• Iberoamérica
• /
• v.23 no.1
• /
• pp.79-111
• /
• 2021

The purpose of this article is to present the potential of balangandãs as a documentary source for intersectional studies of gender and slavery, from the analysis of formal and symbolic aspects of the museum objects in the Museu Nacional Nacional (MHN) and Museu Carlos Costa Pinto (MCCP). Balangandãs are a type of creole jewelry, made in gold or silver and used in Brazil since the 18th century by black women - women who worked, especially in the commercialization of foodstuffs in large urban centers. They are described in printed sources and engravings, and preserved in some museum institutions. In the study, the meanings attributed to the object over the centuries are observed: jewelry, amulet, peculium, document, travel memory and heritage. As a result, the article seeks to highlight the objects as a historical and documentary source, verifying similarities between the pieces musealized at the MHN and at the MCCP, also emphasizing the documental power of the pieces produced today.

• Electronics and Telecommunications Trends
• /
• v.39 no.4
• /
• pp.82-92
• /
• 2024

This paper introduces the technological development trends in on-device SLMs (Small Language Models). Large Language Models (LLMs) based on the transformer model have gained global attention with the emergence of ChatGPT, providing detailed and sophisticated responses across various knowledge domains, thereby increasing their impact across society. While major global tech companies are continuously announcing new LLMs or enhancing their capabilities, the development of SLMs, which are lightweight versions of LLMs, is intensely progressing. SLMs have the advantage of being able to run as on-device AI on smartphones or edge devices with limited memory and computing resources, enabling their application in various fields from a commercialization perspective. This paper examines the technical features for developing SLMs, lightweight technologies, semiconductor technology development trends for on-device AI, and potential applications across various industries.

• Korean Journal of Cognitive Science
• /
• v.22 no.2
• /
• pp.173-192
• /
• 2011

The visual system applies previously learned contextual knowledge to facilitate the perception and encoding of scenes. When extrapolation following the employment of contextual information occurs, it sometimes leads to scene memory distortion: people report as if they saw more than they actually had seen. This phenomenon is called the "boundary extension" effect (Intraub & Richardson, 1989). The present study aimed to clarify the effects of contextual information on boundary extension in a more systematic way. Based on the assumption that it is harder to extract contextual information from inverted scenes compared to intact scenes, we presented inverted scenes either during encoding or retrieval to manipulate the level of contextual information and compared the magnitude of boundary extension effect for upright versus inverted scenes. In a series of experiments, we found that scene inversion during encoding, but not during retrieval, significantly reduced boundary extension. Showing reduced memory distortion for inverted scenes, the current study directly demonstrated that access to contextual information is a critical component of scene extrapolation process.

• The Journal of the Korea Contents Association
• /
• v.16 no.2
• /
• pp.106-116
• /
• 2016

What made stand out among these police procedurals was its focus on the importance of history to understand both the past and contemporary society and culture. Many of the cases involve aspects of history that are not part of mainstream narratives and constitute the unfinished business of America. The flashback sequences throughout the show visually and affectively transport the viewer back to an earlier historical moment. What all the victims have in common is that they refused to be silent or passive in the face of circumstances they believed were wrong, or they refused to conform to dehumanising and destructive social and cultural expectation. Those promoting these lies have the power to establish their version as the power to establish their version as the "truth" of what happened in context, but viewers are positioned to see the truth as the victims experienced it and to identify with them and against the power abusers. The Cold Case detectives pursue the ghosts of American history, restoring to our view stories of lives and experiences rendered invisible and unspeakable by dominant forms of knowledge and power. The series invites viewrs to empathize with the victim and to develop an ethical commitment to social justice in the future.

• Journal of the Korean Society for Library and Information Science
• /
• v.43 no.2
• /
• pp.277-298
• /
• 2009

In order to utilize the operational metadata retrieval system which is based on the semantic association between the research outcomes of the Korean Research Memory (KRM), the systematic integration of the metadata of the research outcomes with the term data registry (DR) was required. This study presents the ontology which can be used effectively to support the KRM. It includes the concept modeling of the knowledge structure of research outcomes, information retrieval service modeling that applies an analytical approach to researchers' needs in conjunction with the term DR, and also some technical and administrative suggestions for future system management.

