- 🔭 My research interests focus on discovering the underlying potential of modern models that can empower an increasingly broad range of downstream tasks, spanning perception, reconstruction, generation. I believe that pretrained foundation models possess remarkable intelligence, yet they often lack the specialized capabilities required for real-world downstream applications. This motivates my work toward bridging this gap directly. Recently, I have been deeply inspired by the impressive progress of generative video models. These models appear to encode rich world knowledge, such as physical dynamics and multi-image interactions, capturing what I would describe as the abstract relationships that govern the world. I believe such models have the potential to evolve into \emph{world models} capable of simulating complex real-world behaviors and powering a wide variety of intelligent tasks. I am actively exploring this direction and welcome discussions or collaboration ideas.
- 📫 How to reach me: haosen.yang.6@gmail.com
Pinned Loading
Something went wrong, please refresh the page to try again.
If the problem persists, check the GitHub status page or contact support.
If the problem persists, check the GitHub status page or contact support.