About

We conduct research in the field of Artificial Intelligence, focusing on deep learning methodologies from foundational studies to practical applications, with an emphasis on perception (recognition), reasoning, and generative models. We aim to develop simple yet effective approaches that are easily understandable without complex knowledge and readily implementable without complicated algorithms, enabling widespread adoption.

Youngsung Kim

Assistant Professor (Tenure-track)
  • Department: Artificial Intelligence (Undergraduate); Electrical and Computer Engineering (Graduate)
  • Institution: Inha University

Previous Positions

  • Principal Researcher at Samsung Advanced Institute of Technology (SAIT)
  • Postdoctoral Associate at SMART, MIT
  • Visting Researcher at MILA/SAIL & at I2R

Education

  • Ph.D., M.S., B.S. in Electrical and Electronic Engineering, Yonsei University

Research Team

Graduate Students
  • [Positions Available]
Undergraduate Interns
  • Hanbyul Kim - Reasoning Models
  • Sangmin Oh - Multimodal Models

Join Our Research Group

We are continually seeking highly motivated undergraduate and graduate students. Both full-time and part-time positions are available with compensation based on qualifications.

Priority Research Areas:
  • Multimodal Models
  • Visual Reasoning
  • Neurosymbolic AI
  • Deep Learning Applications
What We Offer:
  • Collaborative research environment
  • Conference participation support
  • Independent research opportunities
  • Industry collaboration projects

연구원 모집

학부 및 대학원 과정에서의 열정적이고 성실한 학생들을 항상 모집합니다. 정규(Full-time) 및 부분(Part-time) 형태 모두 가능합니다.

주요 연구 분야: Multimodal Models, Visual Reasoning, Neurosymbolic AI를 포함하며, 인공지능(AI), 머신러닝(ML) 분야의 다양한 주제를 다룰 수 있습니다.

Contact: y.kim@inha.ac.kr

Research

Research Interests

Deep Learning

Research focuses on neural network architectures, optimization techniques, and representation learning. We explore self-supervised learning and efficient inference methods for both small and large-scale models.

Machine Learning

Development of novel algorithms for supervised, unsupervised, and self-supervised learning. Focus on interpretability and efficiency of machine learning systems.

Computer Vision

Image and video understanding, object detection and segmentation, and visual scene analysis for real-world applications.

Multimodal Learning

Integration of multiple data modalities including vision, language, and others. Cross-modal retrieval, alignment, and generation with applications in human-device-AI interaction.

Neurosymbolic AI

Bridging neural networks with symbolic reasoning for interpretable and robust AI. Logic-guided learning, knowledge integration, and hybrid architecture design.

HW-SW Co-design

Research on hardware-software co-optimization for efficient AI systems. End-to-end design methodologies for accelerators, model compression, and edge deployment.

Publications

Selected Research Papers & Patents

Conference Papers

  • Protofl: Unsupervised federated learning via prototypical distillation
    H. Kim, Y. Kwak, M. Jung, J. Shin, Y. Kim, C. Kim
    International Conference on Computer Vision (ICCV 2023)
  • Connecting Sphere Manifolds Hierarchically for Regularization
    D. Scieur*, Y. Kim* (*Equal Contribution)
    International Conference on Machine Learning (ICML 2021)
  • Few-shot Visual Reasoning with Meta-Analogical Contrastive Learning
    Y. Kim, J. Shin, E. Yang, and S. J. Hwang
    Neural Information Processing Systems (NeurIPS 2020)
    ★ Samsung Best Paper Award 2021 (Silver)
  • Residual Encoder Decoder Network and Adaptive Prior for Face Parsing
    T. Guo, Y. Kim, H. Zhang, D. Qian, B. Yoo, J. Xu, D. Zou, J.-J. Han, and C. Choi
    AAAI Conference on Artificial Intelligence (AAAI 2018)
  • Activity Recognition for Smartphone Based Travel Surveys Based on Cross-User History Data
    Y. Kim, F. C. Pereira, F. Zhao, A. Ghorpade, P. C. Zegras, and M. Ben-Akiva
    International Conference on Pattern Recognition (ICPR 2014)
  • Evaluating FMS: A preliminary comparison with a traditional travel survey
    C. Carrion, F. Pereira, R. Ball, F. Zhao, Y. Kim, K. Nawarathne, N. Zheng, C. Zegras, and M. Ben-Akiva
    Transportation Research Board 93rd Annual Meeting (TRB 2014)
  • An Online Learning Algorithm for Biometric Scores Fusion
    Y. Kim, K.-A. Toh, and A. B. J. Teoh
    The 4th IEEE Conference on Biometrics: Theory, Applications and Systems (BTAS 2010)
  • A Method to Enhance Face Biometric Security
    Y. Kim and K.-A. Toh
    The First IEEE Conference on Biometrics: Theory, Applications and Systems (BTAS 2007)

