Mingkun Xu

AURA reveals that understanding implicit user intent can dramatically reduce the number of queries needed while enhancing the relevance of responses.

VLMs can achieve state-of-the-art adversarial robustness by iteratively refining visual and textual representations through a closed-loop prompting mechanism, even with frozen encoders.

LLMs can now compress their KV cache more effectively by dynamically synthesizing soft tokens tailored to the input, preserving crucial context that's otherwise lost with static methods.

Even the best vision-language models struggle to diagnose brain tumors from MRI scans, but a new dataset and benchmark reveals a path to significant accuracy gains through instruction tuning.