Analysis of Novelty Filtering for Speech Signal Enhancement

( Hanseok Ko ) 고려대학교 공학기술연구소 1997 공학기술연구소보 Vol.4 No.-

A SHIPBOARD MULTISENSOR SOLUTION FOR THE DETECTION OF FAST MOVING SMALL SURFACE OBJECTS

Hanseok Ko 제어로봇시스템학회 1995 제어로봇시스템학회 국내학술대회 논문집 Vol.1995 No.10

Dynamical Behavior of Autoassociative Memory Performaing Novelty Filtering

Ko, Hanseok The Acoustical Society of Korea 1998 韓國音響學會誌 Vol.17 No.e4

This paper concerns the dynamical behavior, in probabilistic sense, of a feedforward neural network performing auto association for novelty. Networks of retinotopic topology having a one-to-one correspondence between and output units can be readily trained using back-propagation algorithm, to perform autoassociative mappings. A novelty filter is obtained by subtracting the network output from the input vector. Then the presentation of a "familiar" pattern tends to evoke a null response ; but any anomalous component is enhanced. Such a behavior exhibits a promising feature for enhancement of weak signals in additive noise. As an analysis of the novelty filtering, this paper shows that the probability density function of the weigh converges to Gaussian when the input time series is statistically characterized by nonsymmetrical probability density functions. After output units are locally linearized, the recursive relation for updating the weight of the neural network is converted into a first-order random differential equation. Based on this equation it is shown that the probability density function of the weight satisfies the Fokker-Planck equation. By solving the Fokker-Planck equation, it is found that the weight is Gaussian distributed with time dependent mean and variance.

오토 황제의 원수정 강화 정책 연구 - 원로원 및 군대와의 관계를 중심으로

고한석 ( Ko Hanseok ) 한국서양고대역사문화학회 ( 구 한국서양고대사학회 ) 2018 서양고대사연구 Vol.53 No.-

서기 69년 1월, 오토는 근위대의 지지를 바탕으로 갈바를 제거하고 제위를 쟁취하였다. 그가 보여주었던 여러 정책들은 네로 혹은 갈바, 비텔리우스와는 매우 다르게 나타났다는 점에서 새로운 관점에서 주목할 필요가 있다. 본 논문에서는 오토 황제의 치세를 면밀히 분석하여, 그가 자신의 원수정을 강화하기 위해 제국의 두 축이었던 원로원과 인민, 그리고 군대를 적극적으로 포용하려 했음을 규명하고자 했다. 즉위 직후 오토는 네로의 이미지와 자신을 연결시켜 평민의 지지를 획득하면서도, 이전에 숙청된 이들의 복권을 통해 네로와는 달리 국정 운영의 동반자로서 원로원의 권위를 인정하였다. 또한 즉위 전부터 근위대를 물심 양면으로 지원하는 한편, 즉위 이후에는 그들의 현안을 해결해줌으로써 군대의 확고한 지지를 획득하였다. 이와 더불어 군사지도자뿐만 아니라 종교 지도자로서의 이미지를 적극 활용함으로써 정권의 정당성을 적극 홍보하였다. 이는 아우구스투스가 원수정 체제를 확립해가는 과정에서 보여주었던 행보와 유사하다는 점에서, 오토의 치세에 공화정의 회복이라는 명분과 1인 지배 체제라는 실상 사이의 균형이라는 원수정의 핵심이 명확하게 드러났음을 확인할 수 있었다. In January 69 A.D., Otho removed Galba and took the throne with the support of the praetorian guards. It is worth noting that many of the his policies were very different from Nero or Galba or Vitelius. In this paper, I argue that Emperor Otho tried to embrace the two pillars of the Roman empire - the Senate and the army - to strengthen his principate by analyzing his reign and his policies. Soon after the coup, Otho connected himself with Nero’s image to win the support of the common people. However, unlike Nero, he recognized the authority of the Senate as a partner in state administration through the reinstatement of those previously purged. He also gained firm support of army by aiding the praetorian guards before the throne, and by addressing their pending issues after the throne. In addition, he actively used his image as a religious leader as well as a military leader to promote the legitimacy of his principate. This was similar to Augustus’s progress in establishing a system of the Principate, which clearly revealed the core idea of a balance between the cause of a republic’s recovery and the reality of a monarchy in the reign of Otho.

