YongKyung Oh

YongKyung Oh

Visiting researcher at the UCLA Health - David Geffen School of Medicine.

Research experience

Affiliation

  • Postdoctoral researcher at UCLA Health - David Geffen School of Medicine at UCLA
    • Medical & Imaging Informatics (MII) group
    • Principal Investigator: Dr. Alex Bui

Research Interests

  • Integration of domain expertise with artificial intelligence to drive innovative data analytics and modeling efforts.
  • Design of robust and reliable AI models for dynamic environments, with a focus on adapting to dataset shifts.
  • Development and application of advanced time series analysis techniques, particularly for irregularly sampled data.
  • Implementation of cutting-edge machine learning and deep learning techniques, including neural differential equations.
  • Expertise in longitudinal data analysis and pattern detection, especially in medical and healthcare informatics.
  • Proficient in utilizing statistical methods based on physical and engineering models across various domains.

Research Keywords

  • Physical & Engineering Model with simulation
  • Anomaly Detection using machine learning
  • Causality analysis with domain knowledge
  • Multivariate Time-series analysis
  • Graph Neural Network
  • Multi-modal data fusion
  • Neural Differential Equations

Related Working domains

  • Sequence data from industry
  • Maritime surveillance
  • Shipping Logistics & Stowage Planning
  • Urban traffic and accident congestion propagation
  • Gas sensor array and Electronic Nose
  • Nuclear reactor operation and Anomaly detection
  • Medical data including real-time sensor signal and image data

Working Papers

  • Oh, Y., Kim, S. & Bui, A., Deep Interaction Feature Fusion for Robust Human Activity Recognition
  • Oh, Y., Lim, D., Kim, S. & Bui, A., TANDEM: Temporal Attention-guided Neural Differential Equations for Missingness in Time Series Classification
  • Oh, Y., Kam, S., Lim, D. & Kim, S., Advancing Irregular Time Series Classification in Astronomy with Neural Stochastic Differential Equations
  • Oh, Y., Lim, D. & Kim, S., Neural Flow-Based Controlled Differential Equations for Robust Human Activity Recognition with Irregularly Sampled Time Series Data
  • Oh, Y., Lim, D. & Kim, S., DualDynamics: Synergizing Implicit and Explicit Methods for Robust Irregular Time Series Analysis
  • Oh, Y., Koh, G., Shin, K., Moon, J. & Kim, S. Real-Time Risk Assessment of Thyroid Function Abnormality using Irregularly-Sampled Heart Rate Records.
  • Oh, Y., Goh, K., Kwak, J., Shin, K., Kim, G., Lee M., Choung, H., Kim, N., Moon, J. & Kim, S., TAOD-Net: A Data-Driven Approach for Detecting Thyroid-Associated Orbitopathy Symptoms Using Facial Images. [code]
  • Oh, Y., Kim, H. & Kim, S., TSSI: Time Series as Screenshot Images for Multivariate Time Series Classification using Convolutional Neural Networks.
  • Oh, Y., Kwak, J. & Kim, S., A propagation prediction method for non-recurrent traffic congestion. [preprint]

Journal Papers (Published)

