• JSTS:Journal of Semiconductor Technology and Science
• /
• v.4 no.4
• /
• pp.286-292
• /
• 2004

In this paper we present a static performance estimation technique of on-chip bus architectures. The proposed technique requires the static scheduling of function blocks of a task to analyze bus conflicts caused by simultaneous accesses from processing elements to which function blocks are mapped. To apply it to multitask applications, the concurrent execution of the function blocks of different tasks also should be considered. Since tasks are scheduled independently, considering all cases of concurrency in each processing element is impractical. Therefore we make an average estimate on the effects of other tasks with respect to bus request rate and bus access time. The proposed technique was incorporated with our exploration framework for on-chip bus architectures, Its viability and efficiency are validated by a preliminary example.

• Journal of Computing Science and Engineering
• /
• v.5 no.3
• /
• pp.257-268
• /
• 2011

Machine learning and data mining have found many applications in biological domains, where we look to build predictive models based on labeled training data. However, in practice, high quality labeled data is scarce, and to label new data incurs high costs. Transfer and multitask learning offer an attractive alternative, by allowing useful knowledge to be extracted and transferred from data in auxiliary domains helps counter the lack of data problem in the target domain. In this article, we survey recent advances in transfer and multitask learning for bioinformatics applications. In particular, we survey several key bioinformatics application areas, including sequence classification, gene expression data analysis, biological network reconstruction and biomedical applications.

• Journal of KIISE:Computer Systems and Theory
• /
• v.33 no.9
• /
• pp.592-607
• /
• 2006

This paper proposes a technique to select processors and hardware IPs and to map the tasks into the selected processing elements, aming to achieve high performance with minimal system cost when multitask applications with real-time constraints are run on a multicore SoC. Such technique is called to 'Hardware-Software Cosynthesis Technique'. A cosynthesis technique was already presented in our early work [1] where we divide the complex cosynthesis problem into three subproblems and conquer each subproblem separately: selection of appropriate processing components, mapping and scheduling of function blocks to the selected processing component, and schedulability analysis. Despite good features, our previous technique has a serious limitation that a task monopolizes the entire system resource to get the minimum schedule length. But in general we may obtain higher performance in multitask multicore system if independent multiple tasks are running concurrently on different processor cores. In this paper, we present two mapping techniques, task mapping avoidance technique(TMA) and task mapping pinning technique(TMP), which are applicable for general cases with diverse operating policies in a multicore environment. We could obtain significant performance improvement for a multimedia real-time application, multi-channel Digital Video Recorder system and for randomly generated multitask graphs obtained from the related works.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.15 no.2
• /
• pp.421-441
• /
• 2021

Fruit detection in orchards is one of the most crucial tasks for designing the visual system of an automated harvesting robot. It is the first and foremost tool employed for tasks such as sorting, grading, harvesting, disease control, and yield estimation, etc. Efficient visual systems are crucial for designing an automated robot. However, conventional fruit detection methods always a trade-off with accuracy, real-time response, and extensibility. Therefore, an improved method is proposed based on coarse-to-fine multitask cascaded convolutional networks (MTCNN) with three aspects to enable the practical application. First, the architecture of Fruit-MTCNN was improved to increase its power to discriminate between objects and their backgrounds. Then, with a few manual labels and operations, synthetic images and labels were generated to increase the diversity and the number of image samples. Further, through the online hard example mining (OHEM) strategy during training, the detector retrained hard examples. Finally, the improved detector was tested for its performance that proved superior in predicted accuracy and retaining good performances on portability with the low time cost. Based on performance, it was concluded that the detector could be applied practically in the actual orchard environment.

• Journal of the Korean Institute of Telematics and Electronics B
• /
• v.30B no.10
• /
• pp.59-68
• /
• 1993

This paper describes implemation of 'General-purpose ETRI MAP interface module' (GEM) for Mini-MAP network. GEM operates as a Mini-MAP node in our FA system. To communicate between GEM and programmable devices(PD) such as PLC and CNC, serial communication is used. Application programs of a MiNi-MAP host system control and monitor programmable devices via GEM. GEM is implemented and tested on the basis of the MAP 3.0. TBC in the Nini-MAP board performs the function of the MAC sublayer. The LLC sublayer is implemented according to the specification of Class 3 that includes Type 1 and 3. And the MMS services are designed within the scope of implementation class MAP3. All the softwares are implemented under the real-time multitask OS for real-time application of the Mini-MAP and they are loaded into PROMs at the network board of GEM. We tested the LLC functions to make use of a protocol analyzer for the token-passing protocol. Also the MMS conformance test was carried out by exchanging primitives between GEM and a MMS product that had already passed the conformance test. Therefore GEM is proposed as a network tool of Computer Integrated Manufacturing (CIM) to integrate PDs which don't support MAP functions.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.12 no.5
• /
• pp.451-465
• /
• 1987