• Korean Chemical Engineering Research
• /
• v.43 no.1
• /
• pp.76-79
• /
• 2005

Dry etching of NiFe, NiFeCo, and Ta for magnetic random access memory (MRAM) by inductively coupled plasmas (ICPs) of $Cl_2/Ar$ has been carried out. NiFe and NiFeCo showed maximum etch rates at a particular ICP source power, but the etch rate of Ta increased with the ICP source power. The etch rates of the magnetic thin films increased with the RF chuck power, but decreased with the operating pressure and the $Cl_2$ concentration. To avoid a corrosion problem by chlorine, the etched samples were rinsed with de-ionized water for 5 minutes after etching. The etch profile showed a clean and smooth surface at 50% $Cl_2$ concentration.

• Journal of KIISE
• /
• v.42 no.7
• /
• pp.852-859
• /
• 2015

As the amount of knowledge information significantly increases, a lot of progress has been made in the studies focusing on how to reason large scale ontology effectively at the level of RDFS or OWL. These reasoning methods are divided into TBox classifications and ABox realizations. A TBox classification mainly deals with integrity and dependencies in schema, whereas an ABox realization mainly handles a variety of issues in instances. Therefore, the ABox realization is very important in practical applications. In this paper, we propose a realization method for analyzing the constraint of the specified class, so that the reasoning system automatically infers the classes to which instances belong. Unlike conventional methods that take advantage of the object oriented language based distributed file system, we propose a large scale ontology reasoning method using spark, which is a functional programming-based in-memory system. To verify the effectiveness of the proposed method, we used instances created from the Wine ontology by W3C(120 to 600 million triples). The proposed system processed the largest 600 million triples and generated 951 million triples in 51 minutes (696 K triple / sec) in our largest experiment.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2019.05a
• /
• pp.79-82
• /
• 2019

This study aims to design and implement automated SEO tools that has applied the artificial intelligence techniques for search engine optimization (SEO; Search Engine Optimization). Traditional Search Engine Optimization (SEO) on-page optimization show limitations that rely only on knowledge of webpage administrators. Thereby, this paper proposes the metadata generation system. It introduces three approaches for recommending metadata; i) Downloading the metadata which is the top of webpage ii) Generating terms which is high relevance by using bi-directional Long Short Term Memory (LSTM) based on attention; iii) Learning through the Generative Adversarial Network (GAN) to enhance overall performance. It is expected to be useful as an optimizing tool that can be evaluated and improve the online marketing processes.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.11 no.6
• /
• pp.77-87
• /
• 2001

This paper describes a design of cryptographic processor that implements the Rijndael cipher algorithm, the Advanced Encryption Standard algorithm. It can execute both encryption and decryption, and supports only 128-bit block and 128-bit keys. As the processor is implemented only one round, it must iterate 11 times to perform an encryption/decryption. We implemented the ByteSub and InvByteSub transformation using the algorithm for minimizing the increase of area which is caused by different encryption and decryption. It could reduce the memory size by half than implementing, with only ROM. We estimate that the cryptographic processor consists of about 15,000 gates, 32K-bit ROM and 1408-bit RAM, and has a throughput of 1.28 Gbps at 110 MHz clock based on Samsung 0.5um CMOS standard cell library. To our knowledge, this offers more reduced memory size compared to previously reported implementations with the same performance.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.22 no.3
• /
• pp.495-504
• /
• 2012

This paper presents various speed optimization techniques for software implementation of the HIGHT block cipher on CPUs and GPUs. We considered 32-bit and 64-bit operating systems for CPU implementations. After we applied the bit-slicing and byte-slicing techniques to HIGHT, the encryption speed recorded 1.48Gbps over the intel core i7 920 CPU with a 64-bit operating system, which is up to 2.4 times faster than the previous implementation. We also implemented HIGHT on an NVIDIA GPU equipped with CUDA, and applied various optimization techniques, such as storing most frequently used data like subkeys and the F lookup table in the shared memory; and using coalesced access when reading data from the global memory. To our knowledge, this is the first result that implements and optimizes HIGHT on a GPU. We verified that the byte-slicing technique guarantees a speed-up of more than 20%, resulting a speed which is 31 times faster than that on a CPU.