• The Journal of the Acoustical Society of Korea
• /
• v.15 no.2
• /
• pp.79-84
• /
• 1996

In this paper, adaptive correlator model using backpropagation neural network based on complex multilayer perceptron is presented for suppressing interference of narrow-band of direct sequence spread spectrum receiver in CDMA mobile communication systems. Recursive least square backpropagation algorithm with backpropagation error is used for fast convergence and better performance in adaptive correlator scheme. According to signal noise ratio and transmission power ratio, computer simulation results show that bit error ratio of adaptive correlator uswing backpropagation neural network improved than that of adaptive transversal filter of direct sequence spread spectrum considering of co-channel and narrow-band interference. Bit error ratio of adaptive correlator using backpropagation neural network is reduced about $10^{-1}$ than that of adaptive transversal filter where interference versus signal ratio is 5 dB.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.16 no.1
• /
• pp.13-22
• /
• 1991

In this paper we designed 2 bit ECC(Error Checking and Correction) circuit using Single Layer Perceptron type neural networks. We used (11, 6) block codes having 6 data bits and 8 check bits with appling cyclic hamming codes. All of the circuits are layouted by CMOs 2um double metal design rules. In the result of circuit simulation, 2 bit ECC circuit operates at 67MHz of input frequency.

• Journal of Institute of Control, Robotics and Systems
• /
• v.7 no.1
• /
• pp.1211-1217
• /
• 2001

This paper presnts a controller for the heat capacity of thermal storage systems using off-peak electricity which is composed of an identifier using neural networks and a storage time adjuster in order to store exactly the required thermal energy without loss. Since thermal storage systems have nonlinear characteristics and large time constant, even if we predict the heating load accurately, it is very difficult to store exactly the required thermal energy. Thus, in the neural network for the identifier, the adaptive learning rate for high learning speed and bit inputs based on state changes of thermal storage power source are used. Also a hardware for the controller using a microprocessor is developed. The performance of the proposed controller is shown by experiment.

• Proceedings of the IEEK Conference
• /
• 2008.06a
• /
• pp.883-884
• /
• 2008

This paper presents a method for classification of underwater transient signals using, which employs a binary image pattern of the mel-frequency cepstral coefficients(MFCC) as a feature vector and a neural network as a classifier. A feature vector is obtained by taking DCT and 1-bit quantization for the square matrix of the MFCC sequences. The classifier is a feed-forward neural network having one hidden layer and one output layer, and a back propagation algorithm is used to update the weighting vector of each layer. Experimental results with some underwater transient signals demonstrate that the proposed method is very promising for classification of underwater transient signals.

• Journal of the Korea Society of Computer and Information
• /
• v.29 no.7
• /
• pp.11-20
• /
• 2024

In this paper, we propose the BAG (Bit-width Aware Generator) and the Intermediate Layer Knowledge Distillation using Channel-wise Attention to reduce the knowledge gap between a quantized network, a full-precision network, and a generator in GDFQ (Generative Data-Free Quantization). Since the generator in GDFQ is only trained by the feedback from the full-precision network, the gap resulting in decreased capability due to low bit-width of the quantized network has no effect on training the generator. To alleviate this problem, BAG is quantized with same bit-width of the quantized network, and it can generate synthetic images, which are effectively used for training the quantized network. Typically, the knowledge gap between the quantized network and the full-precision network is also important. To resolve this, we compute channel-wise attention of outputs of convolutional layers, and minimize the loss function as the distance of them. As the result, the quantized network can learn which channels to focus on more from mimicking the full-precision network. To prove the efficiency of proposed methods, we quantize the network trained on CIFAR-100 with 3 bit-width weights and activations, and train it and the generator with our method. As the result, we achieve 56.14% Top-1 Accuracy and increase 3.4% higher accuracy compared to our baseline AdaDFQ.

