Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
1
0
2
Forgetting in neural networks is governed more by gradient concentration than by the saliency of selected weights, undermining common assumptions in machine unlearning.