Exploring the
future of visual
compression.
I study how generative models and learned representations can make image compression more efficient, perceptually compelling, and meaningful.
Curiosity-driven
research.
I am particularly interested in the connection between representation learning, rate constraints, and the behavior of diffusion-based decoders. I aim to ground new ideas in careful paper reading and empirical evaluation.
Better representations make better reconstructions possible.
The questions
that drive my work.
Ideas, exploration,
and ongoing work.
These are ongoing research directions and study activities, not claims of published results or released software.
Research takes
time to develop.
Publications will be listed here.
No publications have been added to this page. Future papers and preprints can be included with verified links.
Let's exchange
ideas.
Interested in generative image compression, learned representations, or related research? Feel free to connect.
Contact links can be enabled by adding your email or GitHub username in profile.js.