  • [J10] Oh, Y., Kim, S., & Bui, A.A.T. (2025). Deep Interaction Feature Fusion for Robust Human Activity Recognition. In: Peng, KC., et al. Human Activity Recognition and Anomaly Detection. IJCAI 2024. Communications in Computer and Information Science, vol 2201. Springer, Singapore. https://doi.org/10.1007/978-981-97-9003-6_7. [paper]
  • [J9] Oh, Y., Yoon, K., Park, J., & Kim, S. (2024). Comparative evaluation of VAE-based monitoring statistics for real-time anomaly detection in AIS data. Maritime Policy & Management, 1-18. https://doi.org/10.1080/03088839.2024.2388177. [paper]
  • [J8] Oh, Y., & Kim, S. (2024). Multi-modal lifelog data fusion for improved human activity recognition: A hybrid approach. Information Fusion, 110, 102464. https://doi.org/10.1016/j.inffus.2024.102464. [paper] [code]
  • [J7] Oh, Y., & Kim, S. (2023), Grid-Based Bayesian Bootstrap Approach for Real-Time Detection of Abnormal Vessel Behaviors From AIS Data in Maritime Logistics. IEEE Transactions on Automation Science and Engineering. https://doi.org/10.1109/TASE.2023.3329041. [paper] [code]
  • [J6] Oh, Y., Lee, J. & Kim, S. (2023), Sensor drift compensation for gas mixture classification in batch experiments. Quality and Reliability Engineering International. 2023; 1-16. https://doi.org/10.1002/qre.3354. [paper]
  • [J5] Oh, Y., Kwak, J. & Kim, S. (2023), Time delay estimation of traffic congestion propagation due to accidents based on statistical causality. Electronic Research Archive, 31(2), 691-707. https://doi.org/10.3934/era.2023034. [paper] [code]
  • [J4] Oh, Y., Lee, C., Lee, J., Kim, S. & Kim, S. (2022). Multichannel convolution neural network for gas mixture classification. Annals of Operations Research, 339(1), 261-295. https://doi.org/10.1007/s10479-022-04715-2. [paper] [code]
  • [J3] Lee, J., Kwak, J., Oh, Y. & Kim, S. (2022), Quantifying incident impacts and identifying influential features in urban traffic networks. Transportmetrica B: Transport Dynamics, 11(1), 279-300. https://doi.org/10.1080/21680566.2022.2063205. [paper]
  • [J2] Oh, Y., Kim, H., Lee, D. & Kim, S. (2021), Simulation-based Anomaly Detection in Nuclear Reactors. Journal of the Korean Institute of Industrial Engineers, 47(2), 130-143. https://doi.org/10.7232/JKIIE.2021.47.2.130. [paper]
  • [J1] Oh, Y., Kim, N. & Kim, S. (2021), Transfer Learning based Approach for Mixture Gas Classification. Journal of the Korean Institute of Industrial Engineers, 47(2), 144-159. https://doi.org/10.7232/JKIIE.2021.47.2.144. [paper]

Conference Papers (In-press or Published)

