TLDR AI · RESEARCH & INNOVATION

ENHANCING SECURITY IN MODELS USING DIFFUSION

arxiv.org29 min read

Summary from the TLDR AI newsletter

Researchers are exploring if "diffusion models" can make models that are already trained to be robust even stronger. A new method, "DiffSmooth", cleans up data using diffusion and then uses a special smoothing process to make the models even more reliable.