Seong Hyeon Park

Seong Hyeon Park

Ph.D. Student

KAIST AI

Welcome

I focus on multi-modal representation learning. Most of my recent works are related to few-/zero-shot adaptation of vision-language models for novel tasks and data, including image segmentation and video recognition. Currently, I am a Ph.D. student member of Algorithmic Intelligence Laboratory @ the KAIST AI, advised by Dr. Jinwoo Shin. Previously, I graduated from Hanyang University with M.S. and B.S degrees under Dr. Jun-Won Choi, where I studied machine learning models for motion forecasting and autonomous driving.

Interests
  • Deep Learning
  • Multimodal Representation
  • Predictive Uncertainty
Education
  • M.S. in Electrical Engineering, 2020

    Hanyang University

  • B.S. in Electrical Engineering, 2018

    Hanyang University

Recent Publications

(2023). IFSeg: Image-free Semantic Segmentation via Vision-Language Model. In CVPR 2023.

Cite Code arXiv

(2022). K-centered Patch Sampling for Efficient Video Recognition. In ECCV 2022.

Cite Code

(2021). LaPred: Lane-Aware Prediction of Multi-Modal Future Trajectories of Dynamic Agents. In CVPR 2021.

Cite arXiv

(2020). Diverse and Admissible Trajectory Forecasting through Multimodal Context Understanding. In ECCV 2020.

Cite Code Slides Video DOI

(2018). Sequence-to-Sequence Prediction of Vehicle Trajectory via LSTM Encoder-Decoder Architecture. In IEEE IV'18.

Cite Video DOI