• Journal of the Korean Society for information Management
• /
• v.40 no.2
• /
• pp.115-135
• /
• 2023

The purpose of this study is to assess the effectiveness of using deep learning language models to extract references automatically and create a reference database for research reports in an efficient manner. Unlike academic journals, research reports present difficulties in automatically extracting references due to variations in formatting across institutions. In this study, we addressed this issue by introducing the task of separating references from non-reference phrases, in addition to the commonly used metadata extraction task for reference extraction. The study employed datasets that included various types of references, such as those from research reports of a particular institution, academic journals, and a combination of academic journal references and non-reference texts. Two deep learning language models, namely RoBERTa+CRF and ChatGPT, were compared to evaluate their performance in automatic extraction. They were used to extract metadata, categorize data types, and separate original text. The research findings showed that the deep learning language models were highly effective, achieving maximum F1-scores of 95.41% for metadata extraction and 98.91% for categorization of data types and separation of the original text. These results provide valuable insights into the use of deep learning language models and different types of datasets for constructing reference databases for research reports including both reference and non-reference texts.

• The Journal of the Acoustical Society of Korea
• /
• v.17 no.2E
• /
• pp.31-37
• /
• 1998

This paper proposes a feature extraction method using wavelet transform for speech recognition. Speech recognition system generally carries out the recognition task based on speech features which are usually obtained via time-frequency representations such as Short-Time Fourier Transform (STFT) and Linear Predictive Coding(LPC). In some respects these methods may not be suitable for representing highly complex speech characteristics. They map the speech features with same may not frequency resolutions at all frequencies. Wavelet transform overcomes some of these limitations. Wavelet transform captures signal with fine time resolutions at high frequencies and fine frequency resolutions at low frequencies, which may present a significant advantage when analyzing highly localized speech events. Based on this motivation, this paper investigates the effectiveness of wavelet transform for feature extraction of wavelet transform for feature extraction focused on enhancing speech recognition. The proposed method is implemented using Sampled Continuous Wavelet Transform (SCWT) and its performance is tested on a speaker-independent isolated word recognizer that discerns 50 Korean words. In particular, the effect of mother wavelet employed and number of voices per octave on the performance of proposed method is investigated. Also the influence on the size of mother wavelet on the performance of proposed method is discussed. Throughout the experiments, the performance of proposed method is discussed. Throughout the experiments, the performance of proposed method is compared with the most prevalent conventional method, MFCC (Mel0frequency Cepstral Coefficient). The experiments show that the recognition performance of the proposed method is better than that of MFCC. But the improvement is marginal while, due to the dimensionality increase, the computational loads of proposed method is substantially greater than that of MFCC.

• International Journal of Computer Science & Network Security
• /
• v.21 no.3
• /
• pp.41-47
• /
• 2021

Detection of protein-protein interactions (PPIs) remains essential for the development of therapies against diseases. Experimental studies to detect PPI are longer and more expensive. Today, with the availability of PPI data, several computer models for predicting PPIs have been proposed. One of the big challenges in this task is feature extraction. The relevance of the information extracted by some extraction techniques remains limited. In this work, we first propose an extraction method based on correlation relationships between the physicochemical properties of amino acids. The proposed method uses a correlation matrix obtained from the hydrophobicity and hydrophilicity properties that it then integrates in the calculation of the bigram. Then, we use the SVM algorithm to detect the presence of an interaction between 2 given proteins. Experimental results show that the proposed method obtains better performances compared to the approaches in the literature. It obtains performances of 94.75% in accuracy, 95.12% in precision and 96% in sensitivity on human HPRD protein data.

• Proceedings of the Safety Management and Science Conference
• /
• 2006.11a
• /
• pp.415-426
• /
• 2006

Many technical and nontechnical issues hinder enterprise wide workflow management. The most significant technical issue is the inability to deal with the heterogeneity among users, workflow types, and WFMSs. Not all users demand the same workflow functionality, so user interfaces of different levels of sophistication are required. Because workflow types cannot always be fully predefined, they often need to be adjusted or extended during execution. Unlike relational database management systems, however, each WFMS often has differing workflow metamodels. This leads to incompatibility between WFMSs, making integration into an environment comprising many heterogeneous WFMSs a troublesome and sometimes impossible task. Current Workflow system consists mainly of Database system. It contains some problems like that the integration relationship among system processes cant be expressed properly. This research has been focused on two phases that should be considered in the Workflow system. First of all, the first phase is the analysis phase; one of its role is to figure out independent execution task unit(Workflow component). The second phase is design phase that provides with the framework to execute these task units actively. The Workflow component extraction method in the analysis phase uses a analysis method called C-C Net and, in the design phase, the architecture that makes the these Workflow component executed actively is provided. Through this research, each process is divided into a task unit and more effective Workflow system could be formed by executing these units actively. Current system layer calls task units, on the other hand, the Workflow system this research implemented provides with the architecture that places a layer between them that controls task units actively.