  • [C21] Oh, Y. & Kim, S. (2024). Grid-Based Bayesian Bootstrap Approach for Real-Time Detection of Abnormal Vessel Behaviors from AIS Data in Maritime Logistics, IEEE 20th International Conference on Automation Science and Engineering (CASE) 2024, Institute of Electrical and Electronics Engineers (IEEE), Aug 2024. [slide]
  • [C20] Oh, Y., Lim, D., Kim, S. & Bui, A. (2024). Attention-guided Neural Differential Equations Framework for Missingness in Time Series Classification, The 10th Mining and Learning from Time Series (MILETS) Workshop, held in conjunction with the 2024 ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD) 2024, Aug 2024. [workshop] [poster]
  • [C19] Oh, Y., Kim, S. & Bui, A. (2024). Deep Interaction Feature Fusion for Robust Human Activity Recognition, 4th International Workshop on Deep Learning for Human Activity Recognition, held in conjunction with the 33rd International Joint Conference on Artificial Intelligence (IJCAI) 2024, Aug 2024. [workshop] [proceeding] [slide]
  • [C18] Oh, Y., Lim, D., & Kim, S. (2024), Neural CDE-Flow for Irregular Time Series: Integrating Neural Controlled Differential Equations and Neural Flow for Comprehensive Irregularly-sampled Time Series Analysis, Institute of Industrial and Systems Engineers (IISE) Annual Conference & Expo 2024, May 2024. Data Analytics and Information Systems (DAIS) Division Best Paper Competition Finalist.[slide]
  • [C17] Oh, Y., Kam, S., Lim, D. & Kim, S. (2024), Neural Langevin-type Stochastic Differential Equations for Astronomical time series Classification under Irregular Observations, AI4DifferentialEquations in Science Workshop, The Twelfth International Conference on Learning Representations (ICLR) 2024, May 2024. [workshop] [poster]
  • [C16] Oh, Y., Lim, D., & Kim, S. (2024), Stable Neural Stochastic Differential Equations in Analyzing Irregular Time Series Data, The Twelfth International Conference on Learning Representations (ICLR) 2024, May 2024. Spotlight presentation (Notable Top 5%). [paper] [code] [slide] [poster] [news]
  • [C15] Oh, Y., Lim, D., & Kim, S. (2024), Invertible Solution of Neural Differential Equations for Analysis of Irregularly-Sampled Time Series. Artificial Intelligence for Time Series (AI4TS) Workshop, The 38th Annual AAAI Conference on Artificial Intelligence 2024, February 2024. [workshop] [poster]
  • [C14] Oh, Y., Kam, S., Lim, D., & Kim, S. (2024), Enhancing Astronomical time series Classification with Neural Stochastic Differential Equations under Irregular Observations, Knowledge Guided ML (KGML) Bridge Program, The 38th Annual AAAI Conference on Artificial Intelligence 2024, February 2024. [workshop] [slide] [poster]
  • [C13] Oh, Y., Lim, D., & Kim, S. (2023), Unifying Neural Controlled Differential Equations and Neural Flow for Irregular Time Series Classification. In The Symbiosis of Deep Learning and Differential Equations (DLDE) III Workshop, The 37th Annual Conference on Neural Information Processing Systems (NeurIPS) 2023, December 2023.[workshop] [paper] [poster]
  • [C12] Oh, Y. & Kim, S. (2023), Flow-based Neural Differential Equations for Time series Analysis, Conference of Korea Computer Congress 2023 (KCC 2023), June 2023. Artificial Intelligence Division - Best Paper Award. [slide]
  • [C11] Oh, Y. & Kim, S. (2023), Heterogeneous Model Fusion for Enhanced Sensor-Based Human Activity Recognition, Conference of Korea Computer Congress 2023 (KCC 2023), June 2023. Artificial Intelligence Division - Best Oral Presentation Award. [slide]
  • [C10] Oh, Y. & Kim, S. (2023), Neural Stochastic Differential Equations Based on the Ornstein-Uhlenbeck Process, Institute of Industrial and Systems Engineers (IISE) Annual Conference & Expo 2023, May 2023. Data Analytics and Information Systems (DAIS) Division Best Paper Competition Finalist.[slide]
  • [C9] Oh, Y. & Kim, S. (2022), Lifelog data fusion approach for emotion recognition, Conference of Korea Software Congress (KSC) 2022, December 2022. [slide]
  • [C8] Oh, Y., Koh, G., Shin, K., Moon, J. & Kim, S. (2022), Real-Time Risk Assessment of Thyroid Function Abnormality using Irregularly-Sampled Heart Rate Records, The 7th Joint Conference of Korean Artificial Intelligence Association (JKAIA) 2022, November 2022. [poster]
  • [C7] Oh, Y. & Kim, S. (2022), Irregularly sampled time series classification using neural stochastic differential equation, The 7th Joint Conference of Korean Artificial Intelligence Association (JKAIA) 2022, November 2022. [poster]
  • [C6] Oh, Y., Kwak, J. & Kim, S. (2022), Time Delay Estimation of Traffic Congestion Based on Statistical Causality, The 3rd Workshop on Data-driven Intelligent Transportation (DIT 2022), The 31st ACM International Conference on Information and Knowledge Management (CIKM) 2022, October 2022. [workshop] [slide]
  • [C5] Oh, Y. (2022), Multivariate Times Series Classification Using Multichannel CNN, Proceedings of the 31st International Joint Conference on Artificial Intelligence (IJCAI), Doctoral Consortium. Pages 5865-5866. IJCAI 2022, May 2022. [paper] [slide]
  • [C4] Oh, Y. & Kim, S. (2021), Multi-channel Convolution Neural Network for Gas Mixture Classification. In 2021 International Conference on Data Mining Workshops (ICDMW) (pp. 1094-1095). IEEE, December 2021. [paper] [slide]
  • [C3] Oh, Y. & Kim, S. (2021), Multi-channel Convolution Neural Network for Gas Mixture Classification, The 4th Joint Conference of Korean Artificial Intelligence Association (JKAIA) 2021, July 2021. [poster]
  • [C2] Oh, Y. & Kim, S. (2021), Logistics Anomaly Detection with Maritime Big Data: A Bootstrap Approach, Institute of Industrial and Systems Engineers (IISE) Annual Conference & Expo 2021, May 2021. Logistics & Supply Chain (LSC) Division - Best Paper Award. [slide]
  • [C1] Oh, Y. & Kim, S. (2019), Exploiting Logistics Anomaly Detection using Maritime Big Data. In IIE Annual Conference. Proceedings (pp. 964-969). Institute of Industrial and Systems Engineers (IISE). Institute of Industrial and Systems Engineers (IISE) Annual Conference & Expo, May 2019. [proceeding] [slide]

