Participants are not allowed to modify the training procedure.

Federated Learning (FL) allows machine learning models to train locally on individual mobile devices, synchronizing model updates via a shared server.

Major challenges in scaling self-training are the choice of label space, pseudo-annotation filtering, and training efficiency.

This paper focuses on the methods that explain what neurons learn during the training process.

Through extensive experiments on prevalent image datasets, we demonstrate the superiority of HASSOD over existing methods, thereby advancing the state of the art in self-supervised object detection. Notably, we improve Mask AR from 20.2 to 22.5 on L VIS, and from 17.0 to 26.0 on SA-1B.