• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2001년도 ICCAS
• /
• pp.115.3-115
• /
• 2001

In this paper, discrete-time cellular neural networks are designed in order to function as associative memories by using Hebbian learning rule and non-cloning template. The proposed method has a very simple structure to design and to learn. Weights are updated by the connection between the neuron and its neighborhood. In the simulation, the proposed method is applied to the classification of a traffic sign pattern.

• International Journal of Internet, Broadcasting and Communication
• /
• 제12권2호
• /
• pp.45-50
• /
• 2020

A Language model with neural networks commonly trained with likelihood loss. Such that the model can learn the sequence of human text. State-of-the-art results achieved in various language generation tasks, e.g., text summarization, dialogue response generation, and text generation, by utilizing the language model's next token output probabilities. Monotonous and boring outputs are a well-known problem of this model, yet only a few solutions proposed to address this problem. Several decoding techniques proposed to suppress repetitive tokens. Unlikelihood training approached this problem by penalizing candidate tokens probabilities if the tokens already seen in previous steps. While the method successfully showed a less repetitive generated token, the method has a large memory consumption because of the training need a big vocabulary size. We effectively reduced memory footprint by encoding words as sequences of subword units. Finally, we report competitive results with token level unlikelihood training in several automatic evaluations compared to the previous work.

• 생물정신의학
• /
• 제19권1호
• /
• pp.1-8
• /
• 2012

The mechanism of psychotherapy is explained by the recent developments in neuroscience and neuroimaging. The purpose of this study is to understand the nature of psychotherapy and to discuss the future of psychotherapy improvement with the help of advances of the neurobiological findings in psychotherapy. For this study, we investigated a wide range of materials. We searched for various researches on psychotherapy, brain, and neurobiology. In addition to the conventional psychodynamic psychotherapy, we investigated research findings on cognitive behavioral therapy, interpersonal psychotherapy and eye movement desensitization and reprocessing (EMDR). Moreover, based on the actual experiences of treating patients, we speculated the neurobiological mechanisms of the process and results of psychotherapy. With the development of neuroscience, we are now able to understand the personal consciousness, unconsciousness and developmental process. Also subdividing the disease is made possible. Personalized treatment has become available, and we are able to predict the prognosis of patients. Our memories are composed by implicit memory and explicit memory. By psychotherapy, we can consciously remember explicit memory, and it becomes easier to explore implicit memory through free association. Through psychotherapy, we will also be able to learn the effect of acquired environment and experience. Psychotherapy is able to correct human behaviors by modifying the memories. Through the regulation of emotions, it becomes possible to modify the memories and correct the behaviors. In this process, doctor-patient relationship is the main factor which cause positive treatment effects. Furthermore imagination therapy or unconscious, non-verbal stimuli could bring about positive treatment effects. Now psychotherapy could be explained and studied by neuroscientific researches. In this sense, we could provide the direction of future advances in neuroscience by the neurobiological understanding of psychotherapy.

