Long H. Pham

A new method improves fuzzing by focusing on input relationships and adapting strategies.

CCL ensures neural networks maintain accuracy while learning new tasks.

A new method helps protect language models from harmful backdoor attacks.