Search papers, labs, and topics across Lattice.
1
0
2
7
Token-level visual dependence is the key to preventing catastrophic forgetting in multimodal continual learning, enabling models to adapt without losing prior knowledge.
