• Journal of the Korean Institute of Telematics and Electronics S
• /
• v.35S no.2
• /
• pp.70-78
• /
• 1998

Spike detection in long-term EEG monitoring forepilepsy by wavelet transform(WT), artificial neural network(ANN) and the expert system is presented. First, a small set of wavelet coefficients is used to represent the characteristics of a singlechannel epileptic spikes and normal activities. In this stage, two parameters are also extracted from the relation between EEG activities before the spike event and EEG activities with the spike. then, three-layer feed-forward network employing the error back propagation algorithm is trained and tested using parameters obtained from the first stage. Spikes are identified in individual EEG channels by 16 identical neural networks. Finally, 16-channel expert system based on the context information of adjacent channels is introducedto yield more reliable results and reject artifacts. In this study, epileptic spikes and normal activities are selected from 32 patient's EEG in consensus among experts. The result showed that the WT reduced data input size and the preprocessed ANN had more accuracy than that of ANN with the same input size of raw data. Ina clinical test, our expert rule system was capable of rejecting artifacts commonly found in EEG recodings.

• Proceedings of the KOSOMBE Conference
• /
• v.1998 no.11
• /
• pp.207-208
• /
• 1998

Classification of extracellularly recorded action potential into each unit is an important procedure for further analysis of spike trains as point process. We utilize feedforward neural network structures, multilayer perceptron and radial basis function network to implement spike classifier. For the efficient training of classifiers, nonlinear energy operator that can trace the instantaneous frequency as well as the amplitude of the input signal is used. Trained classifiers shows successful operation, up to 90% correct classification was possible under 1.2 of signal-to-noise ratio.

• The Journal of Korea Robotics Society
• /
• v.1 no.2
• /
• pp.163-171
• /
• 2006

This paper introduces the research progress on the artificial brain in the Telerobotics and Control Laboratory at KAIST. This series of studies is based on the assumption that it will be possible to develop an artificial intelligence by copying the mechanisms of the animal brain. Two important brain mechanisms are considered: spike-timing dependent plasticity and dopaminergic plasticity. Each mechanism is implemented in two coding paradigms: spike-codes and rate-codes. Spike-timing dependent plasticity is essential for self-organization in the brain. Dopamine neurons deliver reward signals and modify the synaptic efficacies in order to maximize the predicted reward. This paper addresses how artificial intelligence can emerge by the synergy between self-organization and reinforcement learning. For implementation issues, the rate codes of the brain mechanisms are developed to calculate the neuron dynamics efficiently.

• Proceedings of the KOSOMBE Conference
• /
• v.1997 no.05
• /
• pp.358-361
• /
• 1997

This paper describes a multichannel epileptic seizure detection algorithm based on wavelet transform(WT), artificial neural network(ANN) and expert system. First, through the WT, a small number of wavelet coefficients is used to represent the single channel epileptic spike. Next, 3-layer feed-forward network employing the error back propagation algorithm is trained and tested using parameters obtained above. Finally, 16 channel expert system which is based on clinical experience is introduced as a artifact rejection and reliable detection. The suggested algorithm was implemented on personal computer(PC). Two main events i.e., epileptiform and normal activities, were selected from 32 person's EEGs(normal: 20, seizure disorder: 12) in consensus among experts. The result was that WT reduced data input size and ANN detected 97 of the 100 EEGs containing definite spike - sensitivity of 97%. Expert rule system was capable of rejecting a wide variety of artifacts commonly found in EEG recordings. It also reduced false positive detections of ANN.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.21 no.3
• /
• pp.107-114
• /
• 2021

Recently artificial intelligence technology has been introduced in various fields and various machine learning models have been operated in various frameworks as academic interest has increased. However, these frameworks have different data formats, which lack interoperability, and to overcome this, the open neural network exchange format, ONNX, has been proposed. In this paper we describe how to transform multiple machine learning models to ONNX, and propose algorithms and inference systems that can determine machine learning techniques in an integrated ONNX format. Furthermore we compare the inference results of the models before and after the ONNX transformation, showing that there is no loss or performance degradation of the learning results between the ONNX transformation.

• Proceedings of the Korea Multimedia Society Conference
• /
• 1998.04a
• /
• pp.385-388
• /
• 1998

The main interest of previous researches for shape from shading was only on Lambertian surface which consists of diffuse reflectance surface. In practice, natural objects have hybrid reflectance, which limits the recovery of shape from intensity distribution[1][2]. In this paper, we propose the method of determining shape using neural network and diffuse illumination. The segmented region of sphere surface is used for training set, which can be determined by shadow line and edge of surface. Diffuse illumination is used to avoid specular spike and highlight which usually cause many problems such as intensity disparities. Diffuse illumination method using flat paper allows us to reduce these phenomena with simple scheme. Neural network and Diffuse illumination method are useful for shape from shading, because it can be applied to objects of unknown reflectance properties, but it is applied only to Torrance-Sparrow style reflectance.

