• Journal of the Korean Institute of Intelligent Systems
• /
• v.14 no.7
• /
• pp.920-926
• /
• 2004

The human being receives a new information from outside and the information shows gradual oblivion with time. But the information remains in memory and isn't forgotten for a long time if the information is read several times over. For example, we assume that we memorize a telephone number when we listen and never remind we may forget it soon, but we commit to memory long time by repeating. If the human being received new information with strong stimulus, it could remain in memory without recalling repeatedly. The moments of almost losing one's life in an accident or getting a stroke of luck are rarely forgiven. The human being can keep memory for a long time in spite of the limit of memory for the mechanism mentioned above. In this paper, we propose a model to explain the mechanism mentioned above using a neural network and fuzzy.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.15 no.3
• /
• pp.837-852
• /
• 2021

Recurrent neural network (RNN) architectures have been used for language modeling (LM) tasks that require learning long-range word or character sequences. However, the RNN architecture is still suffered from unstable gradients on long-range sequences. To address the issue of long-range sequences, an attention mechanism has been used, showing state-of-the-art (SOTA) performance in all LM tasks. A differentiable neural computer (DNC) is a deep learning architecture using an attention mechanism. The DNC architecture is a neural network augmented with a content-addressable external memory. However, in the write operation, some information unrelated to the input word remains in memory. Moreover, DNCs have been found to perform poorly with low numbers of weight parameters. Therefore, we propose a robust memory deallocation method using a limited retention vector. The limited retention vector determines whether the network increases or decreases its usage of information in external memory according to a threshold. We experimentally evaluate the robustness of a DNC implementing the proposed approach according to the size of the controller and external memory on the enwik8 LM task. When we decreased the number of weight parameters by 32.47%, the proposed DNC showed a low bits-per-character (BPC) degradation of 4.30%, demonstrating the effectiveness of our approach in language modeling tasks.

• International Journal of Advanced Culture Technology
• /
• v.5 no.4
• /
• pp.57-62
• /
• 2017

In this paper, we applied long short term memory(LSTM) for classifying political polarity in cyber public sphere. The data collected from the cyber public sphere is transformed into word corpus data through word embedding. Based on this word corpus data, we train recurrent neural network (RNN) which is connected by LSTM's. Softmax function is applied at the output of the RNN. We conducted our proposed system to obtain experimental results, and we will enhance our proposed system by refining LSTM in our system.

• 한국어정보학회:학술대회논문집
• /
• 2016.10a
• /
• pp.119-121
• /
• 2016

최근 사람과 컴퓨터가 대화를 하는 채팅시스템 연구가 활발해지고 있다. 컴퓨터가 사람의 말에 적절한 응답을 하기 위해선 그 의미를 분석할 필요가 있다. 발화에 대한 의미 분석의 기본이 되는 연구로 감정분석과 화행분석이 있다. 그러나 이 둘은 서로 밀접한 연관이 있음에도 불구하고 함께 분석하는 연구가 시도되지 않았다. 본 연구에서는 Long Short-term Memory(LSTM)를 이용하여 대화체 문장의 감정과 화행, 서술자를 동시에 분석하는 통합 대화 분석모델을 제안한다. 사랑 도메인 데이터를 사용한 실험에서 제안 모델은 감정 58.08%, 화행 82.60%, 서술자 62.74%의 정확도(Accuracy)를 보였다.

• Nuclear Engineering and Technology
• /
• v.50 no.4
• /
• pp.582-588
• /
• 2018

Accident diagnosis is one of the complex tasks for nuclear power plant (NPP) operators. In abnormal or emergency situations, the diagnostic activity of the NPP states is burdensome though necessary. Numerous computer-based methods and operator support systems have been suggested to address this problem. Among them, the recurrent neural network (RNN) has performed well at analyzing time series data. This study proposes an algorithm for accident diagnosis using long short-term memory (LSTM), which is a kind of RNN, which improves the limitation for time reflection. The algorithm consists of preprocessing, the LSTM network, and postprocessing. In the LSTM-based algorithm, preprocessed input variables are calculated to output the accident diagnosis results. The outputs are also postprocessed using softmax to determine the ranking of accident diagnosis results with probabilities. This algorithm was trained using a compact nuclear simulator for several accidents: a loss of coolant accident, a steam generator tube rupture, and a main steam line break. The trained algorithm was also tested to demonstrate the feasibility of diagnosing NPP accidents.

• The Korean Journal of Applied Statistics
• /
• v.21 no.6
• /
• pp.1037-1044
• /
• 2008

In this paper, we show the performance of the power transformation GARCH(PGARCH) model to analyze the internet traffic data. The long memory property which is the typical characteristic of internet traffic data can be explained by the PGARCH model rather than the linear GARCH model. Small simulation and the analysis of the real internet traffic show the out-performance of the PARCH MODEL over the linear GARCH one.

• JSTS:Journal of Semiconductor Technology and Science
• /
• v.14 no.6
• /
• pp.755-759
• /
• 2014

We propose an integrate-and-fire neuron circuit and synaptic devices with the floating body MOSFETs. The synaptic devices consist of a floating body MOSFET to imitate biological synaptic characteristics. The synaptic learning is performed by hole accumulation. The synaptic device has short-term and long-term memory in a single silicon device. I&F neuron circuit emulate the biological neuron characteristics such as integration, threshold triggering, output generation, and refractory period, using floating body MOSFET. The neuron circuit sends feedback signal to the synaptic transistor for long-term memory.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2018.10a
• /
• pp.477-479
• /
• 2018

This study was carried out to utilize LSTM(Long Short-Term Memory) technique which is one kind of artificial neural network. Among the 27 types of motion data released by the UTD(University of Texas at Dallas), 3-axis acceleration and angular velocity data were applied to the basic LSTM and Deep Residual Bidir-LSTM techniques to classify the behavior.

• Annual Conference on Human and Language Technology
• /
• 2016.10a
• /
• pp.119-121
• /
• 2016

최근 사람과 컴퓨터가 대화를 하는 채팅시스템 연구가 활발해지고 있다. 컴퓨터가 사람의 말에 적절한 응답을 하기 위해선 그 의미를 분석할 필요가 있다. 발화에 대한 의미 분석의 기본이 되는 연구로 감정분석과 화행분석이 있다. 그러나 이 둘은 서로 밀접한 연관이 있음에도 불구하고 함께 분석하는 연구가 시도 되지 않았다. 본 연구에서는 Long Short-term Memory(LSTM)를 이용하여 대화체 문장의 감정과 화행, 서술자를 동시에 분석하는 통합 대화 분석모델을 제안한다. 사랑 도메인 데이터를 사용한 실험에서 제안 모델은 감정 58.08%, 화행 82.60%, 서술자 62.74%의 정확도(Accuracy)를 보였다.

• ETRI Journal
• /
• v.43 no.2
• /
• pp.288-298
• /
• 2021

Abstractive text summarization is a process of making a summary of a given text by paraphrasing the facts of the text while keeping the meaning intact. The manmade summary generation process is laborious and time-consuming. We present here a summary generation model that is based on multilayered attentional peephole convolutional long short-term memory (MAPCoL; LSTM) in order to extract abstractive summaries of large text in an automated manner. We added the concept of attention in a peephole convolutional LSTM to improve the overall quality of a summary by giving weights to important parts of the source text during training. We evaluated the performance with regard to semantic coherence of our MAPCoL model over a popular dataset named CNN/Daily Mail, and found that MAPCoL outperformed other traditional LSTM-based models. We found improvements in the performance of MAPCoL in different internal settings when compared to state-of-the-art models of abstractive text summarization.