기원전 56년 루카 회동과 속주 중심의 권력구도 재편

고한석 ( Ko Hanseok ) 한국서양고대역사문화학회 ( 구 한국서양고대사학회 ) 2015 서양고대사연구 Vol.43 No.-

기원전 56년, 1차 삼두정을 결성하여 정국의 주도권을 장악했던 폼페이우스와 크라수스, 카이사르는 루카에 모여 권력을 재분배하기로 합의하였다. 삼두의 합의는 원로원이 주도하는 로마의 정치적 전통에 위배되었음에도 관철될 수 있었다. 이는 당시 지중해 제국으로 성장한 로마가 새로운 정치구도로 변화하고 있었음을 보여준다. 따라서 본 논문에서는 루카 회동을 전후로 한 로마의 정치 상황을 고찰함으로써, 이 시기를 기점으로 로마의 권력구조가 질적으로 변화하였음을 규명하고자 한다. 루카 회동을 통해 삼두는 카이사르의 갈리아 원정과 폼페이우스의 ‘곡물수급위원’직 취임 과정에서 심화된 임페리움의 불균형을 해소하고, 실제 군단을 장악하여 자신들의 권력기반을 안정화시키고자 하였다. 이후 삼두는 정치적 폭력과 키케로의 연설, 그리고 각종 입법안들을 통해 원로원을 무력화시키고 합의안을 실현할 수 있었다. 삼두의 권력 재분배는 특히 로마의 중앙 정무관직이 아닌 ‘속주’를 나누어 가짐으로써 현실화되었는데, 이탈리아 반도 밖에 자리한 로마의 점령지에 불과했던 ‘속주’가 중앙 정계의 권력 향배에 중요한 역할을 하게 되었다는 점은 주목할 만하다. 전통적으로 속주에 대한 지배권은 원로원이 장악하고 있었다. 하지만 1차 삼두정의 출현을 전후한 시기에 삼두, 특히 폼페이우스와 카이사르는 속주에서 대규모 전쟁을 벌임으로써 원로원의 통제권을 사실상 무력화시켰다. 또한 그들은 전리품의 명목으로 속주에서 막대한 부를 획득하고, 이를 정치적 지지세력 확보를 위한 정치자금으로 활용하였다. 결국, 1차 삼두정 시기가 되면 속주는 총독이 원로원의 간섭에서 벗어나 자신의 정치적 기반을 구축하는 거점으로서의 의미가 강해진다고 할 수 있다. 요컨대, 폼페이우스와 크라수스, 카이사르 사이의 루카 회동은 속주를 중심으로 하는 새로운 권력 구도의 형성을 의미한다. 속주 중심의 권력 구도는 이후 2차 삼두정과 아우구스투스의 원수정에서도 확인된다는 점에서, 루카 회동을 중심으로 하는 1차 삼두정기는 로마 시를 중심으로 하는 전통적인 공화정 질서가 지중해 세계를 아우르는 신질서로 새롭게 변화하게 되는 기점으로 이해할 수 있는 주요한 시기라 할 수 있을 것이다. In 56 BC, Pompey, Crassus, and Caesar who had already secured their leadership by organizing the First Triumvirate, gathered at Luca. Regarding that the Roman senate, the core of the Roman republic, was completely excluded in this conference, it is fair to say that new circumstances have emerged within the Roman power structure. In this paper, I argue that the Roman power structure underwent a substantial change around this period by analyzing the Roman political situation surrounding the Luca conference. Their arrangement was aimed to unravel the imbalance which had intensified in the course of Caesar’s Gallic campaign and Cura Annonae of Pompey and to stabilize the power base of the Triumvirate by taking control of the army. Subsequently the Triumvirate used political violence, Cicero’s speech, and various legislations to affect the senate and actualize their arrangement. The fact that the power of the Triumvirate was redistributed through the allocation of provinces, and that ‘province’, a location outside the Italian peninsula, and not the central magistrates of Rome, came to play an important role in the central political power struggle is worthy of notice. Pompey and Caesar virtually emasculated the senate by waging a large-scale war. Justifying the spoils as trophies of war, they also accumulated enormous wealth which was used as political funding to secure loyal supporters. In the time of the First Triumvirate, provinces acquired a new meaning - a stronghold in building a political base while eliminating the interference of the senate. In short, the conference held amongst Pompey, Crassus, and Caesar in Luca meant that a new province-based power structure emerged. This province-based power structure was maintained and secured in Roman history by such occurrences as in the Second Triumvirate and the Augustan Principate, thus confirming the suggestion that the time of the First Triumvirate was a crucial point in history, a transition from the traditional republican order to a new political order surrounding the Mediterranean world.

세계사 교과의 ‘시민 혁명’ 서술과 정형화된 근대

高翰奭(Ko, Hanseok) 역사교육연구회 2016 역사교육 Vol.140 No.-

This paper investigates how ‘civil revolution’ is treated in the world history curriculum and textbooks. Status of civil revolution in the world history curriculum has been firmly held in both the material and chapter organization aspects, and the interpretation has been sticking to traditional one centered at British, American, and French revolution. Textbook treatments also have naturally followed the precedent ways. The descriptions of civil revolution are known to be too succinct and superficial, and sometimes erroneous. Moreover, usual textbooks summarize the views presented in Introductory of Western History(『西洋史槪論』) while emphasizing the positive aspects. Formalized descriptions of civil revolution in the world history curriculum and textbooks are attributed to the status of so-called orthodox theory in academia, and how importantly recognized the development process of democracy is among the topics of world history course. In particular, even ‘new world history’ discourse which raises doubt on Eurocentrism takes the historical status of civil revolution at a face value. In conclusion, we must try to transform the shape of absolutized civil revolution in the current world history course in order to overcome the Eurocentrism. However, as the objective of world history course is strongly connected to presentation of how the modern world had formed, it is quite hard to change the current status of Eurocentrism and civil revolution in the curriculum and textbooks. So we have to come up with fresh ways to replace current world history curriculum such as ‘western history as regional history’ or ‘the history of democracy as the topical history’.

