Junoh Kang

I am a third-year Ph.D. student at the SNU Computer Vision Lab, advised by Bohyung Han. I aim to solve problems of machine learning in a theoretically concrete manner by leveraging my expertise: mathematics applied to problem description and solution. My primary research interest is generative models. Specifically, I am interested in 1) unlearning generative models for privacy, 2) imposing more controllability for video generation, and 3) accelerating video generative models.

Besides the research, I am also interested in economics, investment, and playing sports such as tennis, golf, futsal, and boxing.

news

Feb 27, 2024	Our paper, Observation-Guided Diffusion Probabilistic Models, is accepted to CVPR 2024.
Oct 9, 2023	Our paper, Observation-Guided Diffusion Probabilistic Models, is archived.
Jun 22, 2023	Our paper, Open-Set Representation Learning Through Combinatorial Embedding, is presented at CVPR 2023.

publications

(*) denotes equal contribution

2024

CVPR

Observation-Guided Diffusion Probabilistic Models

Junoh Kang*, Jinyoung Choi*, Sungik Choi, and Bohyung Han

In CVPR, 2024

Abstract arXiv PDF Code

We propose a novel diffusion-based image generation method called the observation-guided diffusion probabilistic model (OGDM), which effectively addresses the trade-off between quality control and fast sampling. Our approach reestablishes the training objective by integrating the guidance of the observation process with the Markov chain in a principled way. This is achieved by introducing an additional loss term derived from the observation based on a conditional discriminator on noise level, which employs a Bernoulli distribution indicating whether its input lies on the (noisy) real manifold or not. This strategy allows us to optimize the more accurate negative log-likelihood induced in the inference stage especially when the number of function evaluations is limited. The proposed training scheme is also advantageous even when incorporated only into the fine-tuning process, and it is compatible with various fast inference strategies since our method yields better denoising networks using the exactly the same inference procedure without incurring extra computational cost. We demonstrate the effectiveness of our training algorithm using diverse inference techniques on strong diffusion model baselines. Our implementation is available at https://github.com/Junoh-Kang/OGDM_edm.
arXiv

FIFO-Diffusion: Generating Infinite Videos from Text without Training

Jihwan Kim*, Junoh Kang*, Jinyoung Choi, and Bohyung Han

In arXiv, 2024

Website

2023

CVPR

Open-Set Representation Learning Through Combinatorial Embedding

Geeho Kim, Junoh Kang, and Bohyung Han

In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023

Abstract arXiv PDF

Visual recognition tasks are often limited to dealing with a small subset of classes simply because the labels for the remaining classes are unavailable. We are interested in identifying novel concepts in a dataset through representation learning based on both labeled and unlabeled examples, and extending the horizon of recognition to both known and novel classes. To address this challenging task, we propose a combinatorial learning approach, which naturally clusters the examples in unseen classes using the compositional knowledge given by multiple supervised meta-classifiers on heterogeneous label spaces. The representations given by the combinatorial embedding are made more robust by unsupervised pairwise relation learning. The proposed algorithm discovers novel concepts via a joint optimization for enhancing the discrimitiveness of unseen classes as well as learning the representations of known classes generalizable to novel ones. Our extensive experiments demonstrate remarkable performance gains by the proposed approach on public datasets for image retrieval and image categorization with novel class discovery.