Hank Chi-Hsi Kung 孔啟熙

I am a visiting researcher at Indiana University Bloomington where I am working with Prof. David Crandall and Prof. Linda Smith. My current research focuses on visual representation learning in both human and machine intelligence.

Previously, I was a research assistant at National Chiao Tung University, Taiwan, and was supervised by Prof. Yi-Ting Chen and Dr. Yi-Hsuan Tsai.

I was a research intern at IBM Thomas J. Watson Research Center. I received my M.Sc from National Tsing-Hua University, where I was supervised by Prof. Che-Rung Lee, and B.Sc from National Taipei University.

I am actively seeking a Ph.D. position!

Email / Google Scholar / X / Github / CV

Research Interests

How can we train AI so that fast, intuitive perception and slow, explicit reasoning learn from one another? In humans, Kahneman’s System 1 and System 2 are intertwined: intuition reflects internalized reasoning. My goal is developmental co-evolution: slow reasoning shapes what fast perception learns, and improving perceptual representations in turn reshape reasoning strategies.

My approach is inspired by cognitive development: infants first acquire intuitive representations through sensorimotor exploration, then gradually ground symbolic concepts in those continuous representations. Across development, perception and reasoning reshape each other continually. Based on this progression, I pursue three directions:

(1) Cognitive-inspired world models for intuitive physics that capture physical dynamics, emphasizing object awareness and memory, and evaluated with reasoning-aware benchmarks that test whether internal states support simple inferences and actions;
(2) A learnable neuro-symbolic interface that maps continuous latents into discrete predicates (e.g., stability, friction, collision) with natural supervision, instead of assuming perfect compositional symbols; and
(3) continual co-training where planner failures signal perception to refine its abstractions, and new perceptual affordances expand the planner’s search space over time.

Workshop Organizing

Organized Reverse-Engineering Human Learning Mechanisms: Toward Human-Like AI at NeurIPS 2025

3rd ROAD++ Workshop & Challenge of Compositional Representation for Traffic Activities at ECCV 2024

Publications
	Why Children Prefer Planar Views: A Reinforcement Learning Approach to Simulating Development of View Bias Chi-Hsi Kung, Frangil Ramirez, Juhyung Ha, Yi-Ting Chen, David Crandall, Linda Smith Paper coming soon / Code coming soon
	What Changed and What Could Have Changed? State-Change Counterfactuals for Procedure-Aware Video Representation Learning Chi-Hsi Kung, Frangil Ramirez, Juhyung Ha, Yi-Ting Chen, David Crandall, Yi-Hsuan Tsai, (* Equal Contribution) ICCV 2025, CVPR EgoVIS 2025 ICCV / arxiv / Code
	ATARS: An Aerial Traffic Atomic Activity Recognition and Temporal Segmentation Dataset Zihao Chen, Hsuanyu Wu, Chi-Hsi Kung, Yi-Ting Chen, Yan-Tsung Peng* (* Equal Advising) IROS 2025 arxiv / Code /
	Controllable Scenario-based Collision Generation for Safety Assessment Pin-Lun Chen, Chi-Hsi Kung, Che-Han Chang, Wei-Chen Chiu, Yi-Ting Chen Under review arxiv / Code coming soon
	Action-slot: Visual Action-centric Representations for Atomic Activity Recognition in Traffic Scenes Chi-Hsi Kung, Shu-Wei Lu, Yi-Hsuan Tsai, Yi-Ting Chen CVPR, 2024 project page / CVPR / arxiv / code / TACO dataset
	RiskBench: A Scenario-based Benchmark for Risk Identification Chi-Hsi Kung, Chieh-Chi Yang, Pang-Yuan Pao, Shu-Wei Lu, Pin-Lun Chen, Hsin-Cheng Lu, Yi-Ting Chen ICRA, 2024 project page / video / ICRA / code / dataset
	ADD: A Fine-grained Dynamic Inference Architecture for Semantic Image Segmentation Chi-Hsi Kung and Che-Rung Lee IROS, 2021 & ACML 2021 MRVC workshop IROS / code

Conference Reviewer

IEEE Conference on Computer Vision and Pattern Recognition CVPR 2023-2025

The International Conference on Machine Learning ICML 2025

International Conference on Computer Vision ICCV 2025

Advances in Neural Information Processing Systems NeurIPS 2024

IEEE International Conference on Development and Learning ICDL 2024

IEEE/RSJ International Conference on Intelligent Robots and Systems IROS 2025

Invited Talk

CVGIP 2024

Feel free to steal this website's source code. Do not scrape the HTML from this page itself, as it includes analytics tags that you do not want on your own website — use the github code instead. Also, consider using Leonid Keselman's Jekyll fork of this page.

Research Interests

Workshop Organizing

Publications

Conference Reviewer

Invited Talk