Improved Error Correction Research Articles

Supporting dictation speech recognition error correction: the impact of external information

Although speech recognition technology has made remarkable progress, its wide adoption is still restricted by notable effort made and frustration experienced by users while correcting speech recognition errors. One of the promising ways to improve error correction is by providing user support. Although support mechanisms have been proposed for inline speech recognition error correction, how to develop user support for third-party error correction has been under studied. To address the unique challenges associated with third-party error correction, external information that is obtained outside of an erroneous sentence was employed in this research. Specifically, three types of external information were selected, including word alternative hypotheses, noisy context and accurate context, and their impacts on dictation speech recognition error correction were assessed empirically. As expected, results revealed the importance of context information to improving both the performance and perception of error correction. Additionally, the results also provided insights into possible effects of word error rate and sentence length on error correction. These findings have significant implications to interface design for transcript proofreading systems.

채널상태정보를 이용하는 블록터보 부호화된 OFDM 시스템의 성능 분석

AbstractIn this paper, the new decoding algorithm of Block Turbo Codes using Channel State Information(CSI), which is estimated to compensate for the distorted signal caused by multi-path fading, is proposed in order to improve error correction capacity during decoding procedure in OFDM system. The performance of the new decoding algorithm is compared to that of the conventional decoding algorithm without using channel state information under the Rayleigh fading channel. Experimental results showed that in case of only one iteration coding gains of up to 5.0dB~9.0dB can be obtained by applying the channel state information to the conventional decoding algorithm according to the modulation methods. In addition to that, the new decoding algorithm using channel state information at only one iteration shows a performance improvement of 3.5dB to 5.0dB when compared to the conventional decoding algorithm after four iterations. This leads to reduce the considerable amount of computation. Key Words : OFDM, Channel state information, Block turbo code

ONE-PARTY QUANTUM-ERROR-CORRECTING CODES FOR UNBALANCED ERRORS: PRINCIPLES AND APPLICATION TO QUANTUM DENSE CODING AND QUANTUM SECURE DIRECT COMMUNICATION

In this article, we present unbalanced-quantum-error-correcting codes (one-party QECCs) — a novel idea for correcting unbalanced quantum errors. In some quantum communication tasks using entangled pairs, the error distributions between two parts of the pairs are unbalanced, and one party holds the whole entangled pairs at the final stage, and he or she is able to perform joint measurements on the pairs. In this situation the proposed one-party QECCs can improve error correction by allowing a higher-tolerated error rate. We have established the general correspondence between linear classical codes and the one-party QECCs, and we have given the general definition for these types of quantum-error-correcting codes. It has been shown that the one-party QECCs can correct errors as long as the error threshold is not larger than 0.5. They work even for fidelity less than 0.5 as long as it is larger than 0.25. We give several concrete examples of the one-party QECCs. We provide the applications of the one-party QECCs in quantum dense coding, so that it can function in noisy channels. As a result, a large number of quantum secure direct communication protocols based on dense coding are also able to be protected by this new type of one-party QECCs.

고유 공간 빔형성 알고리즘을 이용한 AT-DMB 수신 방법

고전송률 지상파 디지털멀티미디어 방송 즉 Advanced T-DMB 방송 (이하 AT-DMB)은 기존 T-DMB와 역호환성을 유지하면서 전송률을 최대 2 배로 높이기 위한 방법의 하나로 개발 되고 있다. AT-DMB 시스템은 T-DMB의 DQPSK 변조신호에 새로운 변조가 적용된 신호를 중첩하는 계층변조를 적용 한다. AT-DMB의 계층변조에는 새로 추가하는 향상 계층신호로 이동성에 강한 BPSK 신호를 더하는 B 모드와 보다 많은 데이터 전송을 위한 QPSK 신호를 더하는 Q 모드가 있다. T-DMB와 역호환성 유지 및 방송권역 축소를 최소화하기 위해 향상 계층의 신호를 작게 유지 해야 한다. 이 때문에 새로 추가된 향상계층 신호는 전송채널에 존재하는 페이딩에 의해 쉽게 왜곡될 수 있다. 이를 위해 향상 계층의 신호에는 기존의 길쌈 부호 보다 에러 정정 능력이 뛰어난 터보 부호를 채용하고 있다. 그러나 AT-DMB의 경우 기본계층과 향상계층 사이에 존재하는 QEF(quasi error free) SNR(signal to noise ratio) 차이로 인해 방송권역에 차이가 발생할 수 있다. 따라서 본 논문에서는 별도의 송신 출력의 증가 없이 AT-DMB 신호를 안정적으로 수신할 수 있는 빔형성 이득과 다이버시티 이득을 동시에 얻을 수 있는 고유 공간 빔형성 기법을 적용한 AT-DMB 수신 방법을 제안한다. 제안한 고유 공간 빔형성 기법이 적용된 AT-DMB 수신 방법에 대해 컴퓨터 시뮬레이션을 통해 성능을 검증하였고 그 적용성을 제시하였다. AT-DMB system has been developed to increase data rate up to double of conventional T-DMB in the same bandwidth while maintaining backward compatibility. The AT-DMB system adopted hierarchical modulation which adds BPSK or QPSK signal as enhancement layer to existing DQPSK signal. The enhancement layer signal should be small enough to maintain backward compatibility and to minimize the coverage loss of conventional T-DMB service coverage. But this causes the enhancement layer signal of AT-DMB susceptible to fading effect in transmission channel. A turbo code which has improved error correction capability than convolutional code, is applied to the enhancement layer signal of the AT-DMB system for compensating channel distortion. However there is a need for other solutions for better reception of AT-DMB signal in receiver side without increasing transmitting power. In this paper, we propose adaptive array antenna system with Eigen-space beamforming algorithm which benefits beamforming gain along with diversity gain. We analyzed the reception performances of AT-DMB system in indoor and mobile environments when this new smart antenna system and algorithm is introduced. The computer simulation results are presented along with analysis comments.

