Hessian Geometry of Latent Space in Generative Models
This paper presents a novel method for analyzing the latent space geometry of generative models, including statistical physics models and diffusion models, by reconstructing...
Style Transfer Lab
We build mathematically rigorous generative AI methods grounded in optimal transport, information geometry, and latent space analysis—with applications in color/style transfer, biological data imputation, and lightweight on-device systems.
About our lab
Our work bridges theory and deployment: we develop principled models and algorithms, then validate them on real tasks and publish with clear artifacts (papers, code, and demos where possible).
-
Theory
Optimal transport, geometry of probability, convergence guarantees. -
Methods
Diffusion/flows, guidance, latent space structure, training-free extensions. -
Applications
Color/style transfer, harmonization, biological data imputation, on-device inference.
Recent publications
View all
Color Transfer with Modulated Flows
In this work, we introduce Modulated Flows (ModFlows), a novel approach for color transfer between images based on rectified flows. The primary goal of...
Color Conditional Generation with Sliced Wasserstein Guidance
We propose SW-Guidance, a training-free approach for image generation conditioned on the color distribution of a reference image. While it is possible to generate...
Get in touch
If you’re interested in collaboration, visiting positions, or consulting, email us and include a short summary of your project.
Email: your-email@example.com
- Suggested subject:
Collaboration inquiry - Include: problem statement, timeline, and relevant links (paper/code).
- We typically respond within 24–48 hours.
Prefer browsing first?
(Replace placeholders with your official email + social links when ready.)