Search papers, labs, and topics across Lattice.
School of Electrical Engineering, Korea Advanced Institute of Science and Technology (KAIST), Daejeon, South Korea
2
0
4
Naive attention-based filtering for edge-cloud inference is suboptimal under tight bandwidth constraints; prioritizing semantic diversity in transmitted embeddings yields surprisingly large accuracy gains.
By conditioning on observation residuals, ReCo-Diff achieves more accurate and stable sparse-view CT reconstructions than existing cold diffusion models, even under severe sparsity.