• Journal of Biomedical Engineering Research
• /
• v.42 no.4
• /
• pp.167-174
• /
• 2021

The purpose of this study is to classify TIFF images, PNG images, and JPEG images using deep learning, and to compare the accuracy by verifying the classification performance. The TIFF, PNG, and JPEG images converted from chest X-ray DICOM images were applied to five deep neural network models performed in image recognition and classification to compare classification performance. The data consisted of a total of 4,000 X-ray images, which were converted from DICOM images into 16-bit TIFF images and 8-bit PNG and JPEG images. The learning models are CNN models - VGG16, ResNet50, InceptionV3, DenseNet121, and EfficientNetB0. The accuracy of the five convolutional neural network models of TIFF images is 99.86%, 99.86%, 99.99%, 100%, and 99.89%. The accuracy of PNG images is 99.88%, 100%, 99.97%, 99.87%, and 100%. The accuracy of JPEG images is 100%, 100%, 99.96%, 99.89%, and 100%. Validation of classification performance using test data showed 100% in accuracy, precision, recall and F1 score. Our classification results show that when DICOM images are converted to TIFF, PNG, and JPEG images and learned through preprocessing, the learning works well in all formats. In medical imaging research using deep learning, the classification performance is not affected by converting DICOM images into any format.

• Journal of information and communication convergence engineering
• /
• v.19 no.1
• /
• pp.22-28
• /
• 2021

Convolutional neural networks (CNNs) are one of the most frequently used artificial intelligence techniques. Among CNN-based applications, small and timing-sensitive applications have emerged, which must be reliable to prevent severe accidents. However, as the small and timing-sensitive systems do not have sufficient system resources, they do not possess proper error protection schemes. In this paper, we propose MATE, which is a low-cost CNN weight error correction technique. Based on the observation that all mantissa bits are not closely related to the accuracy, MATE replaces some mantissa bits in the weight with error correction codes. Therefore, MATE can provide high data protection without requiring additional memory space or modifying the memory architecture. The experimental results demonstrate that MATE retains nearly the same accuracy as the ideal error-free case on erroneous DRAM and has approximately 60% accuracy, even with extremely high bit error rates.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 1998.06a
• /
• pp.399-403
• /
• 1998

In this paper, we proposed a new fuzzy supervised learning algorithm. We construct, and train, a new type fuzzy neural net to model the linear activation function. Properties of our fuzzy neural net include : (1) a proposed linear activation function ; and (2) a modified delta rule for learning algorithm. We applied this proposed learning algorithm to exclusive OR,3 bit parity using benchmark in neural network and pattern recognition problems, a kind of image recognition.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.29 no.12A
• /
• pp.1333-1339
• /
• 2004

In IS-95 and IMT-2000 systems using variable code rates and constraint lengths, this paper limits code rate 1/2 and constraint length 3 and states the effective reduction of PM(Path Metric) and BM(Branch Metric) memories and arithmetic comparative calculations with appling PVSL(Prototype Vector Selecting Logic) and LVQ(Learning Vector Quantization) in neural network to simplify systems and to decode forwardly. Regardless of extension of constraint length, this paper presents the new Vierbi decoder and the appied algorithm because new structure and algorithm can apply to the existing Viterbi decoder using only uncomplicated application and verifies the rationality of the proposed Viterbi decoder through VHDL simulation and compares the performance between the proposed Viterbi decoder and the existing.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.20 no.1
• /
• pp.193-198
• /
• 2020

The various uses of the Convolutional Neural Network technology are accelerating the evolution of the computing area, but the opposite is causing serious hardware performance shortages. Neural network accelerators, next-generation memory device technologies, and high-bandwidth memory architectures were proposed as countermeasures, but they are difficult to actively introduce due to the problems of versatility, technological maturity, and high cost, respectively. This study proposes DRAM-based main memory technology that enables read operations to be completed without waiting until the end of the refresh operation using pre-stored XOR bit values, even when the refresh operation is performed in the main memory. The results showed that the proposed technique improved performance by 5.8%, saved energy by 1.2%, and improved EDP by 10.6%.