Search papers, labs, and topics across Lattice.
The Hong Kong University of Science and Technology
2
0
4
Forget scalar rewards: GenEvolve distills structured visual experiences from successful and failed generation trajectories, enabling token-level supervision for self-improving image generation agents.
Ditch raster scanning: this new RWKV-based pan-sharpening method uses semantic prototypes for context-aware token reordering, boosting performance.