Full Professor, University of the Chinese Academy of Sciences
1 paper at NeurIPS 2025
We reformulate the task vector coefficient prediction as a variational inference problem and propose a novel composition framework.
