• International Journal of Advanced Culture Technology
• /
• v.5 no.1
• /
• pp.83-89
• /
• 2017

In this paper, we propose an extended method of one-hot Long Short-Term Memory (LSTM) and evaluate the performance on spam filtering task. Most of traditional methods proposed for spam filtering task use word occurrences to represent spam or non-spam messages and all syntactic and semantic information are ignored. Major issue appears when both spam and non-spam messages share many common words and noise words. Therefore, it becomes challenging to the system to filter correct labels between spam and non-spam. Unlike previous studies on information filtering task, instead of using only word occurrence and word context as in probabilistic models, we apply a neural network-based approach to train the system filter for a better performance. In addition to one-hot representation, using term weight with attention mechanism allows classifier to focus on potential words which most likely appear in spam and non-spam collection. As a result, we obtained some improvement over the performances of the previous methods. We find out using region embedding and pooling features on the top of LSTM along with attention mechanism allows system to explore a better document representation for filtering task in general.

• KIPS Transactions on Computer and Communication Systems
• /
• v.7 no.4
• /
• pp.99-102
• /
• 2018

Caffe is a neural net learning software which is widely used in academic researches. The GPU memory capacity is one of the most important aspects of designing neural net architectures. For example, many object detection systems require to use less than 12GB to fit a single GPU. In this paper, we extended Caffe to allow to use more than 12GB GPU memory. To verify the effectiveness of the extended software, we executed some training experiments to determine the learning efficiency of the object detection neural net software using a PC with three GPUs.

• Proceedings of the Earthquake Engineering Society of Korea Conference
• /
• 2003.03a
• /
• pp.329-336
• /
• 2003

The box system that is composed only of reinforced concrete walls and slabs are adopted on many high-rise apartment buildings recently constructed in Korea. And the framed structure with shear wall core that can effectively resist horizontal forces is frequently adopted for the structural system for high-rise building structures. In these structures, a shear wall may have one or more openings for functional reasons. It is necessary to use subdivided finite elements for accurate analysis of the shear wall with openings. But it would take significant amount of computational time and memory if the entire building structure is subdivided into a finer mesh. An efficient analysis method that can be used regardless of the number, size and location of openings is proposed in this study. The analysis method uses super element, substructure, matrix condensation technique and fictitious beam technique. Three-dimensional analyses of the box system and the framed structure with shear wall core having various types of openings were peformed to verify the efficiency of the proposed method. It was confirmed that the proposed method have outstanding accuracy with drastically reduced time and computer memory from the analyses of example structures.

• Proceedings of the Earthquake Engineering Society of Korea Conference
• /
• 2003.03a
• /
• pp.50-58
• /
• 2003

We have developed a locally variable time-step scheme matching with discontinuous grids in the flute-difference method for the efficient simulation of seismic wave propagation. The first-order velocity-stress formulations are used to obtain the spatial derivatives using finite-difference operators on a staggered grid. A three-times coarser grid in the high-velocity region compared with the grid in the low-velocity region is used to avoid spatial oversampling. Temporal steps corresponding to the spatial sampling ratio between both regions are determined based on proper stability criteria. The wavefield in the margin of the region with smaller time-step are linearly interpolated in time using the values calculated in the region with larger one. The accuracy of the proposed scheme is tested through comparisons with analytic solutions and conventional finite-difference scheme with constant grid spacing and time step. The use of the locally variable time-step scheme with discontinuous grids results in remarkable saving of the computation time and memory requirement with dependency of the efficiency on the simulation model. This implies that ground motion for a realistic velocity structures including near-surface sediments can be modeled to high frequency (several Hz) without requiring severe computer memory

• Journal of the Society of Naval Architects of Korea
• /
• v.56 no.1
• /
• pp.47-57
• /
• 2019

Recent CFD solvers in engineering have to treat geometrically complex domains and moving body problems. In ship hydrodynamics, flow around the stern and ship motions in waves are examples of such cases mentioned before. The unstructured grid scheme is successfully applied for these problems, but it has weakness of inefficient memory usage and intensive computational time as compared to the structured grid method. Overset grid scheme is one of the alternatives for structured grid system taking advantage of fast and memory efficiency. Overset grid scheme is especially useful for moving body problem because there is no need to re-mesh around the body. In this paper, we adopted the Suggar++, the grid connectivity and interpolation utility for the overlapping grid, to WAVIS which is the in-house flow solver of KRISO. Then we introduced some applications using the overset grid method for flow analysis around the ships. The computed results show that WAVIS with Suggar++ is practically feasible and has an advantages for moving geometry cases.