Coding and Computing Join Forces

Researchers who develop theories of computation are using their insights to create improved error correction codes.

Design of Improved Error Correction Decoder Using Error Detecting Information of Modulation Code in Digital Versatile Disc Systems

We present a powerful error control decoder which can be used in all kinds of digital versatile disk (DVD) systems. The decoder exploits the error information from the modulation decoder in order to increase the error correcting capability. We can identify that the modulation decoder in DVD system can detect errors more than 60% of total errors when burst errors are occurred. In results, for a decoded block, error correcting capability of the proposed scheme is improved up to 25% more than that of the original error control decoder. Also, a pipeline-balanced Reed–Solomon Product Code (RSPC) decoder with a low hardware complexity is designed to maximize the throughput. The maximum throughput of the RSPC decoder is 740 Mbps at 100 MHz and the number of gate counts is 20.3 K for RS(182,172,11) decoder and 30.7 K for RS(208,192,17) decoder, respectively.

Data mining for detecting errors in dictation speech recognition

The efficiency promised by a dictation speech recognition (DSR) system is lessened by the need for correcting recognition errors. Error detection is the precursor of error correction. Developing effective techniques for error detection can thus lead to improved error correction. Current research on error detection has focused mainly on transcription and/or domain-specific speech. Error detection in DSR has been studied less. We propose data mining models for detecting errors in DSR. Instead of relying on internal parameters from DSR systems, we propose a loosely coupled approach to error detection based on features extracted from the DSR output. The features mainly came from two sources: confidence scores and linguistics parsing. Link grammar was innovatively applied to error detection. Three data mining techniques, including Na/spl inodot//spl uml/ve Bayes, neural networks, and Support Vector Machines (SVMs), were evaluated on 5M DSR corpora. The experimental results showed that significant performance was achieved in that F-measures for error detection ranged from 55.3% to 62.5%. This study provided insights into the merit of different data-mining techniques and different types of features in error detection.

Sensitivity to Noise in Bidirectional Associative Memory (BAM)

Original Hebbian encoding scheme of bidirectional associative memory (BAM) provides a poor pattern capacity and recall performance. Based on Rosenblatt's perceptron learning algorithm, the pattern capacity of BAM is enlarged, and perfect recall of all training pattern pairs is guaranteed. However, these methods put their emphases on pattern capacity, rather than error correction capability which is another critical point of BAM. This paper analyzes the sensitivity to noise in BAM and obtains an interesting idea to improve noise immunity of BAM. Some researchers have found that the noise sensitivity of BAM relates to the minimum absolute value of net inputs (MAV). However, in this paper, the analysis on failure association shows that it is related not only to MAV but also to the variance of weights associated with synapse connections. In fact, it is a positive monotone increasing function of the quotient of MAV divided by the variance of weights. This idea provides an useful principle of improving error correction capability of BAM. Some revised encoding schemes, such as small variance learning for BAM (SVBAM), evolutionary pseudorelaxation learning for BAM (EPRLAB) and evolutionary bidirectional learning (EBL), have been introduced to illustrate the performance of this principle. All these methods perform better than their original versions in noise immunity. Moreover, these methods have no negative effect on the pattern capacity of BAM. The convergence of these methods is also discussed in this paper. If there exist solutions, EPRLAB and EBL always converge to a global optimal solution in the senses of both pattern capacity and noise immunity. However, the convergence of SVBAM may be affected by a preset function.

