effort to provide a unified guideline for development of the ALM Protocols and provide an environment for running multiple protocols simultaneously in a system. In this paper, we propose socket APIs to be a reference in developing new ALM Protocols which enables a system to support multiple protocols in a system with other ALM Protocols and which gives an environment to support efficient protocol management.

• Journal of KIISE:Computing Practices and Letters
• /
• v.14 no.1
• /
• pp.52-62
• /
• 2008

As various types of embedded devices are widely used, a technology that supports reuse of applications on multiple platforms is needed in order for time-to-market development of the applications. The interface middleware is one of such technology and it hides platform dependency from application programmers. Existing interface middleware such as the MT project, Xenomai and Legacy2linux have limitation in that the APIs provided by each of these middleware are fixed to a specific operating system, and the middleware does not provide dynamic expansion of its API set. In this paper, we propose a middleware which hides operating system dependencies and enables porting of applications on various operating systems. In addition, the middleware has scalable structure so that it is suitable for resource-limited embedded systems. The overhead of the middleware, i.e., the time delay occurred by the middleware is between $0.3{\mu}sec\;and\;5{\mu}sec$ in most cases. We believe that the amount of overhead is reasonable and does not hurt the performance of applications.

• Journal of Digital Convergence
• /
• v.11 no.12
• /
• pp.479-485
• /
• 2013

Virtualization is a promising approach to consolidating multiple online services onto a smaller number of computing resources. A virtualized server environment allows computing resources to be shared among multiple performance isolated platforms called virtual machines. Through server virtualization software, applications servers are encapsulated into VMs, and deployed with APIs on top generalized pools of CPU and memory resources. Networking and security have been moved to a software abstraction layer that transformed computing, network virtualization. And it paves the way for enterprise to rapidly deploy networking and security for any application by creating the virtual network. Stochastic reward net (SRN) is an extension of stochastic Petri nets which provides compact modeling facilities for system analysis. In this paper, we develop SRN model of network virtualization based on virtual switch. Measures of interest such as switching delay and throughput are considered. These measures are expressed in terms of the expected values of reward rate functions for SRNs. Numerical results are obtained according to the virtual switch capacity and number of active VMs.

• Journal of the Korean Society of Environmental Restoration Technology
• /
• v.20 no.3
• /
• pp.55-64
• /
• 2017

The problem of the population number of honeybees that is decreasing not only domestically but also globally, has a great influence on human beings and the entire ecosystem. The habitat of honeybees is recognized to be superior in urban environment rather than rural environment, and predicting for habitat assessment and conservation is necessary. Based on this, we targeted Cheonan City and neighboring administrative areas where the distribution of agricultural areas, urban areas, and forest areas is displayed equally. In order to predict the habitat preferred by honeybees, we apply the Maxent model what based on the presence information of the species. We also selected 10 environmental variables expected to influence honeybees habitat environment through literature survey. As a result of constructing the species distribution model using the Maxent model, 71.7% of the training data were shown on the AUC(Area Under Cover) basis, and it was be confirmed with an area of 20.73% in the whole target area, based on the 50% probability of presence of honeybees. It was confirmed that the contribution of the variable has influence on land covering, distance from the forest, altitude, aspect. Based on this, the possibility of honeybee's habitat characteristics were confirmed to be higher in wetland environment, in agricultural land, close to forest and lower elevation, southeast and west. The prediction of these habitat environments has significance as a lead research that presents the habitat of honeybees with high conservation value of ecosystems in terms of urban space, and it will be useful for future urban park planning and conservation area selection.