• 한국비블리아학회지
• /
• 제23권2호
• /
• pp.45-68
• /
• 2012

기록물은 직간접적으로 기억과 관련되어 있다. 한 사회가 과거를 어떻게 기억하는지는 무엇이 기록관에 수집되고, 그것이 이용자들에게 어떻게 이해되고 이용되는지에 달려있다. 기억은 종종 현시점의 개인적, 사회적 관심에 의해 촉발된다. 따라서 기록물의 해석은 현재의 관심에 따라 달라질 수 있다. 이런 관점에서, 무엇이 이용자들로 하여금 기록물을 찾게 만들고 이용자들이 기록물을 어떻게 이용하는 지의 이용맥락을 안다면, 기록물/기록관이 사회의 기억에 어떤 영향을 미치는지를 알 수 있을 것이다. 웹에서는 이용자들간, 이용자와 기록물/기록관 간의 관계를 쉽게 관찰할 수 있다. 기록관들은웹 2.0 기술 및 디지털 문화를 이용하여 이용자들이 기록물과 어떤 상호작용하는지, 기록에 어떤 기여를 하는지를 관찰함으로써, 사회의 기억을 위한 기록관으로 새로운 자리매김을 할 수 있을 것이다. 본 연구는 기록학이라는 관점에서 기억에 관한 담론을 조명하고, 특히 웹 2.0이라는 새로운 환경에서 어떻게 기록관이 기억을 위한 발판이 될 수 있는지에 대해 논하였다. 이용자와 이용맥락에 중점을 두어 기록학문헌에 비추어진 집단 기억을 논하고, 집단기억이 기록물, 역사, 증거라는 담론과 어떻게 연결되어 설명되어왔는지를 개관하였다. 이러한 이론적 배경을 바탕으로, 웹 2.0 기술이 집단기억을 위해 어떤 기술적인 발판을 제공하는지를 고찰하였다. 또한, 기록물을 둘러싼 포괄적인 내러티브를 만들어가는데 이용자들이 웹 2.0 어플리케이션을 통해 어떻게 자신의 기억을 나누고 집단 기억을 만들어가는지에 관한 사례를 살펴보았다.

• International Journal of Contents
• /
• 제15권3호
• /
• pp.32-38
• /
• 2019

We usually suffer from difficulties in treating or managing Big Data generated from various digital media and/or sensors using traditional mining techniques. Additionally, there are many problems relative to the lack of memory and the burden of the learning curve, etc. in an increasing capacity of large volumes of text when new data are continuously accumulated because we ineffectively analyze total data including data previously analyzed and collected. In this paper, we propose a general-purpose classifier and its structure to solve these problems. We depart from the current feature-reduction methods and introduce a new scheme that only adopts changed elements when new features are partially accumulated in this free-style learning environment. The incremental learning module built from a gradually progressive formation learns only changed parts of data without any re-processing of current accumulations while traditional methods re-learn total data for every adding or changing of data. Additionally, users can freely merge new data with previous data throughout the resource management procedure whenever re-learning is needed. At the end of this paper, we confirm a good performance of this method in data processing based on the Big Data environment throughout an analysis because of its learning efficiency. Also, comparing this algorithm with those of NB and SVM, we can achieve an accuracy of approximately 95% in all three models. We expect that our method will be a viable substitute for high performance and accuracy relative to large computing systems for Big Data analysis using a PC cluster environment.

• 인지과학
• /
• 제14권1호
• /
• pp.1-9
• /
• 2003

본 연구에서는 EPAM(Elementary Perceiver and Memorizer) Ⅳ를 사용하여 인공 문법이 사용된 암묵적 학습에서의 인간 수행을 모사하였다. 암묵 학습(implicit learning) 과제에서 참가자들은 인공 문법(rtificial grammar)을 사용해 만들어진 '문법적' 문자열과 무선적으로 만들어진 '비문법적' 문자열을 학습하였는데, 이 때 비문법적 문자열보다 문법적 문자열의 학습이 더 우수하였다. 또한 참가자들은 이전에 본 적이 없었던 새로운 문자열에 대해서도 그 문법성을 판단할 수 있었다. 단순 기억 시스템인 EPAM Ⅳ에 항목 내 군집화(within-item chunking) 기능을 추가하여 암묵 학습 과제에서의 인간수행을 모사한 결과, EPAM Ⅳ 또한 무선적인 문자열보다 문법적인 문자열을 보다 잘 학습하였고, 비문법적 문자열과 문법적 문자열을 구별할 수 있었다. 이러한 결과는 인공 문법을 사용한 암묵 학습 과제에서의 수행이 규칙 추상화보다는 군집화(chunking)에 근거한 재인 기억을 바탕으로 이루어짐을 시사한다.