The detection and correction of snow water equivalent pressure sensor errors

AbstractSnow water equivalent (SWE) sensors can experience errors when the base of the snow cover is at the melting temperature, the snow can support shear stresses (assumed to occur at densities greater than 200 kg m−3), and the rate of snowmelt on the sensor is different than on the surrounding ground. Either undermeasurement or overmeasurement errors may occur at critical times when the snow cover transitions from winter to spring conditions and at the start of periods of rapid snowmelt. Parameters to determine the onset of SWE sensor undermeasurement errors are defined by a negative rate of change for SWE, a negative rate of change for snow density, and an increasing snow depth. For the onset of overmeasurement errors, the rate of change for SWE will be positive while snow depth decreases and the snow density rate of change exceeds a defined positive threshold. When the snow temperature and density error conditions and the three under‐ or over‐measurement error‐indicator parameters are satisfied at the same time, an SWE sensor error has started. Real‐time correction of the errors is done by multiplying the average snow cover density, set at the start of the error, with the snow depth. Once the error event ends, when the corrected SWE and SWE sensor data intersect, SWE is again determined from SWE sensor measurements. SWE sensor errors were accurately detected and corrected for five different sensors located in maritime and intermountain climatic zones when high‐quality SWE sensor, snow or air temperature, and snow depth measurements were available. Implementation of the error detection and correction method requires simultaneous measurements of SWE, snow depth, and snow temperature near the ground. Improved error correction can be achieved by incorporating precipitation data and estimates of snow density due to retained rain or snow melt. Copyright © 2004 John Wiley & Sons, Ltd.

Memory annihilation of structured maps in bidirectional associative memories

Structured sets comprise Boolean vectors with equal pair-wise Hamming distances, h. An external vector, if it exists at an equidistance of h/2 from each vector of the structured set, is called the centroid of the set. A structured map is a one-one onto mapping between structured sets. It is a set of associations between Boolean vectors, where both domain and range vectors are drawn from structured sets. Associations between centroids are called centroidal associations. In this paper we show that when structured maps are encoded into bidirectional associative memories using outer-product correlation encoding, the memory of these associations annihilates under certain mild conditions. When annihilation occurs, the centroidal association emerges as a stable association, and we call it an alien attractor. For the special case of maps where h=2, self-annihilation can take place when either the domain or range dimensions are greater than five. In fact, we show that for dimensions greater than eight, as few as three associations suffice for self-annihilation. As an example shows, annihilation occurs even for the case of bipolar decoding which is well known for its improved error correction capability in such associative memory models.

Multilevel block codes with interdependency among component codes for Rayleigh fading channels

In this paper, in order to improve error correction capacity, the following method is proposed for constructing multilevel codes with interdependencies among the levels. First, one of the component codes of a basic multilevel code is divided into 2i (i = 1, 2, …) subcodes. Then, the number of codewords of the other component code is expanded by a factor of 2i and partitioned into 2i subcode groups. One-to-one assignments among the levels are made between these subcode groups, a code array is constructed by using only the codewords contained in these subcodes, and an interdependency among the levels is introduced. In QPSK and 8PSK modulation, multilevel codes that have superior performance in fading channels are determined through a computer search. Furthermore, the bit error rate performance in the Rayleigh fading channel is determined by computer simulation. Then, although the decoder complexity of the multistage decoding method is increased after the introduction of the interdependencies, the bit error rate becomes equivalent to that of maximum likelihood decoding. Compared with previously published results, it is demonstrated that codes can be constructed with higher coding gain. © 1998 Scripta Technica. Electron Comm Jpn Pt 3, 81(1): 91–100, 1998

Improved error correction technique for on-wafer lightwave measurements of photodetectors

An accurate correction technique for on-wafer small-signal lightwave measurements of photodetectors is presented. This technique is an improvement of the conventional calibration methods for on-wafer lightwave measurements. Mathematical expressions for the dominant error sources that exist in the measurement system are derived. Experimental results for an InGaAs-InP pin photodiode show a smoother modulation response characteristic when the presented technique is used.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>

Novel two-layer neural multiassociative memory model

A novel, two-layer VLSI realizable, modular and expandable neural network multiassociative memory model is proposed. When compared to the Hopfield model, it requires <i>0(N)</i> connections for <i>N</i>-bit long vectors; yields improved error correction and storage capabilities and a faster convergence rate, avoids the storage of false memories, and possesses analogies to biological neural networks.

PPM Performance for Reed-Solomon Decoding Over an Optical-RF Relay Link

Consideration is presently being given to an optical-RF relay deep space communication link that transmits optical PPM data from spacecraft to orbital relay, then retransmits the data via microwave to ground. It is generally advantageous to use Reed-Solomon (RS) encoding over the PPM optical link for improved error correction. In this paper several demodulating schemes are considered for generating the RS symbols at the relay photodetector output, and the effect of each on overall RS decoding performance is computed. Both Poisson and Gaussian optical noise models are used in the evaluation. The effect on performance of bit errors in the relay downlink is also examined.