전임교수 겸임교수 fnctId=prof,fnctNo=484 성진택(Jin-Taek Seong) 직위(직급) 교수 연구실 데이터사이언스대학원 306호 교수전공 및 연구분야 Information Theory / Artificial Intelligence / Bigdata Analysis 연락처 062-530-5798 이메일 jtseong@jnu.ac.kr 홈페이지 https://sites.google.com/view/jin-taek/ Education (1) Ph.D.(Aug. 2014), Gwangju Institute of Science and Technology (GIST), School of Information and Communication Engineering, Ph.D. Dissertation “Inverse Problems of Compressed Sensing and Cooperative Networks over Finite Fields” advised by Prof. Heung-No Lee (2) M.S.(Feb. 2008), Gwangju Institute of Science and Technology (GIST), School of Information and Mechatronics Engineering, M.S. Thesis “Design and Performance Analysis of Wideband Analog Correlator for Fully Polarimetric Radiometer at W-band” advised by Prof. Yong-Hoon Kim (3) B.S.(Feb. 2006), University of Seoul (UOS), School of Electrical Engineering and Computer Science Work/Research Experience (1) Associate Professor (Mar. 2018 – Feb. 2023), Department of Convergence Software, Mokpo National University (2) Assistant Professor (Apr. 2017 – Feb. 2018), Department of Information and Communication Engineering, Honam University (3) Assistant Deputy Director (Sep. 2016 – Mar. 2017), Project Management Division, Defense Acquisition Program Administration (DAPA, 방위사업청) (4) Researcher (Sep. 2014 – Sep. 2016), Medical Device Development Center, Daegu-Gyeongbuk Medical Innovation Foundation (DGMIF, 대구경북첨단의료산업진흥재단) (5) Junior Researcher (Mar. 2008 – Jan. 2010), Air Conditioning Division, LG Electronics Inc. (6) IEEE(member), 대한전자공학회(정회원), 한국정보과학회(정회원), 한국인공지능학회(정회원), 소프트웨어과업심의위원(Oct. 2022 – current), 전남건설기술심의위원(Dec. 2021, - current), 전북건설기술심의위원(Dec. 2023, - current), 익산지방국토관리청 기술자문위원(Feb. 2023 – current), 서울교통공사 기술자문위원(Jan. 2024 – current) Research Areas Anomaly Detection In data analysis, anomaly detection (also referred to as outlier detection and sometimes as novelty detection) is generally understood to be the identification of rare items, events or observations which deviate significantly from the majority of the data and do not conform to a well defined notion of normal behaviour. Such examples may arouse suspicions of being generated by a different mechanism, or appear inconsistent with the remainder of that set of data. Detection of Electricity Theft based Learning Technieques Electricity theft refers to the unauthorized and illegal consumption of electricity without proper billing or payment. This illicit activity poses significant challenges for power utilities, leading to revenue losses and increased operational costs. It involves various techniques, such as tampering with meters, illegal connections, and meter bypassing, which not only undermines the financial stability of utility companies but also disrupts the overall power distribution system. Combating electricity theft requires implementing robust monitoring systems, raising public awareness, and enforcing stringent penalties to deter such practices Label Noise in Supervised LearningLabel noise in supervised learning refers to the presence of incorrect or inaccurate labels in the training data. In this paradigm, the algorithm learns from data where each input is associated with a corresponding target label. However, real-world datasets can be prone to errors, mislabeling, or ambiguity, which introduces label noise. Addressing label noise is crucial as it can severely impact the model's performance and generalization ability, necessitating the use of various data cleaning, regularization, and ensemble techniques to mitigate its adverse effects and improve the reliability of the learned model. Group Testing Group testing is a technique where multiple samples are combined into pools and tested together to identify positive cases efficiently. By testing multiple samples at once, group testing reduces the number of tests required and saves resources. Group testing is used in various fields such as healthcare, public health, and agriculture for screening, surveillance, and control of diseases. Group testing can detect positive cases with high sensitivity and specificity while reducing the burden on the healthcare system. Group testing requires a certain level of prevalence, and additional confirmatory tests may be needed for positive pools. COVID-19 testing of asymptomatic individuals in a community can be done using group testing to identify positive cases efficiently. Latent Factor Analysis Latent factor analysis is a statistical technique used to identify the underlying structure of a set of variables. It assumes that there are underlying factors that influence the observed variables and seeks to identify those factors. It is commonly used in fields such as psychology, education, marketing, and finance. The technique involves several steps, including data collection, data preprocessing, factor extraction, and factor rotation. The results of the analysis can provide insights into the relationships between the observed variables and the underlying factors. It can also be used for predictive modeling and dimensionality reduction. Compressive Sensing Compressive sensing is a signal processing technique used to efficiently acquire and process high-dimensional data with fewer measurements than traditional methods. It assumes that the signal to be acquired is sparse, meaning it has few non-zero coefficients in a known or unknown basis. It allows for data to be acquired at a lower rate and then reconstructed with high accuracy, reducing data acquisition and storage costs. The technique involves several steps, including signal acquisition, signal processing, and sparse signal recovery. Compressive sensing has applications in a wide range of fields, including image and video processing, wireless communication, medical imaging, and more. The technique is still an active area of research, with ongoing efforts to improve its efficiency and accuracy.