Three papers published to arXiv on July 3, 2026, address different aspects of memory and reasoning in AI systems.
MMIR-TCM for Traditional Chinese Medicine
According to arxiv.org, researchers introduced MMIR-TCM, a framework designed to address challenges in Traditional Chinese Medicine (TCM) diagnosis. The paper notes that “Traditional Chinese Medicine (TCM) diagnosis, particularly through tongue inspection, faces persistent challenges in subjectivity and reproducibility.” The framework aims to bridge “the semantic gap between visual tongue features and textual reasoning” for clinical tasks including syndrome differentiation and prescription generation.
DRIFTLENS Measures Personalization Effects
A second arxiv.org paper introduced DRIFTLENS, a framework for measuring “memory-induced reasoning drift” in language models. According to the researchers, personalization through stored user attributes “can also change the reasoning trajectory used to justify the response.” The study found that “across four LLMs and 10 user-attribute categories, including age, occupation, and disability, user-attribute memory induces medium-to-large reasoning drift above each model’s pragmatic-noise floor.”
Procedural Memory Distillation
A third arxiv.org paper proposed Procedural Memory Distillation (PMD), which “converts cross-episode signals into reusable procedural memory and distills it into the policy’s weights during training.” According to the paper, PMD improved performance over baseline methods by 3.8-5.5% on SCIKNOWEVAL and 7.9-13.6% on LIVECODEBENCH across tested models.