Search papers, labs, and topics across Lattice.
University of Central Florida
4
0
9
Safety-aligned text-to-image models may appear effective, but they suffer substantial semantic fidelity losses that standard metrics fail to capture.
ReasonCLIP-58M shows that structured reasoning supervision can dramatically boost CLIP's reasoning abilities while maintaining efficiency.
CaMBRAIN unlocks real-time, continuous EEG analysis with a Mamba-based architecture, finally making long-range brain activity monitoring practical.
Dataset distillation is more effective when you focus on *what the model can currently learn*, not just visual diversity or average training dynamics.
