Reproduce “InfoDPP-PAC: Principled Patch Selection for Whole Slide Image Analysis” 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
Prateek Mittal, Ayush Srivastava, Joohi Chauhan
InfoDPP-PAC selects whole-slide image patches by combining teacher-derived relevance, diversity, and a PAC-style rule that stops when residual information gain is low. A Gaussian process estimates patch quality from a small labelled seed set. On 22 held-out slides, it selected 48.9 patches on average under a 300-patch cap, reducing counts 83.7% while retaining 97.9% of composite score; this measures selection quality, not diagnostic accuracy.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.