Reproduce “Beyond Sparse Captions: Aligning Slide-Level Text and Patch-Level Vision in Pathology” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Daniel Shao, Dongmin Bang, Luca L. Weishaupt, Sophia J. Wagner, Nic G. Reitsam, Ming Y. Lu, Long P. Le, Richard J. Chen and 1 more
LIMIT adapts the patch-level vision–language model CONCH v1.5 using routine whole-slide reports, without hand-labeled patch locations or trainable image-side parameters. It freezes the image encoder and uses report-conditioned attention to weight patches, while adapting the text encoder. On benchmarks, CONCH-Z reaches 96.7% binary patch-level AUC and 78.8% zero-shot balanced accuracy on organ-constrained TCGA-OT; these results do not establish clinical readiness.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.