Presentations and Talks

[Archive link]

Industry Cooperation Projects

UCLA Depression Grand Challenge

  • Topic: AI-based solutions to advance depression research and treatment strategies
  • Participation period: 2023.09. ~ Present
    • The UCLA Depression Grand Challenge is a pioneering initiative that aims to cut the global burden of depression in half by 2050 through innovative research and treatment development.
    • Collaborate with Apple and UCLA Health in pioneering research using mobile technology and wearable devices to identify early indicators of depression and track treatment effectiveness
    • Develop and implement AI-based solutions for several studies, including OPTIMA (Operationalizing Digital PhenoTyping In the Measurement of Anhedonia), analyzing digital biomarkers to measure depression symptoms
  • Technology Stack: Python, Deep learning, Computer Vision

THYROSCOPE Inc.

  • Topic: Classification for Eye disease related to Thyroid illness
  • Participation period: 2021.09. ~ 2023.08.
    • Develop a classification model for eye disease related to thyroid illness, using pictures taken by a DSLR camera
    • Propose a visual symptom recognition model using a vision transformer (TAOD-Net)
  • Technology Stack: Python, Deep learning, Computer Vision

THYROSCOPE Inc.

  • Topic: Thyroid hormone prediction with wearable device
  • Participation period: 2021.09. ~ 2023.08.
    • Develop prediction model for thyroid hormone using data collected from wearable device
    • Sequence data modeling with different record frequencies and different modalities
  • Technology Stack: Python, Sequential machine learning & deep learning

NAVER maps

  • Topic: Traffic prediction and congestion modeling
  • Participation period: 2020.09. ~ 2022.11.
    • Develop a traffic prediction model using historic & real-time data
    • Algorithm for predicting traffic congestion from quantitative and qualitative perspectives
    • • Causal relationship analysis using traffic incident-oriented speed data
  • Technology Stack: Python, SQL, Spark, Graph based machine learning & deep learning

CyberLogitec

  • Topic: Development of optimization algorithm and solver for large-size problems
  • Participation period: 2019.06. ~ 2019.12.
    • Implement optimization preprocess and solve the algorithm
    • Develop a pipeline for large-size optimization problems using open source solvers including CBC
  • Technology Stack: Python, R, CPLEX, CBC

Ulsan Port Authority

  • Topic: Development of smart port and logistics system
  • Participation period: 2019.01. ~ 2019.12.
    • Analyze automatic identification system (AIS) data for maritime surveillance
  • Technology Stack: Python, R

Taesung Environmental Research Institute

  • Topic: Development of predictive models for sensor data analysis
  • Participation period: 2018.07. ~ 2019.03.
    • Construct research setup and develop pipeline including database setup to deployment
      • Phase 1. Develop linear predictive model (e.g. Generalized Linear Model)
      • Phase 2. Develop non-linear predictive model (e.g. SVM, RF, XGBOOST)
      • Phase 3. Develop multi-layer model
  • Technology Stack: Python, R, MySQL, H2O AI library

