Search papers, labs, and topics across Lattice.
Purdue University
3
0
6
Achieving high task performance with ultra-low-rank adapters, SALT recovers accuracy while slashing memory usage by up to 16x.
Jointly optimizing visual token selection and LLM computation can drastically reduce inference costs while enhancing performance in multimodal tasks.
Trigger color can dramatically alter the success rates of backdoor attacks in federated learning, revealing a critical vulnerability in current defenses.
