• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.28 no.10C
• /
• pp.936-941
• /
• 2003

Density evolution was developed as a method for computing the capacity of low-density parity-check(LDPC) codes under the sum-product algorithm [1]. Based on the assumption that the passed messages on the belief propagation model can be approximated well by Gaussian random variables, a modified and simplified version of density evolution technique was introduced in [2]. Recently, the min-sum algorithm was applied to the density evolution of LDPC codes as an alternative decoding algorithm in [3]. Next question is how the min-sum algorithm is combined with a Gaussian approximation. In this paper, the capacity of various rate LDPC codes is obtained using the min-sum algorithm combined with the Gaussian approximation, which gives a simplest way of LDPC code analysis. Unlike the sum-product algorithm, the symmetry condition [4] is not maintained in the min-sum algorithm. Therefore, the variance as well as the mean of Gaussian distribution are recursively computed in this analysis. It is also shown that the min-sum threshold under a gaussian approximation is well matched to the simulation results.

• The Transactions of the Korea Information Processing Society
• /
• v.7 no.6
• /
• pp.1852-1860
• /
• 2000

Distinct from the Markov random field or pseudo 2D HMM models for image analysis, this paper proposes a new model of 2D hidden Markov mesh chain(HMMM) model which subsumes the definitions of and the assumptions underlying the conventional HMM. The proposed model is a new theoretical realization of 2D HMM with the causality of top-down and left-right progression and the complete lattice constraint. These two conditions enable an efficient mesh decoding for model estimation and a recursive maximum likelihood estimation of model parameters. Those algorithms are developed in theoretical perspective and, in particular, the training algorithm, it is proved, attains the optimal set of parameters.

• The Journal of the Acoustical Society of Korea
• /
• v.23 no.1
• /
• pp.75-81
• /
• 2004

Gaussian selection (GS) is a popular approach in the continuous density hidden Markov model for fast decoding. It enables fast likelihood computation by reducing the number of Gaussian components calculated. In this paper, we propose a new GS method for the phonetic tied-mixture (PTM) hidden Markov models. The PTM model can represent each state of the same topological location with a shared set of Gaussian mixture components and contort dependent weights. Thus the proposed method imposes constraint on the weights as well as the number of Gaussian components to reduce the computational load. Experimental results show that the proposed method reduces the percentage of Gaussian computation to 16.41%, compared with 20-30% for the conventional GS methods, with little degradation in recognition.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.20 no.5
• /
• pp.895-901
• /
• 2016

The next generation wireless and satellite communications require high transmission efficiency and high reliability to provide various services with subscribers. To satisfied these requirements, incorporated MIMO (Multiple Input Multiple Output) system with FTN (Faster Than Nyquist) techniques based on layered space time coded method are considered in the paper. To improve performance, STTC (Space Time Trellis Code) was employed as an inner code. As the same as SISO (Single Input Single Output) system, the outer codes are turbo codes. In receiver side, BCJR algorithm is used for STTC decoding in order to eliminate interferences induced by FTN transmission. They can yield significantly increased the data rates and improved link reliability without additional bandwidth. Therefore, we proposed a new decoding model for MIMO FTN model and confirmed that performance was improved compared to conventional SISO model according to amount of interference for FTN.

• International Journal of Internet, Broadcasting and Communication
• /
• v.12 no.2
• /
• 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.

• MALSORI
• /
• no.65
• /
• pp.93-103
• /
• 2008

In this paper, we propose a method for optimizing a multiple pronunciation dictionary used for modeling pronunciation variations of non-native speech. The proposed method removes some confusable pronunciation variants in the dictionary, resulting in a reduced dictionary size and less decoding time for automatic speech recognition (ASR). To this end, a confusability measure is first defined based on the Levenshtein distance between two different pronunciation variants. Then, the number of phonemes for each pronunciation variant is incorporated into the confusability measure to compensate for ASR errors due to words of a shorter length. We investigate the effect of the proposed method on ASR performance, where Korean is selected as the target language and Korean utterances spoken by Chinese native speakers are considered as non-native speech. It is shown from the experiments that an ASR system using the multiple pronunciation dictionary optimized by the proposed method can provide a relative average word error rate reduction of 6.25%, with 11.67% less ASR decoding time, as compared with that using a multiple pronunciation dictionary without the optimization.