Journal Papers

  • Deep Self-Supervised Diversity Promoting Learning on Hierarchical Hyperspheres for Regularization
    Y. Kim, Yoonsuk Hyun, Jae-Joon Han, Eunho Yang, Sung Ju Hwang, Jinwoo Shin
    IEEE Access, vol. 11, pp. 146208-146222, 2023
  • Deep stochastic logic gate networks
    Y. Kim
    IEEE Access, vol. 11, pp. 122488-122501, 2023
  • Activity Recognition for a Smartphone and Web Based Human Mobility Sensing System
    Y. Kim, A. Ghorpade, F. Zhao, F. C. Pereira, P. C. Zegras, and M. Ben-Akiva
    IEEE Intelligent Systems, vol. 33, no. 04, pp. 5-23, 2018
  • Deep Facial Age Estimation using Conditional Multitask Learning with Weak Label Expansion
    B. Yoo, Y. Kwak, Y. Kim, C. Choi, and J. Kim
    IEEE Signal Processing Letters, vol. 25, no. 6, pp. 808-812, 2018
  • Background Subtraction Using Illumination-Invariant Structural Complexity
    W. Kim and Y. Kim
    IEEE Signal Processing Letters, vol. 23, no. 5, pp. 634--638, May 2016.
  • Exploratory Analysis of a Smartphone-Based Travel Survey in Singapore
    F. Zhao, F. C. Pereira, R. Ball, Y. Kim, Y. Han, C. Zegras, and M. Ben-Akiva
    Transportation Research Record, vol. 2, no. 2494, pp. 45-56, 2015
    ★ Pyke Johnson Award 2015 (TRB)
  • An Online Learning Network for Biometric Scores Fusion
    Y. Kim, K.-A. Toh, A.B.J. Teoh, H.-L. Eng, and W.-Y. Yau
    Neurocomputing, vol. 102, pp. 65-77, February 2013.
  • An Online AUC Formulation for Binary Classification
    Y. Kim, K.-A. Toh, A.B.J. Teoh, H.-L. Eng, and W.-Y. Yau
    Pattern Recognition, vol. 45, no. 6, pp. 2266-2279, June 2012.
  • A Performance Driven Methodology for Cancelable Face Templates Generation
    Y. Kim, A.B.J. Teoh, and K.-A. Toh
    Pattern Recognition, vol. 43, no. 7, pp. 2544-2559, July 2010.
  • Fusion of visual and infra-red face scores by weighted power series
    K.-A. Toh, Y. Kim, S. Lee, and J. Kim
    Pattern Recognition Letters, vol. 29, no. 5, pp. 603-615, April 2008.

Pre-prints

  • Standard Neural Computation Alone Is Insufficient for Logical Intelligence
    Y. Kim
    arXiv preprint arXiv:2502.02135
  • Towards Narrowing the Generalization Gap in Deep Boolean Networks
    Y. Kim
    arXiv preprint arXiv:2409.05905
  • Deep hierarchical-hyperspherical learning
    Y. Kim and J.J. Han
    submitted to conference 2019
  • Deep generative-contrastive networks for facial expression recognition
    Y. Kim, B. Yoo, Y. Kwak, C. Choi, and J. Kim
    (arXiv preprint 2017, arXiv:1703.07140)

Intellectual Property

Patent Portfolio

15 Registered US Patents | 7 Pending Applications

Our patent portfolio covers biometric authentication, facial recognition, emotion detection, and neural network architectures. Notable deployment includes authentication technology embedded in Samsung Galaxy smartphones (S6+, Note 5, S7, S8, and more).

  • Biometric Authentication: US Patents 10,248,835; 10,691,918; 10,521,642
  • Facial Recognition & Expression: US Patents 10,891,468; 10,387,716; 11,093,734
  • Neural Network Methods: US Patent Apps 17/026,951; 16/902,299

Contact

Get in Touch

Email

y.kim@inha.ac.kr

Office

Inha-ro 100, Michuhol-gu
Incheon 22212, Korea

Phone

+82 32 860 9519

Academic Profiles

Copyright © 2025 Youngsung Kim | AI Research Group