Seongmin Hong

I am a Postdoctoral Researcher at the Institute of New Media and Communications (INMC) at Seoul National University (SNU), currently completing alternative military service.

I earned my Ph.D. from the Department of Electrical and Computer Engineering at Seoul National University, advised by Prof. Se Young Chun (ICL). I view generative AI not as a one-shot sampler, but as a differentiable optimization process, and my research focuses on endowing modern generative architectures with invertibility and differentiability. This perspective enables physically grounded inference in imaging and integrity-by-design generation for public-facing media. My doctoral thesis was awarded the Distinguished Ph.D. Dissertation Award.

Seongmin Hong

News

Selected Publications

* denotes equal contribution.

2026

DiffBMP thumbnail

DiffBMP: Differentiable Rendering with Bitmap Primitives

Seongmin Hong*, Junghun James Kim*, Daehyeop Kim, Insoo Chung, Se Young Chun

Conference on Computer Vision and Pattern Recognition (CVPR), 2026

CVPR Co-1st Author Differentiable Rendering
Paper Project BibTeX

2024

Gradient-free Decoder Inversion thumbnail

Gradient-free Decoder Inversion in Latent Diffusion Models

Seongmin Hong, Suh Yoon Jeon, Kyeonghyun Lee, Ernest K. Ryu, Se Young Chun

Neural Information Processing Systems (NeurIPS), 2024

NeurIPS 1st Author Generative Model Optimization
Paper arXiv Project Video GitHub BibTeX
Adaptive Selection thumbnail

Adaptive Selection of Sampling-Reconstruction in Fourier Compressed Sensing

Seongmin Hong, Jaehyeok Bae, Jongho Lee, Se Young Chun

European Conference on Computer Vision (ECCV), 2024

ECCV 1st Author Computational Imaging Generative Model
Paper arXiv Project GitHub BibTeX
On Exact Inversion of DPM-Solvers thumbnail

On Exact Inversion of DPM-Solvers

Seongmin Hong, Kyeonghyun Lee, Suh Yoon Jeon, Hyewon Bae, Se Young Chun

Conference on Computer Vision and Pattern Recognition (CVPR), 2024

CVPR 1st Author Generative Model Optimization
Paper Supp arXiv Project GitHub BibTeX

2023

Normalizing Flows robustness thumbnail

On the Robustness of Normalizing Flows for Inverse Problems in Imaging

Seongmin Hong, Inbum Park, Se Young Chun

International Conference on Computer Vision (ICCV), 2023

ICCV 1st Author Computational Imaging Generative Model
Paper Supp arXiv Project GitHub BibTeX
Neural Diffeomorphic Non-uniform B-spline Flows thumbnail

Neural Diffeomorphic Non-uniform B-spline Flows

Seongmin Hong, Se Young Chun

Thirty-Seventh AAAI Conference on Artificial Intelligence (AAAI), 2023  (oral)

AAAI 1st Author Generative Model
Paper arXiv Project GitHub BibTeX

Exhibitions & Art Projects

Digital Assemblage 1 Digital Assemblage 2 Digital Assemblage 3

Engineer's AI Media Art: Digital Assemblage

Seoul Science Museum  ·  February 2026

Led a special public workshop at Seoul Science Museum demonstrating DiffBMP technology to create digital assemblage artworks. Taught participants how to transform everyday objects into artistic expressions using simple recognition methods, exploring the intersection of engineering, technology, and art.

IPA exhibition 1 IPA exhibition 2

IPA (Interactive Perspective Art): Ciao, Technology

SNU Powerplant  ·  January 2026

Collaborative kinetic art exhibition at SNU Powerplant. Served as lead for software infrastructure and typography for the signature piece Ordered Volume—a robotic sculpture utilizing 160 motors and spherical voxels to translate digital data into physical forms.

ARRCO

ARRCO: Multi-view 3D Crystal Point-Cloud Art

Independent crowd-funded project  ·  2025

Developed the full computational pipeline for high-density point-cloud laser engraving to create multi-view optical illusions where distinct images emerge from different observer perspectives.

Other Publications