Patents

  • Apparatus and method for human activity recognition through hybrid fusion of dynamic and static data. | Co-inventor, (KR 10-2024-0033382, applied March 8, 2024)
  • Encoding apparatus and method for converting time series data into images | Co-inventor (KR 10-2024-0101063, granted July 2, 2024) (KR 10-2022-0183341, applied December 23, 2022).
  • Method, computer device, and computer program to predict road congestion propagation using pattern matching | Co-inventor (KR 10-2608343-0000, granted November 27, 2023) (JP2023039925A, published March, 20, 2023) (JP 2022-137508, applied August 31, 2022) (KR 10-2021-0120229, applied September 9, 2021).
  • Method, computer device, and computer program to predict propagation time delay lag of road congestion using transfer entropy | Co-inventor (KR 10-2604575-0000, granted November 16, 2023) (JP2023039418A, published March, 20, 2023) (JP 2022-137509, applied August 31, 2022) (KR 10-2021-0119772, applied September 8, 2021).
  • Method of Anomaly detection of vessels applying bayesian bootstrap | Co-inventor (KR 10-2534357-0000, granted May 16, 2023) (KR 10-2020-0178651, applied December 18, 2020).
  • Sensor drift compensation for mixed gas classification in E-nose system | Co-inventor (US 18/012,626, applied December 22, 2022) (KR 10-2364019-0000, granted February 14, 2022) (PCT/KR2021/010729, applied August 12, 2021) (KR 10-2020-0101037, applied August 12, 2020).
  • Method and apparatus for determining delay possibility of shipment | Co-inventor (KR 10-2250354-0000, granted May 4, 2021) (PCT/KR2020/015854, applied November 12, 2020) (KR 10-2019-0158913, applied December 3, 2019).

Research Awards and Honors

  • IISE 2024 DAIS Division Best Paper Competition Finalist
    Oh, Y., Lim, D. & Kim, S. (2024), Neural CDE-Flow for Irregular Time Series: Integrating Neural Controlled Differential Equations and Neural Flow for Comprehensive Irregularly-sampled Time Series Analysis, Data Analytics and Information Systems (DAIS) Division, Institute of Industrial and Systems Engineers (IISE) Annual Conference & Expo 2024, May 2024. [slide]
  • KCC 2023 AI Division Best Research Paper Award
    Oh, Y. & Kim, S. (2023), Flow-based Neural Differential Equations for Time series Analysis, Artificial Intelligence Division, Korea Computer Congress (KCC) 2023, Korean Institute of Information Scientists and Engineers (KIISE), June 2023.
  • KCC 2023 AI Division Best Oral Presentation Award
    Oh, Y. & Kim, S. (2023), Heterogeneous Model Fusion for Enhanced Sensor-Based Human Activity Recognition, Artificial Intelligence Division, Korea Computer Congress (KCC) 2023, Korean Institute of Information Scientists and Engineers (KIISE), June 2023.
  • IISE 2023 DAIS Division Best Paper Competition Finalist
    Oh, Y. & Kim, S. (2023), Neural Stochastic Differential Equations Based on the Ornstein-Uhlenbeck Process, Data Analytics and Information Systems (DAIS) Division, Institute of Industrial and Systems Engineers (IISE) Annual Conference & Expo 2023, May 2023. [slide]
  • IISE 2021 LSC Division Best Paper Award
    Oh, Y. & Kim, S. (2021), Logistics Anomaly Detection with Maritime Big Data: A Bootstrap Approach, Logistics & Supply Chain (LSC) Division, Institute of Industrial and Systems Engineers (IISE) Annual Conference & Expo 2021, May 2021. [slide] [news]
  • Distinguished Paper Award
    Oh, Y., Lim, D., Hong, D., Lee, J. & Park, J. (2019), Strategy for Maritime Startup Ecosystem, Graduate Student Academic paper competition by Ulsan Development Institute, December 2019.
  • Distinguished Paper Award
    Oh, Y., Lim, D., Hong, D., Lee, J. & Park, J. (2019), Suggestion for Smart Port in industry 4.0, International Conference on Maritime - Academic paper competition, Dongseo University and Ministry of Maritime Affairs and Fisheries, November 2019.

Dissertations

  • Oh, Y. (2023), A Comprehensive Study of Deep Learning for Real-World Multivariate Time Series Classification, Doctoral dissertation paper advised by Prof. Kim, S.
  • Oh, Y. (2017), Discussion of the meaning and components of the fourth industrial revolution, and recommendations for Korea to prepare for the fourth industrial revolution, Master dissertation paper advised by Prof. Gang, K.