• The Korean Journal of Physiology
• /
• v.27 no.1
• /
• pp.79-91
• /
• 1993

The present study was carried out to determine the effects of cocaine (0.25, 1.0, 10.0 mg/kg, i.p.) on the interactions between spontaneously active neurons within ensembles of simultaneously recorded neurons in the primary somatosensory cortex (Sl, n= 20) and the ventroposterolateral (VPL, n= 16) thalamic nucleus of awake rats. Spike triggered cross correlation histograms were constructed between pairs of simultaneously recorded neurons. Among 101 neuronal pairs analyzed, 22.7% showed correlations indicative of various functional connections among the cortical cells, two corticothalamic interactions and one thalamocortical excitatory interaction. There were also 15 cofiring activities among SI cortical cells. These functional connectivities appeared to be modulated (weakened, abolished, or strengthened) during the 5 to 30 min following cocaine injection. The effects of saline were tested as a control, but it did not appear to alter the functional connectivities. In general, cocaine-induced changes of the functional interactions were mainly due to the concomitant alterations of the uncorrelated background discharges. These results suggest that the biphasic effects of cocaine on the spontaneously established neural networks among the SI cortical and the VPL thalamic cells of conscious rat were mainly indirect. However, various changes of the functional interactions by different doses of cocaine appeared to be a possible neural network mechanism for the cocaine induced modulation of afferent somatosensory transmission.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.16 no.1
• /
• pp.334-349
• /
• 2022

The effectiveness of Wi-Fi networks is greatly influenced by the optimization of contention window (CW) parameters. Unfortunately, the conventional approach employed by IEEE 802.11 wireless networks is not scalable enough to sustain consistent performance for the increasing number of stations. Yet, it is still the default when accessing channels for single-users of 802.11 transmissions. Recently, there has been a spike in attempts to enhance network performance using a machine learning (ML) technique known as reinforcement learning (RL). Its advantage is interacting with the surrounding environment and making decisions based on its own experience. Deep RL (DRL) uses deep neural networks (DNN) to deal with more complex environments (such as continuous state spaces or actions spaces) and to get optimum rewards. As a result, we present a new approach of CW control mechanism, which is termed as contention window threshold (CW_Threshold). It uses the DRL principle to define the threshold value and learn optimal settings under various network scenarios. We demonstrate our proposed method, known as a smart exponential-threshold-linear backoff algorithm with a deep Q-learning network (SETL-DQN). The simulation results show that our proposed SETL-DQN algorithm can effectively improve the throughput and reduce the collision rates.

• Journal of Software Assessment and Valuation
• /
• v.17 no.2
• /
• pp.47-57
• /
• 2021

Neuromorphic technology is proposed to complement the shortcomings of existing artificial intelligence technology by mimicking the human brain structure and computational process with hardware. NA-IDE has also been proposed for developing neuromorphic hardware-based IoT applications. To implement an SNN model in NA-IDE, commonly used input data must be transformed for use in the SNN model. In this paper, we implemented a neural coding method encoder component that converts image data into a spike train signal and uses it as an SNN input. The decoder component is implemented to convert the output back to image data when the SNN model generates a spike train signal. If the decoder component uses the same parameters as the encoding process, it can generate static data similar to the original data. It can be used in fields such as image-to-image and speech-to-speech to transform and regenerate input data using the proposed encoder and decoder.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.20 no.1
• /
• pp.163-169
• /
• 2020

Recently, research on automation and unmanned operation of machines in the industrial field has been conducted with the advent of AI, Big data, and the IoT, which are the core technologies of the Fourth Industrial Revolution. The machines for these automation processes are controlled based on the data collected from the sensors attached to them, and further, the processes are managed. Conventionally, the abnormalities of sensors are periodically checked and managed. However, due to various environmental factors and situations in the industrial field, there are cases where the inspection due to the failure is not missed or failures are not detected to prevent damage due to sensor failure. In addition, even if a failure occurs, it is not immediately detected, which worsens the process loss. Therefore, in order to prevent damage caused by such a sudden sensor failure, it is necessary to identify the failure of the sensor in an embedded system in real-time and to diagnose the failure and determine the type for a quick response. In this paper, a deep neural network-based fault diagnosis system is designed and implemented using Raspberry Pi to classify typical sensor fault types such as erratic fault, hard-over fault, spike fault, and stuck fault. In order to diagnose sensor failure, the network is constructed using Google's proposed Inverted residual block structure of MobilieNetV2. The proposed scheme reduces memory usage and improves the performance of the conventional CNN technique to classify sensor faults.