6 papers
Equivariant Latent Alignment via Flow Matching under Group Symmetries
Sunghyun Kim, Jaehoon Hahm, Jeongwoo Shin +1
Geometry-aware generative models and novel view synthesis approaches have shown strong potential in visual fidelity and consistency. In parallel, equivariant representation learnin…
Efficient Adjoint Matching for Fine-tuning Diffusion Models
Jeongwoo Shin, Dongsoo Shin, Yuchen Zhu +5
Reward fine-tuning has become a common approach for aligning pretrained diffusion and flow models with human preferences in text-to-image generation. Among reward-gradient-based me…
Efficient Generative Modeling beyond Memoryless Diffusion via Adjoint Schrödinger Bridge Matching
Jeongwoo Shin, Jinhwan Sul, Joonseok Lee +2
Diffusion models often yield highly curved trajectories and noisy score targets due to an uninformative, memoryless forward process that induces independent data-noise coupling. We…
Latent Diffusion Models with Masked AutoEncoders
Junho Lee, Jeongwoo Shin, Hyungwook Choi +1
In spite of the remarkable potential of Latent Diffusion Models (LDMs) in image generation, the desired properties and optimal design of the autoencoders have been underexplored. I…
Scalable Frame Sampling for Video Classification: A Semi-Optimal Policy Approach with Reduced Search Space
Junho Lee, Jeongwoo Shin, Seung Woo Ko +2
Given a video with frames, frame sampling is a task to select frames, so as to maximize the performance of a fixed video classifier. Not just brute-force search, but…
Self-Guided Masked Autoencoder
Jeongwoo Shin, Inseo Lee, Junho Lee +1
Masked Autoencoder (MAE) is a self-supervised approach for representation learning, widely applicable to a variety of downstream tasks in computer vision. In spite of its success,…