Associate Professor, Korea Advanced Institute of Science and Technology
5 papers at NeurIPS 2025
We propose a new algorithm for deep generative modeling of sequence data in continuous spaces based on a novel adaptation of operator theory for probabilistic dynamical systems.
We propose a modular inductive bias for disentangled representation learning, which we term a compositional bias, decoupled from both learning objectives and model architectures.
Few-shot spatial control for Text-to-Image Diffusion models by leveraging the analogy between query and support spatial conditions to construct task-specific control features.
Rectified Flow for Discrete Flow-based Models
We propose language-grounded visual concept learning in real-world scenes