• 동의생리병리학회지
• /
• 제17권1호
• /
• pp.57-63
• /
• 2003

Background: Rehmanniae radix preparata is a medicinal herb widely used in Asian countries, to replenish Yin and Vital Essence through strengthen the kidney function. In Oriental Medicine the brain has physiological relationship with the Vital Essence. And central nerve system (CNS) is also refered to as 'Sea of the Vital Essence'. So if the Vital Essence is strong it will nourish the brain and memory will all be keen. The purpose of this study was to investigate the effect of 7 to 14 days treatment with Rehmanniae radix preparata (3.4 g /100 g, per os) on retention performance of rats in a passive avoidance situation. Methods: Male Sprague-Dawley rats were trained on a one-trial passive avoidance task using a two-way shuttle box by giving a foot shock. They were tested for retention of 6h, 24h, 72h, 168h, and 336h after training. Results: Rehmanniae radix preparata treated rats showed a increased performance in retention test as compared to saline at 6hr (P<0.05) and 24hr (P<0.05). Conclusion: These data suggest that Rehmanniae radix preparata treatment can improve memory in the rat.

• 전자통신동향분석
• /
• 제34권4호
• /
• pp.43-54
• /
• 2019

Currently, a majority of artificial intelligence is used to secure big data; however, it is concentrated in a few of major companies. Therefore, automatic data augmentation and efficient learning algorithms for small-scale data will become key elements in future artificial intelligence competitiveness. In addition, it is necessary to develop a technique to learn meanings, correlations, and time-related associations of complex modal knowledge similar to that in humans and expand and transfer semantic prediction/knowledge inference about unknown data. To this end, a neural memory model, which imitates how knowledge in the human brain is processed, needs to be developed to enable knowledge expansion through modality cooperative learning. Moreover, declarative and procedural knowledge in the memory model must also be self-developed through human interaction. In this paper, we reviewed this essential methodology and briefly described achievements that have been made so far.

• 한국멀티미디어학회논문지
• /
• 제20권5호
• /
• pp.769-781
• /
• 2017

Deep learning networks like Convolutional Neural Networks (CNNs) show successful performances in many computer vision applications such as image classification, object detection, and so on. For implementation of deep learning networks in embedded system with limited processing power and memory, deep learning network may need to be simplified. However, simplified deep learning network cannot learn every possible scene. One realistic strategy for embedded deep learning network is to construct a simplified deep learning network model optimized for the scene images of the installation place. Then, automatic training will be necessitated for commercialization. In this paper, as an intermediate step toward automatic training under fisheye camera environments, we study more precise human localization in fisheye images, and propose an accurate human localization method, Automatic Ground-Truth Labelling Method (AGTLM). AGTLM first localizes candidate human object bounding boxes by utilizing GoogLeNet-LSTM approach, and after reassurance process by GoogLeNet-based CNN network, finally refines them more correctly and precisely(tightly) by applying saliency object detection technique. The performance improvement of the proposed human localization method, AGTLM with respect to accuracy and tightness is shown through several experiments.

• 비교문화연구
• /
• 제38권
• /
• pp.351-384
• /
• 2015

This paper is concerned with the effectiveness of songs in the acquisition of formulaic sequences in the college EFL classroom. The existing research mentions the use of songs in terms of the power of their melodies (Fonseca-Mora, 2000), linguistic features in song lyrics (Abbott, 2002), and the emotional basis of memory (LI & Brand, 2009). Learners' opinions about the use of songs has been ignored, however. In this paper, seven subjects with English ability ranging from advanced (one) intermediate-high (three), intermediate-middle (two), and intermediate-low (one) studied five different pop songs. The results showed that they did not agree with the existing research findings. Rather, they were negative about using songs in the classroom. Their complaints were the burden of using too many hours to memorize lyrics, few language expressions to learn, and too much emphasis on expressions about love and feelings. Students at all levels expressed similar negativity about the use of songs. When their complaints were discussed during interviews, however, their attitude changed from negative to positive. The case study in this paper was on a small-scale but it is suggested that through further research the use of songs could be activated in the EFL classroom. Considering college language learners disregard most existing EFL materials, it appears to be worthwhile to continue further with this kind of research.