Search papers, labs, and topics across Lattice.
Laboratoire de Recherche d'EPITA, Computing and Mathematical Sciences Division, Mohamed, Zayed University of Artificial Intelligence
4
0
6
Finite-sample guarantees reveal how localized conformal prediction can significantly reduce miscalibration while preserving coverage.
Uniform Diffusion Models aren't optimized by the objective you thought they were, and fixing this mismatch closes the gap with Masked Diffusion Models.
Q-learning regret bounds can be achieved without optimism, but are highly sensitive to the suboptimality gap, motivating a new smoothed exploration strategy.
Skip the expensive gradients: a simple VJP-free approximation lets you edit images and videos with diffusion models just as well as training-heavy approaches.