• Speech Sciences
• /
• v.11 no.1
• /
• pp.7-20
• /
• 2004

In this paper, we propose a two-step noise compensation algorithm in feature extraction for achieving robust speech recognition. The proposed method frees us from requiring a priori information on noisy environments and is simple to implement. First, in frequency domain, the Harmonics-based Spectral Subtraction (HSS) is applied so that it reduces the additive background noise and makes the shape of harmonics in speech spectrum more pronounced. We then apply a judiciously weighted variance Feature Vector Normalization (FVN) to compensate for both the channel distortion and additive noise. The weighted variance FVN compensates for the variance mismatch in both the speech and the non-speech regions respectively. Representative performance evaluation using Aurora 2 database shows that the proposed method yields 27.18% relative improvement in accuracy under a multi-noise training task and 57.94% relative improvement under a clean training task.

• Proceedings of the Korean Society for Language and Information Conference
• /
• 2002.02a
• /
• pp.271-280
• /
• 2002

A system to assist call routing task for telephone operators at the Directorate General of Telecommunications (DGT) in Taiwan is reported in this paper. The system was developed based on DGT organization profile with description of its six divisions instead of a corpus of recorded and transcribed call-routing dialogs. An acoustic module and an information retrieval module were built specifically for this task. The construction of IR module was based on term extraction and thesaurus discovery processes. By integrating acoustic and IR module, the system achieves satisfactory performance and provides a promising approach to call routing. Simulation results indicated that the proposed algorithm outperforms standard classification methods. A working system based on the proposed approach has been implemented and experimental results are presented.

• Journal of information and communication convergence engineering
• /
• v.9 no.2
• /
• pp.161-165
• /
• 2011

Recently, According to increasing interest to original sound Karaoke instrument, MIDI type karaoke manufacturer attempt to make more cheap method instead of original recoding method. Separating technique for singing voice from music accompaniment is very useful in such equipment. We propose a system to separate singing voice from music accompaniment for stereo recordings. Our system consists of three stages. The first stage is a spectral change detector. The second stage classifies an input into vocal and non vocal portions by using GMM classifier. The last stage is a selective frequency separation stage. The results of removed by listening test from the results for computer based extraction simulation, spectrogram results show separation task successfully. Listening test with extracted MR from proposed system show vocal separating and removal task successfully.

• ETRI Journal
• /
• v.40 no.5
• /
• pp.643-652
• /
• 2018

Dysarthria is a degenerative disorder of the central nervous system that affects the control of articulation and pitch; therefore, it affects the uniqueness of sound produced by the speaker. Hence, dysarthric speaker recognition is a challenging task. In this paper, a feature-extraction method based on deep belief networks is presented for the task of identifying a speaker suffering from dysarthria. The effectiveness of the proposed method is demonstrated and compared with well-known Mel-frequency cepstral coefficient features. For classification purposes, the use of a multi-layer perceptron neural network is proposed with two structures. Our evaluations using the universal access speech database produced promising results and outperformed other baseline methods. In addition, speaker identification under both text-dependent and text-independent conditions are explored. The highest accuracy achieved using the proposed system is 97.3%.

• ETRI Journal
• /
• v.42 no.6
• /
• pp.899-911
• /
• 2020

The data in tables are accurate and rich in information, which facilitates the performance of information extraction and question answering (QA) tasks. TableQA, which is based on tables, solves problems by understanding the table structure and searching for answers to questions. In this paper, we introduce both novice and intermediate Korean TableQA tasks that involve deducing the answer to a question from structured tabular data and using it to build a question answering pair. To solve Korean TableQA tasks, we use S3-NET, which has shown a good performance in machine reading comprehension (MRC), and propose a method of converting structured tabular data into a record format suitable for MRC. Our experimental results show that the proposed method outperforms a baseline in both the novice task (exact match (EM) 96.48% and F1 97.06%) and intermediate task (EM 99.30% and F1 99.55%).

• Journal of Computational Design and Engineering
• /
• v.3 no.3
• /
• pp.286-294
• /
• 2016

In this paper we present an efficient technique for sketch-based 3D modeling using automatically extracted image features. Creating a 3D model often requires a drawing of irregular shapes composed of curved lines as a starting point but it is difficult to hand-draw such lines without introducing awkward bumps and edges along the lines. We propose an automatic alignment of a user's hand-drawn sketch lines to the contour lines of an image, facilitating a considerable level of ease with which the user can carelessly continue sketching while the system intelligently snaps the sketch lines to a background image contour, no longer requiring the strenuous effort and stress of trying to make a perfect line during the modeling task. This interactive technique seamlessly combines the efficiency and perception of the human user with the accuracy of computational power, applied to the domain of 3D modeling where the utmost precision of on-screen drawing has been one of the hurdles of the task hitherto considered a job requiring a highly skilled and careful manipulation by the user. We provide several examples to demonstrate the accuracy and efficiency of the method with which complex shapes were achieved easily and quickly in the interactive outline drawing task.