With the introduction of ISDN network, the D channel protocol has bee defined as a signalling method for ISDN user-network interface. Therefore the NT2(Network Termination 2) which carry out concentration and switching function, must process the D channel related information. This paper describes how the D channel protocoal is applied and implented in a small ISDN subscriber concentrating system that has NT2 functions. The application protocol proposed is addressed taking into consideration the compatibility with ISDN standard facilities, TE(Terminal Equipment) or ET(Exchange Terminator), This protocol has been implementes using a general multitask operating system and it has the features of the minimized information processing and the simpified algorithm which are suitable for a small system. Its application programs are divided into various tasks to facilitate the addition and the modification of function. In this paper, we briefly outline the protocol defined in CCITT and show the application protocol that has fitted in a small concentrating system with NT2 functions. Also we present the experimental results and implementation method of this protocol.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.32 no.8B
• /
• pp.509-520
• /
• 2007

In this paper, we propose a real-time sensor node platform that guarantees the real-time scheduling of periodic and aperiodic tasks through a multitask-based software decomposition technique. Since existing sensor networking operation systems available in literature are not capable of supporting the real-time scheduling of periodic and aperiodic tasks, the preemption of aperiodic task with high priority can block periodic tasks, and so periodic tasks are likely to miss their deadlines. This paper presents a comprehensive evaluation of how to structure periodic or aperiodic task decomposition in real-time sensor-networking platforms as regard to guaranteeing the deadlines of all the periodic tasks and aiming to providing aperiodic tasks with average good response time. A case study based on real system experiments is conducted to illustrate the application and efficiency of the multitask-based dynamic component execution environment in the sensor node equipped with a low-power 8-bit microcontroller, an IEEE802.15.4 compliant 2.4GHz RF transceiver, and several sensors. It shows that our periodic and aperiodic task decomposition technique yields efficient performance in terms of three significant, objective goals: deadline miss ratio of periodic tasks, average response time of aperiodic tasks, and processor utilization of periodic and aperiodic tasks.

• Korean Journal of Radiology
• /
• v.24 no.11
• /
• pp.1061-1080
• /
• 2023

Artificial intelligence (AI) in radiology is a rapidly developing field with several prospective clinical studies demonstrating its benefits in clinical practice. In 2022, the Korean Society of Radiology held a forum to discuss the challenges and drawbacks in AI development and implementation. Various barriers hinder the successful application and widespread adoption of AI in radiology, such as limited annotated data, data privacy and security, data heterogeneity, imbalanced data, model interpretability, overfitting, and integration with clinical workflows. In this review, some of the various possible solutions to these challenges are presented and discussed; these include training with longitudinal and multimodal datasets, dense training with multitask learning and multimodal learning, self-supervised contrastive learning, various image modifications and syntheses using generative models, explainable AI, causal learning, federated learning with large data models, and digital twins.

• The KIPS Transactions:PartC
• /
• v.14C no.4
• /
• pp.371-382
• /
• 2007

In this paper, we propose a real-time sensor node platform that efficiently manages periodic and aperiodic tasks. Since existing sensor node platforms available in literature focus on minimizing the usage of memory and power consumptions, they are not capable of supporting the management of tasks that need their real-time execution and fast average response time. We first analyze how to structure periodic or aperiodic task decomposition in the TinyOS-based sensor node platform as regard to guaranteeing the deadlines of ail the periodic tasks and aiming to providing aperiodic tasks with average good response time. Then we present the application and efficiency of the proposed real-time sensor node platform in the sensor node equipped with a low-power 8-bit microcontroller, an IEEE802.15.4 compliant 2.4GHz RF transceiver, and several sensors. Extensive experiments show that our sensor node platform yields efficient performance in terms of three significant, objective goals: deadline miss ratio of periodic tasks, average response time of aperiodic tasks, and processor utilization of periodic and aperiodic tasks.