• Smart Structures and Systems
• /
• v.32 no.3
• /
• pp.179-193
• /
• 2023

This study explores an alternative to the existing centralized process for data anomaly detection in modern Internet of Things (IoT)-based structural health monitoring (SHM) systems. An edge intelligence framework is proposed for the early detection and classification of various data anomalies facilitating quality enhancement of acquired data before transmitting to a central system. State-of-the-art deep neural network pruning techniques are investigated and compared aiming to significantly reduce the network size so that it can run efficiently on resource-constrained edge devices such as wireless smart sensors. Further, depthwise separable convolution (DSC) is invoked, the integration of which with advanced structural pruning methods exhibited superior compression capability. Last but not least, quantization-aware training (QAT) is adopted for faster processing and lower memory and power consumption. The proposed edge intelligence framework will eventually lead to reduced network overload and latency. This will enable intelligent self-adaptation strategies to be employed to timely deal with a faulty sensor, minimizing the wasteful use of power, memory, and other resources in wireless smart sensors, increasing efficiency, and reducing maintenance costs for modern smart SHM systems. This study presents a theoretical foundation for the proposed framework, the validation of which through actual field trials is a scope for future work.

• Journal of the Institute of Electronics Engineers of Korea TC
• /
• v.43 no.5 s.347
• /
• pp.137-146
• /
• 2006

IP address lookup has become a major bottleneck of packet forwarding and a critical issue for high-speed networking techniques in routers. In this paper, we propose an efficient high-speed IP address lookup scheme using adaptive multiple hashing and prefix grouping. According to our analysis results based on routing data distributions, we grouped prefix lengths and selected the number of hash functions in each group adaptively. As a result, we can reduce collisions caused by hashing. Accordingly, a forwarding table of our scheme has good memory efficiency, and thus is organized with the proper number of memory modules. Also, the proposed scheme has the fast building and searching mechanisms to develop the forwarding table only during a single memory access.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.16 no.4
• /
• pp.15-20
• /
• 2016

Write amplification is a critical factor that limits the stable performance of flash-based storage systems. To reduce write amplification, this paper presents a new technique that cooperatively manages data in flash storage and nonvolatile memory (NVM). Our scheme basically considers NVM as the cache of flash storage, but allows the original data in flash storage to be invalidated if there is a cached copy in NVM, which can temporarily serve as the original data. This scheme eliminates the copy-out operation for a substantial number of cached data, thereby enhancing garbage collection efficiency. Experimental results show that the proposed scheme reduces the copy-out overhead of garbage collection by 51.4% and decreases the standard deviation of response time by 35.4% on average.

• Journal of Korea Multimedia Society
• /
• v.7 no.8
• /
• pp.1086-1096
• /
• 2004

In case of a multimedia application which deals with streaming data, in terms of cache management, cache loses its efficiency due to weak temporal locality of the data. This means that when data have been brought into cache, much of the data are supposed to be replaced without being accessed again during its service. However, there is a good chance that such multimedia data has a commanding locality in it. In this paper, to take advantage of the memory reference regularity which typically innates even in the multimedia data showing up its weak temporal locality, a method is suggested. The suggested method with the feature of dynamic regular-stride reference prefetching can identify for 2-dimensional array format(block pattern). The suggested method is named as block-reference-prediction-technique (BRPT) since it identifies a block pattern and place an address to be prefetched by the regulation of the block format. BRPT proved to be reassuring to reduce memory reference time significantly for applications having abundant block patterns although new rule has complicated the prefetching system even further.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.8 no.8
• /
• pp.1828-1834
• /
• 2004

With the rapid development of wireless communications and mobile technologies, requirements of various services based on moving objects like location-based services and mobile applications services have been increased. In this paper, we propose an index structure which can improve the performance on trajectory-based query especially, one of the various query types for moving objects in mobile applications. It maintains link table(L-Table) to obtain good efficiency on retrieval and insertion performance of the existing TB(Trajectory Bundle)-tree proposed for trajectory-based query of moving objects. The L-Table contains page number in disk and memory pointers pointing the leaf node with the first and last line segment of moving objects in order to directly access preceding node. In addition, we design to reside a part of whole index in main memory by preserving a fixed size of buffer in case of being restricted by available main memory. Finally, experimental results with various data sets show that the proposed technique is superior to the existing index structures with respect to insertion and trajectory-based query.