• Journal of Computing Science and Engineering
• /
• v.4 no.3
• /
• pp.207-224
• /
• 2010

This paper presents a compression scheme for mesh geometry, which is suitable for mobile graphics. The main focus is to enable real-time decoding of compressed vertex positions while providing reasonable compression ratios. Our scheme is based on local quantization of vertex positions with mesh partitioning. To prevent visual seams along the partitioning boundaries, we constrain the locally quantized cells of all mesh partitions to have the same size and aligned local axes. We propose a mesh partitioning algorithm to minimize the size of locally quantized cells, which relates to the distortion of a restored mesh. Vertex coordinates are stored in main memory and transmitted to graphics hardware for rendering in the quantized form, saving memory space and system bus bandwidth. Decoding operation is combined with model geometry transformation, and the only overhead to restore vertex positions is one matrix multiplication for each mesh partition. In our experiments, a 32-bit floating point vertex coordinate is quantized into an 8-bit integer, which is the smallest data size supported in a mobile graphics library. With this setting, the distortions of the restored meshes are comparable to 11-bit global quantization of vertex coordinates. We also apply the proposed approach to compression of vertex attributes, such as vertex normals and texture coordinates, and show that gains similar to vertex geometry can be obtained through local quantization with mesh partitioning.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.28 no.6C
• /
• pp.570-579
• /
• 2003

The performance of turbo codes is investigated in the direct detection optical PPM channel. We assume that an ideal photon counter is used as an optical detector and that the channel has background noise as well as quantum noise. Resulting channel model is M-ary PPM Poisson channel. We propose the structure of the transmitter and receiver for applying turbo codes to this channel. We also derive turbo decoding algorithm for the proposed coding system, by modifying the calculation of the branch metric inherent in the original turbo decoding algorithm developed for the AWGN channel. Analytical bounds are derived and computer simulation is performed to analyze the performance of the proposed coding scheme, and the results are compared with the performances of Reed-Solomon codes and convolutional codes.

• ETRI Journal
• /
• v.27 no.6
• /
• pp.691-696
• /
• 2005

In asynchronous transfer mode (ATM) networks, fixed length cells of 53 bytes are transmitted. A cell may be discarded during transmission due to buffer overflow or a detection of errors. Cell discarding seriously degrades transmission quality. The quality degradation can be reduced by employing efficient forward error control (FEC) to recover discarded cells. In this paper, we present the design and implementation of decoding equipment for FEC in ATM networks based on a single parity check (SPC) product code using very-large-scale integration (VLSI) technology. FEC allows the destination to reconstruct missing data cells by using redundant parity cells that the source adds to each block of data cells. The functionality of the design has been tested using the Model Sim 5.7cXE Simulation Package. The design has been implemented for a $5{\times}5$ matrix of data cells in a Virtex-E XCV 3200E FG1156 device. The simulation and synthesis results show that the decoding function can be completed in 81 clock cycles with an optimum clock of 56.8 MHz. A test bench was written to study the performance of the decoder, and the results are presented.

• Proceedings of the IEEK Conference
• /
• 2000.11a
• /
• pp.165-168
• /
• 2000

The bit error rate(BER)of the data passed through Wideband-Code Division Multiple Access (W-CDMA) system with turbo-codes structure is presented. The performance of turbo-codes under W-CDMA system is estimated for various users and iteration numbers of decoding. The channel model is Additive White Gaussian Noise(AWGN) and Rayleigh fading channel. When iteration number increases, bit error probability of turbo-codes decreases. and when the number of users increase, bit error probability of turbo-codes increases.