개체명 인식을 이용한 반려동물질병 질의응답시스템

오한석(Hanseok Oh),서정우(Jungwoo Seo),김해준(Haejun Kim),김효진(Hyojin Kim),김성서(Seongseo Kim),이현아(Hyunah Lee) 한국정보기술학회 2021 Proceedings of KIIT Conference Vol.2021 No.11

반려동물에 관한 관심이 높아지고 있지만 반려동물의 질병에 관한 정보를 얻기는 쉽지 않다. 본 논문에서는 반려동물의 증상을 질의로 입력하면 관련되는 반려동물 질병을 응답하는 질의응답 시스템을 제안한다. 반려동물질병 질의응답 시스템은 사용자가 입력한 문장과 유사한 제목을 가진 문서를 찾은 후 해당 문서 내에서 개체명 인식 모델을 통해 질병을 인식하여 사용자에게 출력한다. anaGo를 이용하여 구축한 질병 개체명 인식 모델을 음절 단위로 학습을 진행한 결과 0.75의 F1 Score를 얻었다. Although interest in pets is increasing, it is not easy to obtain information about pet diseases. In this paper, we propose a Q&A system of pet disease that can answer pet disease name by inputting symptoms of companion animal. The our system finds a document with a title similar to an user query, recognizes diseases in a document using a Named Entity Recognition(NER) model, and outputs it to the user. Our disease NER model was created using anaGo and trained with syllables, and the result shows 0.75 as F1 Score.

Robust Relative Transfer Function Estimation for Dual Microphone-Based Generalized Sidelobe Canceller

KIM, Kihyeon,KO, Hanseok The Institute of Electronics, Information and Comm 2009 IEICE transactions on information and systems Vol.92 No.9

<P>In this Letter, a robust system identification method is proposed for the generalized sidelobe canceller using dual microphones. The conventional transfer-function generalized sidelobe canceller employs the non-stationarity characteristics of the speech signal to estimate the relative transfer function and thus is difficult to apply when the noise is also non-stationary. Under the assumption of W-disjoint orthogonality between the speech and the non-stationary noise, the proposed algorithm finds the speech-dominant time-frequency bins of the input signal by inspecting the system output and the inter-microphone time delay. Only these bins are used to estimate the relative transfer function, so reliable estimates can be obtained under non-stationary noise conditions. The experimental results show that the proposed algorithm significantly improves the performance of the transfer-function generalized sidelobe canceller, while only sustaining a modest estimation error in adverse non-stationary noise environments.</P>

Spectral Subtraction Using Spectral Harmonics for Robust Speech Recognition in Car Environments

Beh, Jounghoon,Ko, Hanseok The Acoustical Society of Korea 2003 韓國音響學會誌 Vol.22 No.e2

This paper addresses a novel noise-compensation scheme to solve the mismatch problem between training and testing condition for the automatic speech recognition (ASR) system, specifically in car environment. The conventional spectral subtraction schemes rely on the signal-to-noise ratio (SNR) such that attenuation is imposed on that part of the spectrum that appears to have low SNR, and accentuation is made on that part of high SNR. However, these schemes are based on the postulation that the power spectrum of noise is in general at the lower level in magnitude than that of speech. Therefore, while such postulation is adequate for high SNR environment, it is grossly inadequate for low SNR scenarios such as that of car environment. This paper proposes an efficient spectral subtraction scheme focused specifically to low SNR noisy environment by extracting harmonics distinctively in speech spectrum. Representative experiments confirm the superior performance of the proposed method over conventional methods. The experiments are conducted using car noise-corrupted utterances of Aurora2 corpus.

Real-Time Continuous Phoneme Recognition System Using Class-Dependent Tied-Mixture HMM With HBT Structure for Speech-Driven Lip-Sync

Junho Park,Hanseok Ko IEEE 2008 IEEE transactions on multimedia Vol.10 No.7

<P>This work describes a real-time lip-sync method using which an avatar's lip shape is synchronized with the corresponding speech signal. Phoneme recognition is generally regarded as an important task in the operation of a real-time lip-sync system. In this work, the use of the Head-Body-Tail (HBT) model is proposed for the purpose of more efficiently recognizing phonemes which are variously uttered due to co-articulation effects. The HBT model effectively deals with the transition parts of context-dependent models for small-sized vocabulary tasks. These models provide better recognition performance than general context-dependent or context-independent models for the task of digit or vowel recognition. Moreover, each phoneme is categorized into one among four classes and the class-dependent codebook is generated to further improve the performance. Additionally, for the clear representation of the context dependency information in the transient parts, some Gaussians are excluded from class-dependent codebook. The proposed method leads to a lip-sync system that performs at a level that is similar to previous designs based on HBT and continuous hidden Markov models (CHMMs). However, our method reduces the number of model parameters by one-third and enables real-time operation.</P>