PathMem: Toward Cognition-Aligned Memory Transformation for Pathology MLLMs

Jinyue Li, Yuci Liang, Qiankun Li, Xinheng Lyu, Jiayu Qian, Huabao Chen, Kun Wang, Zhigang Zeng, Anil Anthony Bharath, Yang Liu · 2026

Read PathMem: Toward Cognition-Aligned Memory Transformation for Pathology MLLMs by Jinyue Li, Yuci Liang, Qiankun Li, Xinheng Lyu, Jiayu Qian, Huabao Chen, Kun Wang, Zhigang Zeng, Anil Anthony Bharath, Yang Liu (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-09943.txt.

Description

Jinyue Li and coauthors present PathMem, a memory based framework that helps multimodal models reason about pathology slides. Diagnosis needs both visual pattern reading and formal knowledge such as taxonomy and grading rules, which current models apply unevenly. PathMem stores that structured knowledge as long term memory. A memory transformer then moves selected parts into working memory, grounded by the case at hand, lifting report scores.

Catalog record

CREATORJinyue Li, Yuci Liang, Qiankun Li, Xinheng Lyu, Jiayu Qian, Huabao Chen, Kun Wang, Zhigang Zeng, Anil Anthony Bharath, Yang Liu
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE35958
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2603.09943

Browse the full library · About the Tower of Babel