Junsung Park

I am a Ph.D. candidate at KAIST AI, advised by Prof. Hyunjung Shim. I received my B.S. in Mechanical Engineering & Artificial Intelligence from Korea University.

My research focuses on Physical AI & LLMs. I have been working on autonomous driving, vision-language models (VLMs), and LiDAR perception. I am interested in building AI systems that can understand and interact with the physical world. Recently, I’m interested in unifying various tasks and embodiments on a single VLA/WAM model.

News

Publications

* equal contribution

Grounding Driving VLA via Inverse Kinematics

Junsung Park, Hyunjung Shim

Conference on Neural Information Processing Systems (NeurIPS), 2026

PDF Code

WaymoQA: A Multi-View Visual Question Answering Dataset for Safety-Critical Reasoning in Autonomous Driving

Seungjun Yu, Seonho Lee, Namho Kim, Jaeyo Shin, Junsung Park, Wonjeong Ryu, Raehyuk Jung, Hyunjung Shim

Findings of the Association for Computational Linguistics: EMNLP, 2026

PDF Code

SUPER-AD: Semantic Uncertainty-aware Planning for End-to-End Robust Autonomous Driving

Wonjeong Ryu, Seungjun Yu, Seokha Moon, Hojun Choi, Junsung Park, Jinkyu Kim, Hyunjung Shim

Preprint, 2025

PDF

Scene, Class, Signal: Tri-Level Adaptation for Synthetic-to-Real LiDAR Segmentation

Junsung Park, Hyunjung Shim

Preprint, 2025

PDF

Robust Driving QA through Metadata-Grounded Context and Task-Specific Prompts

Seungjun Yu*, Junsung Park*, Youngsun Lim, Hyunjung Shim

IROS 2025 Workshop, RoboSense Challenge, 2025

PDF

3D-Aware Vision-Language Models Fine-Tuning with Geometric Distillation

Seonho Lee*, Jiho Choi*, Inha Kang, Jiwook Kim, Junsung Park, Hyunjung Shim

Findings of the Association for Computational Linguistics: EMNLP, 2025

PDF arXiv Code

No Thing, Nothing: Highlighting Safety-Critical Classes for Robust LiDAR Semantic Segmentation in Adverse Weather

Junsung Park, Hwijeong Lee, Inha Kang, Hyunjung Shim

IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2025

PDF Code

Rethinking Data Augmentation for Robust LiDAR Semantic Segmentation in Adverse Weather

Junsung Park, Kyungmin Kim, Hyunjung Shim

European Conference on Computer Vision (ECCV), 2024 Oral

PDF Code

Precision Matters: Precision-aware Ensemble for Weakly Supervised Semantic Segmentation

Junsung Park, Hyunjung Shim

AAAI Workshop, 2024

PDF Code

Education & Experience

• May 2026 – Nov 2026
  Research Intern, Qualcomm AI Research, Qualcomm Korea YH
  Investigate reasoning faithfulness in driving vision-language-action (VLA) models.
• 2023 – Present
  Ph.D. Candidate in Artificial Intelligence, Korea Advanced Institute of Science and Technology (KAIST)
• 2017 – 2023
  B.S. in Mechanical Engineering & Artificial Intelligence, Korea University

Talks

• Apr 2025
  Invited Talk, AI Tutorial Series, Seoul AI Hub
  Rethinking Data Augmentation for Robust LiDAR Semantic Segmentation in Adverse Weather
• Sep 2024
  Oral Presentation, ECCV 2024
  Rethinking Data Augmentation for Robust LiDAR Semantic Segmentation in Adverse Weather

Patents

• 2024
  Barbell with Built-in Weight Fixing Device
  Korea Patent, Reg. No. 1026314710000 (App. No. 10-2021-0092158)

Projects

• 2025
  IROS 2025 RoboSense Challenge, Track 1: Driving with Language [PDF]
  Developed a prompting method that helps VLMs understand driving scenes. Placed 4th.
• 2022
  Open Manipulator Grasping Task: Vision for Robot Arm Control [Code]
  Undergraduate project. Vision modules for robot-arm grasping; 3D grasp position from FCN segmentation and a pinhole camera model.
• 2021
  Capstone Design: Motor-Driven Screw Mechanism [Code] [Video]
  Undergraduate capstone project. Designed a helical screw mechanism in CAD and implemented Arduino PWM motor control.
• 2021
  Mobile Robot Project: ROS Programming & Motor Control [Code]
  Undergraduate project. ROS1 programming and embedded motor control on an OpenCM board.
• 2018
  Capstone Design: Stair Treads Wheels [PDF]
  Undergraduate project. Designed stair-climbing wheels and their rotation control. Bronze Prize.

Awards & Honors

• 2022 – 2023
  Scholarship, Lee Myungbak & Kim Yoonok Foundation
• 2021 – 2023
  Scholarship, Korea University Alumni Association
• 2018
  Bronze Prize & Social Contribution Prize, Capstone Design Project, Korea Univ. Innovation Center for Engineering

Teaching

• Spring 2026
  AI 621: Computational Photography, Teaching Assistant, KAIST
• Fall 2025
  AI 504: Programming for AI, Teaching Assistant, KAIST
• Spring 2025
  AI 621: Computational Photography, Teaching Assistant, KAIST
• Fall 2024
  AI 503: Math for AI, Teaching Assistant, KAIST
• Spring 2024
  AI 611: Deep Reinforcement Learning, Teaching Assistant, KAIST

Academic Services

Conference Reviewer

• NeurIPS 2026 (Top Reviewer, Main Track)
• ICLR 2026
• AAAI 2025

Lab Service

• Student Lab Leader, Prof. Hyunjung Shim's Lab, KAIST (Jul 2024 – May 2025)
• Research Server Administrator (sole administrator), Prof. Hyunjung Shim's Lab, KAIST (Aug 2023 – May 2025)