Search papers, labs, and topics across Lattice.
Institute of Science Tokyo
3
0
6
Misclassifications in biomedical imaging can be drastically reduced with BioMedVR's confusion-aware approach, leading to superior model performance in critical healthcare applications.
Reversing input text can unlock hidden contextual information, leading to richer embeddings and substantial performance gains in training-free LLMs.
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.
