Reproduce “PathoJev: Efficient Pathology Decisions from Shared Visual Context” 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

Chengyang Zhang, Wenchuan Zhang, Bo Li, Xinyu Liu, Jiaming Yang, Mengran Li, Jie Chen, Wei Ju and 3 more

PathoJev evaluates pathology questions about one whole-slide image by reusing its slide representation and scoring supplied choices in parallel. Its hybrid backbone carries fixed-size recurrent states into separate question branches, while attention still reads the shared prefix. Across three WSI benchmarks, PathoJev-9B achieves 77.2% mean accuracy versus SlideChat’s 71.8%; it supports categorical or binary choices, not open-ended pathology conversations.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.