Search papers, labs, and topics across Lattice.
University of Cambridge
2
0
4
Achieve ~20% gains on difficult 3D medical image segmentation by explicitly removing noisy activations in U-Net skip connections with a novel proximal-sparse attention mechanism.
Lipschitz-constrained Transformers, built from gradient flows, can provably approximate any Lipschitz-continuous function, offering a path to more robust and stable architectures.
